> For the complete documentation index, see [llms.txt](https://evidano.gitbook.io/ailyze-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://evidano.gitbook.io/ailyze-docs/reference/settings/document-analysis.md). # Document Analysis Settings Every control in document analysis: files and context, methodology, focus groups, the table builder, the review dialog, the results page, and the project list. This page lists every control in document analysis and says what each one changes, screen by screen: picking files, the four setup steps (**Files and context**, **Methodology**, **Focus groups**, **Your table**), creating the project, the results page, and the project list. It is for anyone who has run [Analyze Your First Transcript](/ailyze-docs/getting-started/analyze-your-first-transcript.md) and wants to know what a specific option does. The walkthrough is [Analyze Interview Transcripts](/ailyze-docs/common-research-workflows/analyze-interview-transcripts.md); the ideas behind the table are explained in [Design Your Analysis Table](/ailyze-docs/analyze-data/design-your-analysis-table.md). Numbered red badges in the figures match the numbered rows in the table under each figure. The figures use the guide's invented sample files: three staff interviews and a parents' focus group about a mobile health clinic. Plan rules in one line: the Free plan runs the analysis but unlocks only the first theme and the first question and disables exports and edits on the results page; the only wallet charge is for files beyond the included count. Numbers are on [Plans, Limits & Credits](/ailyze-docs/reference/plans-limits-and-credits.md). Last verified against the app: main @ 531b3b28b, 2026-11-02. ## Screen 1: Pick the files From your account page, on the **Document analysis** card, click **New document project**. The card's **View previous** opens the project list and **Manage files** the file manager. The page is headed "New document project", with **All projects** and **Dashboard** above the heading, and opens on the shared file library described in [Upload and Manage Files](/ailyze-docs/getting-started/upload-and-manage-files.md). Other ways in: in Image to Text, **Use in a document project** saves the transcribed pages to your files as a Word document and opens a new document project with it ticked. With no files of your own and no past projects, the library is one dashed card, "No files yet", with **Upload files** (dropping files on the card opens the upload dialog with them) and **Show sample files**, which shows the full library with the shared samples.

File library with upload, recent, manage and view buttons, source tabs, and search — Typing in the search box filters every folder at once; here it narrows the library to the four sample transcripts.

| # | Control | What it does | Notes | | - | ------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------- | | 1 | **Upload files** | Opens the upload dialog (below). Reads **Add more files** once a file is selected. | Uploaded files are ticked for you. | | 2 | **Select from recent** | Picks from files you used in recent projects. | A red badge counts them. | | 3 | **Manage files** | Opens the file manager. **Back to analyze files** returns. | | | 4 | **Grid view** / **List view** and **Sort by date** | Arrange the library. | **List view** is best for ticking many files. | | 5 | Tabs: **All data**, **Manually uploaded**, **AI transcriptions**, **AI translations**, **AI avatar interviews**, **Demo files** | Narrow the library by where a file came from. A tab appears only when it has files. | **AI avatar interviews** has a **Responses** / **Files** switch: **Responses** lists interview transcripts per AI Avatar project. | | 6 | Search box (Ctrl+K) | Filters files and folders by name. | | | | Checkbox on a file | Adds it to **Files to analyze**. | You can also drag a file from the library onto **Files to analyze**. | | | Bar under the library | Counts the ticked files ("4 files selected", or "for analysis" and "for context"). **Clear all** unticks them; **Continue** scrolls down to the setup steps and shows only while they are off screen. | On a paid plan, files beyond the included count show as "N extra files: USD X". |

Upload Files dialog with Upload files and Paste text tabs, four file sources, and the accepted formats line — Files start uploading as soon as you add them; the line at the bottom lists the accepted formats and the size cap.

| # | Control | What it does | | - | ------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------- | | 1 | **Upload files** tab | Upload from your computer or a cloud drive. | | 2 | **Paste text** tab | Saves pasted text as a TXT file in your library. | | 3 | **Local Files**, **Google Drive**, **Dropbox**, **OneDrive** | Where the files come from. Cloud sources ask you to sign in to that service the first time. | | 4 | "Accepted formats" line | DOCX, PDF, and TXT, with the per-file size cap. See [Supported File Types](/ailyze-docs/reference/supported-file-types.md). |

Paste text tab of the Upload Files dialog — Paste text is the quickest way in for notes or answers you have in an email or a web page.

**Restoring a draft.** Your setup is saved in this browser as you go (the step bar shows "Saved" and how long ago). When you come back, **Continue where you left off?** names the project, says when the draft was saved, and lists what it holds (methodology, files to analyze, supporting files, focus-group participants, table columns, themes, questions, segments) and the step it reopens on. If the files ticked now differ from the draft's, a note says restoring keeps the files ticked now. A reload of the page keeps the files you ticked without asking.

Continue where you left off dialog listing the saved files, columns, questions and segments — The sentence under the name says when the draft was saved and which step it reopens on: the files, the methodology, the focus groups, or the table.

| # | Control | What it does | | - | ----------------- | ------------------------------------------------------------------------------------------- | | 1 | **Restore draft** | Reopens the draft on the step where you left it. | | 2 | **Start fresh** | Discards the draft. Drafts expire after 7 days and live only in the browser that made them. | There is no close button: Esc and clicks outside do nothing, so a stray click cannot discard the draft. ## Screen 2: Files and context Ticking a file opens the setup steps below the library. The step bar lists **Files and context**, **Methodology**, **Focus groups** (only when focus groups are found), and **Your table**; click a completed step to go back to it.

Four files to analyze, the Focus groups select, a typed brief, the method suggestion button and Continue — Everything under Files to analyze is analyzed. Background material such as a proposal or an interview guide goes under Supporting files instead.

| # | Control | What it does | Notes | | - | ----------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | **Preview file** (eye) | Opens the file in a preview. | | | 2 | **Use as context instead** (arrow) | Moves the file to **Supporting files**, where it is read for setup only. **Put back** returns it. | PDF, DOCX, TXT, XLSX, and CSV files. At least one file must stay in **Files to analyze**. | | 3 | **Remove file** (x) | Takes the file out of the project (not out of your library). | **Clear all** at the top right empties the list, supporting files included, after "Remove all files?". | | 4 | **Focus groups:** | **Auto-detect** (default), **Yes, some are focus groups**, or **No focus groups**. | Focus-group transcripts get one row per participant. Leave it on **Auto-detect** unless you know. With **Auto-detect**, a note beside the menu says "Checking…", then how many focus groups it found ("1 focus group found", "None found"). Switching to **No focus groups** after participants were found offers **Undo**. | | 5 | **Supporting files** drop zone | Drop or browse for terms of reference, proposals, evaluation frameworks, interview guides, codebooks, or briefs (PDF, DOCX, TXT, XLSX, CSV). | Optional; up to 5 files, "Read by the AI to set up your project, not coded or quoted." An attached codebook is imported word for word as the theme hierarchy. An uploaded file is also saved to your library. Supporting files have **Preview file** too. | | 6 | **Input manually** | A free-text brief: the project, the material, and what you need. | Optional; the counter shows how much of the 6,000 characters you used. A longer paste is cut, with a warning. | | 7 | **Suggest a method from my files and notes** | Reads your supporting files, your brief, and the file names, suggests a methodology, and opens **Methodology** on it. | Needs at least one supporting file or some typed text. It does not draft the table yet. | | 8 | **Continue** | Goes to **Methodology**. If you have not chosen a methodology yet, the AI first suggests one from the file names and the opening of a few files (plus any supporting files and brief), and the step opens on its pick; when no pick comes within 15 seconds, the step opens on the catalogue. | Disabled until at least one file is listed. **Cancel** in the overlay stays on this step. | | | File note "Very little text (N words): results may be thin" | Flags a file with almost no text, such as a cover page or an empty template. | A warning only; the file can still be analyzed. | | | "Scanned images: N images found" | Counts pages without a text layer among the files to analyze. | They are read with OCR, at a per-page charge, or left out, when you start the analysis (Screen 7). | | | Wallet line | On a paid plan, when the extra files cost more than your wallet holds: "N files over the 50 included cost USD X; your wallet has USD Y." with **Top up now**. | Shown before you build the table, not only at the end. | With exactly one file that is not a focus group, **Analyze just one file?** appears when you leave this step (with **Continue** or the suggestion button): **Upload more files**, **Select more files** (back to **Files and context** and the library), or **Skip and analyze 1 file only** (you are not asked again for that file). Comparing segments needs at least two files.

Finding a method for your files overlay with Your notes, Files to analyze and Choosing a method stages — The suggestion usually takes a few seconds; Cancel (or Esc) keeps you on this step. The first stage is named for what you added: Your notes, Supporting files, or Supporting files and notes.

{% hint style="warning" %} **Costs money:** On a paid plan, files beyond the included count are charged per file; the bar under the library, the files step and the review dialog show the extra files and their cost first, and **Confirm charge** shows the amount and your wallet balance before the project is created. Reading scanned pages with OCR is charged per page from your wallet. The Free plan cannot go past the included count. Rates are on [Plans, Limits & Credits](/ailyze-docs/reference/plans-limits-and-credits.md). {% endhint %} ## Screen 3: Methodology The methodology tells the AI what kind of analysis you are doing, which shapes the table it drafts. It is optional.

Methodology step opened on the AI's pick, Framework analysis, with its reason, See other methods and Next — The step opens on the AI's pick; keep it, or see the other methods before you click Next.

| # | Control | What it does | | - | ---------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | Lead line | "The AI suggested a method from your files and notes. Use it, or see other methods." With the catalogue open it reads "… Keep it, or choose another." | | 2 | **Suggested for your files** | The AI's pick with a one-line summary and the AI's reason. A method the AI wrote for your project shows as "A method written for your project", with its text open for editing. | | 3 | Info icon (overview and citations) | Opens an overview with key references and citations. | | 4 | **See other methods** | Opens the catalogue (below). The AI's pick stays selected and is marked **Suggested by AI**, with its reason under the list while it is selected. | | 5 | **Next** | Continues with the selected methodology and starts the draft. Once you have browsed, the footer names your pick ("Selected: Framework analysis"). |

Methodology with no suggestion: search, Common methods, Browse all methods by field, No specific method, disabled Next — With nothing selected, Next stays disabled until you pick a methodology or write your own; No specific method goes on without one.

With nothing suggested, the lead reads "Choose how your files should be analyzed. With no specific method, the AI designs the table from your files and notes." | # | Control | What it does | | - | ------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | 1 | Search box | Searches every methodology by name, alias, or keyword, for example "grounded theory", "outcome harvesting", or "thematic analysis". Enter picks the top result; Esc clears the search. With no match, **Write your own** opens a text box for your own method. | | 2 | **Common methods** | The methods most projects choose (reflexive, codebook and framework thematic approaches, conventional content analysis, document analysis), with the AI's pick first when there is one. | | 3 | **Browse all methods by field** | Replaces the list with three columns, Domain, Family, and Methodology: **Academia**, **Evaluation**, **Market research**, **UX and product research**, **Public policy and consultation**, **Health and social care**, **Legal**. Picking a domain fills the families, and a family fills the methods (aliases in grey). | | 4 | **Describe your own method instead** | Replaces the catalogue with a text box for your own method (below). | | 5 | **Back** | Returns to **Files and context**. | | 6 | **No specific method** | Continues without a methodology; the AI designs the table from your files and notes. | | 7 | **Next** | Continues with the selected methodology. |

Methodology overview dialog with Overview, Key references, Citations, Close and Use this methodology — Use the overview to check that a method fits before you build on it; the citations are ready to paste into a methods section.

The overview's **Use this methodology** selects the method; **Close** leaves the selection as it was. Citation links open in a new tab. If the overview does not load, click **Try again**.

Methodology search results for outcome — Search results list each method with its summary, so similar names are easy to tell apart.

Your methodology text box with a character counter — A method of your own replaces the catalogue entry as the AI's starting point.

| # | Control | What it does | Notes | | - | ----------------------- | --------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------ | | 1 | **2. Your methodology** | Paste or type the method you follow: its steps, the framework, checklist or coding scheme it uses, and what each document should yield. | Up to 12,000 characters. A longer paste is trimmed ("Trimmed to 12,000 characters"). |

The methodology catalogue in detail

* **Academia:** thematic and content approaches (codebook, reflexive, framework, template, directed and conventional content analysis, document analysis, qualitative description); evidence synthesis (data extraction, screening, quality appraisal such as CASP, JBI, MMAT, risk of bias, GRADE-CERQual, and reporting checklists such as PRISMA, CONSORT, COREQ); grounded theory; phenomenology; narrative, incident and biographical; discourse and language; ethnography and observation. * **Evaluation:** theory-based and causal evaluation (theory of change, logframe, realist evaluation, contribution analysis, process tracing); evaluation criteria (OECD DAC, Kirkpatrick, social return on investment); outcome-centered approaches (outcome harvesting, outcome mapping, most significant change, ripple effects mapping); case-based and comparative; implementation and process (CFIR, RE-AIM, process evaluation); participatory and strengths-based. * **Market research:** consumer interviews and focus groups, jobs to be done, journey mapping, brand, concept, message and pricing tests, reviews, win/loss and churn interviews. * **UX and product research:** interview synthesis, diary studies, contextual inquiry, feedback triage, usability tests, heuristic evaluation, think-aloud. * **Public policy and consultation:** consultation responses, stakeholder analysis, complaints, deliberation quality, policy document review, frame analysis, narrative policy framework, policy transfer. * **Health and social care:** patient experience feedback, safety incident review, case note review, significant event analysis, health needs assessment, Donabedian quality framework, guideline appraisal, GRADE evidence-to-decision. * **Every domain:** **Split a document bundle** (family "Document bundles"), described below.

**Split a document bundle** is for PDFs that bundle many separate documents, such as a disclosure bundle or a records dump. With it selected, **Next** opens a split workspace instead of the draft: you agree how to split the pages (where a new sub-document starts, what to keep together), click **Confirm and start splitting**, then review the sub-documents' page ranges (**Merge into previous**, **Start earlier**, **Exclude**, and so on) and send the ones you want to your files with **Send N to a project**. They land in a "(split …)" folder, and **Create a project from these N** starts a document project with one row per sub-document. It works on PDF files only; you can leave the page while it splits, and you get an email when the sub-documents are ready to review. Evidano starts checking the files for focus groups as soon as you tick them. If the check is still running when you leave **Methodology**, "Checking N documents for focus group discussions" appears; **Skip: treat as no focus groups** cancels it and turns focus groups off (you can turn them back on under **Files and context** › **Focus groups**). With **No focus groups** chosen, no check runs. If the check fails, a warning says so and the **Focus groups** step asks which files are focus groups. ## Screen 4: Focus groups This step appears only when focus-group transcripts are found (or when you chose **Yes, some are focus groups**, or the check could not run).

Focus groups step with the detected file ticked, one row per participant chosen and four participants identified — Check the participant list: the moderator should not be on it, and every speaker you want counted should be.

| # | Control | What it does | | - | -------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | Detected files | "Tick the files that are focus group discussions. AI pre-ticked the ones it detected." Untick any that are one-to-one interviews. With none ticked, the button reads **Continue without focus groups**. | | 2 | "Treat each focus-group participant as a separate respondent?" | "No, one row per transcript (e.g., 3 of 10 documents said it was effective)" counts each file once; "Yes, one row per participant (e.g., 3 of 20 participants said it was effective)" gives each participant a row. | | 3 | **Identify participants in focus group discussion documents** | Shown with "Yes": runs the participant detection. | | 4 | Participant list | Headed "Participants in" and the file name. One box per participant the AI found; the minus removes a name, the plus adds an empty box right below it. A file with no speakers found offers **Add speaker**. Fill in or remove empty boxes to continue. | | 5 | **Continue to output configuration** | Saves your answer and starts the draft, or returns to the table when it is already drafted. **Back** returns to **Methodology**. | With one row per participant, each focus-group participant becomes a row of the table, so the builder later says "Each row is one focus-group participant", or "Each row is one document or focus-group participant" when interviews are in the mix. ## Screen 5: Drafting your project The draft starts when you leave **Methodology** (with **Next** or **No specific method**), or with **Continue to output configuration** on **Focus groups**. It usually takes under a minute.

Drafting your project dialog with a timer and five progress stages — The stages show what the draft is built from; nothing is created yet.

| Control | What it does | | ------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Timer and stages | Progress through **Methodology**, **Files to analyze**, **Supporting files**, **Supporting files and notes**, or **Your notes** (named for what you added; left out when you added nothing), **Objectives and unit of analysis**, and **Your table**. | | **Cancel** (or Esc) | Stops the draft; you stay on the step you came from, with everything kept. | If you go back after the draft, change the methodology, the files, or the context, and leave **Methodology** again, **Redraft the output configuration?** asks first: **Redraft** drafts again from the methodology and your documents; **Keep my output configuration** returns to the table you have. If nothing changed, **Next** goes straight back to the table. Reopening **Focus groups** from the table also returns to it without a redraft. ## Screen 6: Your table (the table builder) The draft opens as a report section, the unit of analysis, and two tables, top to bottom: **You will receive a report**, **What is your unit of analysis?**, **You will receive themes generated from the entire dataset**, and **You will receive a table of answers and metadata**. Read [Design Your Analysis Table](/ailyze-docs/analyze-data/design-your-analysis-table.md) for when to use each column type.

Builder bar with back, project name, undo, Saved, Review and start analysis, and the Why this draft notice — Read the Why this draft notice first: it says what the AI understood the project to be and what it proposes to extract.

| # | Control | What it does | Notes | | - | ----------------------------- | ----------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | Back arrow | Returns to **Methodology** (or **Focus groups** when that step ran). | The table is kept. **Back** at the bottom of the page does the same. | | 2 | Project name | The name shown on the project list and report. | Pre-filled by the AI; up to 150 characters, with a counter near the limit. Left empty, it reads "Untitled (the AI will name it)", and the AI names the project when the analysis finishes. You can also name it in the review dialog. | | 3 | **Undo** / **Redo** | Steps back or forward through every change in the builder. | Ctrl+Z and Shift+Ctrl+Z (Cmd on a Mac). | | 4 | "Saved" | Confirms the autosave to this browser and says how long ago. | Wide screens only. | | 5 | **Review and start analysis** | Checks the configuration, then opens the review (Screen 7). | Also at the bottom of the page. Ctrl+Enter does the same. While the AI drafts the themes it reads "Drafting themes and sub-themes…" and waits. | | 6 | **Why this draft:** | The AI's reasons for what it proposed, for example which columns it left under the table. | **Hide** folds it away; "Why this draft" in the bar shows it again. | ### You will receive a report

Report options: summaries, group comparison, question mapping, report language, word cloud and the standard outputs line — The outputs every report has are folded under one line; only the choices stay in view.

The line under the heading says what one row, and one count, will be, for example "Each row is one document. Codes count documents; questions are answered for each document." | # | Option | What it adds | Default | | - | ---------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | | **Executive summary** | A summary of the whole project. | On. | | 1 | Pencil (**Instructions for the executive summary**) | How the summary should read, for example "two paragraphs for a minister". | Often pre-filled by the AI; up to 15,000 characters. Clicking it on an unticked summary ticks it again. | | 2 | **Summary of each document** (or **Summary of each transcript**) | One summary per file. | On. | | 3 | **Compare groups (e.g., by gender, year)** | A comparison section inside every theme and question, by the segments you set up. | On when the draft has segments. The AI tags the files with them as soon as the builder opens. The pencil (**Edit the groups to compare and tag the files**) opens the tagging dialog (below) and is marked in amber with what is still missing while some files are untagged. A value no file carries gets a note with **Remove unused values**. | | 4 | **Research questions mapped to themes** | Answers each question through the themes it is mapped to. | Off. Its hint reads "Recommended when your themes answer your research questions"; needs both themes and questions. Ticking it opens the mapping (below); "N unmapped" beside it counts questions without a theme. | | 5 | **Report in** | The language of themes, summaries, and answers. Quotes stay in the original language. | English. | | 6 | **Word cloud** | A word-frequency visualization. | On. | | 7 | "Also included: N standard outputs" | Unfolds the outputs every report has: **Chatbot to generate custom reports and analyses**; a summary per theme and sub-theme and per research question; each document (or participant) summarized per theme and per question; the hierarchy chart / mindmap, frequency bar charts, and the code co-occurrence table and network. | Always included. Items marked with an asterisk come whenever themes or questions are extracted; the charts need at least two codes. | When questions are mapped to themes, the two per-question summaries are listed unticked and marked †: the summaries are written for the themes instead. Untick **Research questions mapped to themes** if you want a summary for each question.

Research questions mapped to themes dialog listing each question with its themes and Map all again with AI — The AI maps the questions as the dialog opens; change any match by hand, then click Done.

| # | Control | What it does | | - | ------------------------- | ------------------------------------------------------------------------------------------------------------------------------ | | 1 | **Map all again with AI** | Redoes the matching, with **Undo** in the message. Questions without a theme are answered directly from the documents instead. | ### What is your unit of analysis? Asked once, above both tables: a row is the same unit in each. The sentence **Each row is one …** says what one row of the output is; click the word to change it.

What is each row popover with A document or focus-group participant, the example units and Something else — Most projects keep one row per document (or per participant for focus groups); pick something else only when each document holds many units, such as outcomes.

| # | Control | What it does | | - | ---------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | | **A document**, **A focus-group participant**, or, with interviews and focus groups mixed, "A document or focus-group participant" | One row per file, per participant, or per participant and per other document. | | 1 | **Or something the documents describe** | **Program**, **Intervention**, **Outcome**, **Recommendation**, **Case**, **Organization**, **Stakeholder**. Picking one replaces the unit and adds a text column for it; **Undo** in the message takes it back. Switching away again removes that column ("The Outcome column is removed."). When columns already have descriptions, a message offers **Rewrite** to rewrite them for the new unit. | | 2 | **Something else…** | Type any other unit, for example "Finding", and press Enter. | | | **evidence from** | Appears after the unit for units that are not documents: **One or more files** (a row can cite several documents) or **Exactly one file**. Some methodologies fix it to exactly one file. | ### You will receive themes generated from the entire dataset This section appears when the project has a theme hierarchy. If yours has none, **Add** › **Add theme** above the table of answers and metadata adds one.

Themes section with the AI notice, numbered draft themes, Edit themes and sub-themes, and the sample table — Draft themes are a starting point: rename, merge, or delete them before you start the analysis.

| # | Control | What it does | | - | ------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | AI notice | "AI has suggested a draft set of themes and sub-themes for your review. Edit to lock in your choices, or keep them unchanged for AI to validate and refine during the full analysis." **Hide** folds it into a link, "About these AI-suggested themes and sub-themes". | | 2 | "Edit themes and sub-themes" ("Edit themes" with two levels) | Opens the theme set-up (below). | | 3 | **Or add codes manually** | Opens the set-up with a first code started and the cursor in it. Otherwise "The AI creates the codes (short labels attached to quotes) during the analysis; edit them in the project's Codebook afterwards." | | 4 | **Sample table output** | The shape of the theme columns (Theme, Sub-theme, Code), with sample values that are not results. | The AI may still be drafting the themes when the builder opens ("Drafting themes and sub-themes…"); they appear here when ready, and **Cancel** stops it. Meanwhile **Review and start analysis** shows the same words and waits. If the draft fails, a message says so with a way to try again.

Themes set-up with no themes: Levels, Configuration required, Generate draft themes and sub-themes now, and Add manually — Nothing is analyzed until the themes are drafted or added, so the empty set-up is marked Configuration required.

| # | Control | What it does | | - | -------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------- | | 1 | **Levels** | Two or three levels (see the set-up below). | | 2 | **Configuration required** | The section still needs themes. | | 3 | **Generate draft themes and sub-themes now** | Opens the generator (below). | | 4 | **Add manually** | Shows four ways to bring your own themes: **Type**, **Paste**, **Upload**, **Use previous** (described under the set-up below). |

Generate draft themes and sub-themes dialog with pre-filled instructions, Cancel and Generate now — The instructions are pre-filled from your brief; edit them to steer the themes, for example toward barriers or a particular group.

The instructions take up to 3,000 characters. **Generate draft themes and sub-themes now** reads the selected files and fills the section; it creates nothing. When themes already exist, the dialog warns "Replaces your current N themes, including your edits. Undo brings them back."

Theme set-up with Levels, Regenerate, theme cards, Replace with, Codes, Add instructions and Done — Each theme and sub-theme has an optional description; the sparkle icon writes one with AI.

| # | Control | What it does | | - | ------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | **Levels** | **2 levels** (themes and codes) or **3 levels** (themes, sub-themes, and codes). Switching keeps the themes; only the level below them goes, with **Undo**. | | 2 | **Regenerate draft themes and sub-themes** | Opens the generator again; the result replaces the current themes. | | 3 | **+ sub-theme** / **+ code**, x | Add a level underneath, or remove a row. **+ Add theme** sits under the last theme. Arrows on hover move a row up or down; removing one offers **Undo**. | | 4 | **Replace with** | **Paste** (a list, a codebook, or a paragraph; the AI arranges it), **Upload** (PDF, Word, text, or spreadsheet), or **Use previous** (an earlier project's themes, codes, and definitions). **Type** is offered only on an empty set-up, under **Add manually**. | | 5 | **Codes** | "Let the AI create codes during the analysis (edit them in the Codebook afterwards)" (default) or **Add manually now**. With codes of your own, **Use both AI-generated and manually entered codes** keeps the AI's additions. | | 6 | **Add instructions** | Guidance for the whole hierarchy, for example "Keep parents' and staff views apart". It applies when the analysis codes the documents; to steer the draft itself, use **Regenerate draft themes and sub-themes**. | | 7 | **Done** | Closes the set-up (Esc does the same). | Theme and sub-theme names need at least 5 characters and must all differ. A problem shows on the theme itself, for example "Too short: use at least 5 characters" or "Needs at least one sub-theme". A project holds up to 150 themes, sub-themes, and questions together. ### You will receive a table of answers and metadata Everything the AI extracts besides the themes: a column per thing you want from each row.

Table of answers and metadata with Reorder, Add, a group column, a column type, sample rows and More — The grey sample rows only show the shape of the output; they are not results from your files.

| # | Control | What it does | | - | ------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | **Reorder** | Opens **Reorder columns** (below). | | 2 | **Add** | **Add column**, **Add question**, **Add theme**, and **Add group to compare** (menu below). The **+** after the last column opens the same menu; **Add column** asks for a name there ("Type a name, press Enter"). | | 3 | Segment column | Its name and "Segment · N values". Click it to edit the segment and tag the files; hover for **Edit group to compare** and **Remove group to compare**. Shown when **Compare groups (e.g., by gender, year)** is ticked and each row is a document or participant. It reads "Tagging files with AI…" while the AI tags the files; the sample rows then show each file's own value, or "Not tagged". Otherwise hover a file name to see its segment values. | | 4 | Column type | Click it to change the type (menu below). Click the name to rename. Hover a header for **Edit column** (pencil) and **Remove column** (bin). | | 5 | Sample output | Up to three illustrative rows under "Sample output · for illustration only"; the first row shows sample evidence, the way real results cite quotes. | | 6 | **More** | Scrolls right to the remaining columns. | | | **Describe N columns with AI** | Appears when columns have no description; the AI writes them. | Click a header to open it in place: simple and choice columns edit their description there (choice columns also show their options); a column with none says "Needs a description"; and question columns note "AI generates codes based on your question, then synthesizes the findings into an overall answer." **Done** closes it. A column name that matches a compared segment asks first ("“District” is already a group you compare by.", **Add anyway**).

Add menu with Add column, Add question, Add theme and Add group to compare — Add opens on hover or click.

#### More columns for a method Outside literature reviews, the draft keeps a methodology's core columns in the table and puts its other columns under the table, so a first draft stays short. A column your brief names, or one that a linked column or a segment relies on, stays in the table. The draft note says "Framework analysis's other columns are one click away, under the table." (with your method's name).

More columns for Framework analysis, open under the table, with one button per column and Add all — Each button adds one column after the table's other columns; hover a button to read the column's description.

| # | Control | What it does | | - | ------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | **More columns for Framework analysis (N)** | Opens or closes the list of the method's other columns, one **+** button per column. | | 2 | **Add all N** | Adds every column in the list. Added columns go after the table's other columns, before the questions and the theme levels, and leave the list. | #### Column types

Column type menu listing eight types, each with a grey example — Every type shows a grey example of what the AI will write.

| Type | What the AI writes in each row | Example | | ----------------------------------------- | ----------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------- | | **Text** | One free-text value. | "Account description" | | **List of text** | Several short values. | "Every barrier named" | | **Single choice (pick one)** | One option from a list. | "Overall tone": Positive, Negative, Neutral, Mixed | | **Multiple choice (pick all that apply)** | Every option that applies. | "Data sources used" | | **Question & Code** | A written answer to the question, coded into answer codes and summarized across files. | "What barriers to using the clinic does this account describe?" | | **Number** | A number exactly as stated, never estimated. | "Number of villages served" | | **Date** | A date as written. | "March 2024" | | **Yes / No** | Yes or No. **Allow “Unclear”** adds a third answer for when the document does not say; off leaves the cell blank. | "Mentions transport costs" | Changing the type of a described column, or renaming it, offers **Rewrite** for its description. A theme hierarchy is not a column type: add one with **Add** › **Add theme**. A table holds up to 30 data columns. #### Edit column **Edit column** on any header except the theme levels opens the column window. The theme levels' pencil opens the theme set-up instead.

Edit column window for a Text column with name, type, description, instructions, link, extraction pass and Done — The description is the instruction the AI follows for this column, so it is worth reading and tightening.

| # | Control | What it does | Notes | | - | ---------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------- | | 1 | Arrows (**Previous column** / **Next column**) | Step to the previous or next column. | | | 2 | **Name of column** | The header in the table and report. | | | 3 | **Column type** | The column type (menu above). | | | 4 | **Column description** | What the AI should write in this column. The sparkle icon (**Write with AI** / **Rewrite with AI**) drafts it for you. | A check before the summary asks for descriptions on data columns that have none. | | 5 | **Add instructions** | Extra guidance, for example "Focus only on the Results section." | | | 6 | **Link to another column** | Generates the options of linked columns together, so they fit each other (for example context, mechanism, and outcome). | Linking a column that is not a choice column turns it into one. | | 7 | **Extraction pass** | **Core** columns are read in a first pass over each document, **Additional** columns afterwards. | The first five data columns are core by default; the unit column is always core. | | 8 | **Move** | Moves the column left or right. **Remove column** sits next to it. | | | 9 | **Done** | Closes the window. | |

Edit column window for a Question and Code column with Answers are coded as and the two Codes choices — A question column both answers in words and codes the answers, so it feeds the frequency charts too.

| # | Control | What it does | | - | ------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------- | | 1 | **Answers are coded as** | **Single choice (pick one)** or **Multiple choice (pick all that apply)**: whether one answer can carry one code or several. | | 2 | **Let AI generate codes later** | The AI finds the codes while it reads (default). | | 3 | **Specify my own codes manually now** | Lets you set the codes up front, with the same editor as choice options (below). |

Edit column window for a Single choice column with its options, Use both, Add option and the import buttons — Descriptions are optional; add them when two options could be confused.

| # | Control | What it does | | - | ------------------------------------------------------ | --------------------------------------------------------------------------------------------------------- | | 1 | **Let AI generate options later** | The AI proposes the options as it reads (default). | | 2 | **Specify my own options manually now** | Opens two empty rows; Enter adds a row. Drag to reorder; the x removes one. | | 3 | **Add option** | Adds a row. **Add descriptions** gives each option an optional description; **Sort** puts them A to Z. | | 4 | **Or bring options in:** | **Paste** (a list or a codebook), **Upload** (CSV, spreadsheet, PDF, or Word), or **Copy from a column**. | | 5 | **Use both AI-generated and manually entered options** | Lets the AI add options yours do not cover. Off by default. | #### Segments Segments (for example District or Stakeholder type) let the report compare groups. Each file, or each participant, is tagged with one value per segment.

Which participant or document belongs to which group dialog with four ways to tag, a warning, and a Respondent group column — Untagged files are listed in amber at the top; with focus-group participants as rows, the title reads "participant or document".

The dialog is titled "Which file belongs to which group" ("participant" for focus groups). | # | Control | What it does | | - | ------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | **Auto-populate with AI** | The AI reads each file and tags it. | | 2 | **Add manually** | Adds a segment column from a list (Gender, Time Period, Age Group, Region, Income Level, Education Level, and more) or a name of your own. | | 3 | **Import XLSX/CSV** | Reads a sheet of file names and segment values. | | 4 | **Use previous** | Reuses segments from an earlier project. | | 5 | **Optional instructions** | Guides the comparison. | | | Tagging table | Click a cell to add values. Column headers filter and the x removes a segment. **Done** saves, keeping values nobody is tagged with yet, and offers to remove them (**Remove it** / **Remove them**); while the AI is still tagging rows, **Done** waits for them. **Cancel**, or closing with nothing to compare, unticks the comparison. A segment with one value gets a note, not a block. | When the AI tags the files on its own, a message says how many it tagged, with **Review** to check them. #### Questions and reordering

Add questions dialog with two pasted questions, Convert into list, Upload and Use previous — Each question becomes a Question and Code column.

| # | Control | What it does | | - | --------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | Paste box | Paste or type questions, one per line or as prose. | | 2 | **Convert into list** | Splits the text into separate questions; untick any you do not want. | | 3 | **Upload** | Reads questions from a file. | | 4 | **Use previous** | Reuses questions from an earlier project. | | | **Add N questions** | Adds them to the table. A project takes up to 35 questions; a note appears only when no more fit. Questions already in the table are greyed out ("Already in the table") and left out. |

Reorder columns dialog listing columns with their types — The row column stays first, then the files column and the segments (locked); theme levels are not listed. Everything else can move, one at a time or several ticked together.

## Screen 7: Confirm and start the analysis Click **Review and start analysis** in the builder bar or at the bottom of the page. When something is missing, nothing is sent. A message at the bottom of the screen names the first problem ("+N more" counts the rest), and the column, section, or report row involved opens or is highlighted in red. Examples: "Add the themes before creating the project", "Comparing groups needs at least two files; untick it or add files", and, for files without a segment value, "… Tag the files, or untick “Compare groups”." with **Tag files**. A theme problem names the theme, and the theme set-up opens on it. While the AI is still mapping the questions to the themes, the message asks you to wait a few seconds.

1 column needs a description dialog with an empty description box, Write it with AI and Continue — Question columns can stay without a description; other data columns need one.

| # | Control | What it does | | - | --------------------------------------------------------- | ----------------------------------------------------------------- | | 1 | **Write it with AI** (**Write them with AI** for several) | Writes the missing descriptions; edit any, then continue. | | 2 | **Continue** | Goes on to the summary once every required description is filled. |

Create project dialog with the Project name field, the list of rows, columns, themes and report, Keep editing and Confirm — Read the summary as a checklist: the unit, the columns, the themes, the questions, and the report.

| # | Control | What it does | | - | -------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | | Title and **Project name** | **Create “Project name”**, or **Create project** when unnamed, with "Here is what the AI will do with every document." Type in **Project name** to name the project here; left empty ("Untitled (the AI will name it)"), the AI names it when the analysis finishes. | | | Summary | What the AI will do: read the documents (and the focus-group participants between them), the cost line on a paid plan (the extra files and their cost, "confirmed on the next screen", or **No charge** for up to the included files, with your wallet balance), the rows, the data columns, your options, the themes, the questions and their mapping, the segments, "Back every theme, answer and value with quotes and an explanation.", and the report it writes. The footer gives a run-time estimate and says the analysis runs in the background. | | 1 | **Keep editing** | Goes back to the builder. | | 2 | **Confirm configuration and start analysis** | Creates the project and starts the analysis. | After that: 1. If some themes, questions, or documents would be coded exactly as in an earlier run, **Same inputs, same coding** explains that the coding is deterministic and only the summaries are written afresh; **Start analysis** goes ahead, **Go back** returns. 2. If files to analyze have scanned pages, **Scanned images found** says how many pages need OCR and what it costs: **Run OCR and start**, **Don't use OCR…** (tick the files to leave out, then **Remove N files and start**), or **Back to builder**. 3. On a paid plan with more files than the included count, **Confirm charge** shows the amount and your balance before and after; **Start analysis** confirms it. 4. "Analysis starting" counts through the stages until the project is accepted; the bar then reads "Sent" and the badge "QUEUED", and **View project** opens the new project (on its own after a short countdown). If the project cannot be created, the message gives the reason. You get an email when the analysis is done. ## Screen 8: The results page The left rail, **Contents**, lists every section; click one to jump to it. **Hide contents** folds the rail and **Show contents** opens it again. Under the list, **Expand all cells** shows every summary and quote list in full.

Results header with Back to projects, the project name, the documents and Completed chips, View changes and Project details — The chips under the name record what the analysis ran on and when.

| # | Control | What it does | | - | ------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | **Back to projects** | Returns to the project list. | | 2 | Pencil (**Edit project name**) | Renames the project. | | 3 | Documents chip | **Project documents**: the file list, with copy and CSV buttons for the file names and **Add files**. A research documents chip appears when you added context documents. | | 4 | **Completed** | When the analysis finished. | | 5 | **View changes** | The **Change log** of edits, with restore where available. | | 6 | Info icon (**Project details**) | The project ID, the Evidano model version, and when the project was created. |

Project documents popover listing three files and Add files — The quickest way to check which files a finished project used.

Change log dialog saying No changes yet — Themes, questions, codes, and files you add, edit, or delete are logged here.

Contents rail listing the sections and each question with its subsections — Every theme and question is numbered, with its subsections underneath.

While the analysis runs, a banner says "Project is still in progress" with the current stage, the results so far are marked **Partial Results**, and the actions are disabled. A section the analysis has not reached yet says "Still analysing. These summaries appear here once the analysis reaches them." with **Check again**. **Actions**

Action row with Export, Ask AI, Share and Edit project — Free plan users see the downloads and edits greyed out with "Only allowed for paid users".

| # | Action | What it does | Plan | | - | ---------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- | | 1 | **Export** | A menu: **Word report (DOCX, ZIP)** or **Excel workbooks (XLSX, ZIP)**. Each opens a download dialog (below). | Pro. | | 2 | **Ask AI** | Opens the project chat beside the results. See [Chat with Data Settings](/ailyze-docs/reference/settings/chat-with-data.md). | Typing needs Pro. | | 3 | **Share** | Opens **Share project** (below). | Pro. | | 4 | **Edit project** | A menu: **Add file** (adds documents and re-runs the parts that change; with segments, new files must be tagged), **Add theme** and **Add question** (the theme editors, then **Save to project**), and **Delete files** (removes files and re-runs without them; **Components to update** chooses which summaries are redone; at least one file must remain). | **Add file** on all plans, up to the plan's file count; the others Pro. Disabled while an analysis runs. | The download dialog (**Download Word report** or **Download Excel report**) asks **What to include**: **Brief** (the executive summary and the summary of each theme and question; the Word report adds their quotes as endnotes; no annexes), **Full** (every section and annex the project has), or **Custom** (tick the **Sections** and **Annexes**). **Documents** chooses all documents or **Only the documents I choose**. A theme, question, or annex with nothing in it yet is left out of the ZIP.

Add Files dialog with the accepted formats line — New files are analyzed with the same table, themes, and questions as the rest.

**The sections** The page reads top to bottom: the executive summary, the table of all rows, the codebook, the themes and questions, the summary of each document, then charts and tables.

Executive Summary with its outline, Show full summary, and its action row — The outline comes first; Show full summary opens the whole text.

**Executive Summary** opens on its outline, with **Show full summary** for the whole text, stamped Generated, Edited, Refreshed, or "Updated for added files". Its icons offer **Chat with this analysis**, **Edit**, and **Copy**; **More actions for this section** holds **Rewrite with AI** (optionally in another language or with new instructions) and **Delete**. After you edit the themes, "Summary may be out of date: the themes were edited after it was written (date)." appears above it; close it, and it returns after the next theme edit.

Table of all participants: one row per participant, a Role column, question columns of codes, Download CSV and the pager — The first column names each row; here, one row per participant. With outcomes as the unit, it reads "All outcomes".

The table of all rows is headed after the unit, for example "All documents" or "All outcomes", and its block is "Table of all documents" (or outcomes, participants). | # | Part | What it shows | | - | ---------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | Row column | **Outcome**, **Participant**, or **Document**, depending on the unit you chose; a document row shows the document's label, with its file name on hover. Click a row for its quotes. The eye on a column header shows all of that column's quotes grouped by row. | | 2 | **Download CSV** | Downloads the table (Pro). | | 3 | Pager | **Rows** per page and a count, for example "1–15 of 15 outcomes". | | Table control | What it does | | ------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Search box | "Search documents, themes, questions, or codes..." filters the rows. Your browser's Find does not see rows that are scrolled out of view, so use this box. | | **Columns** | Shows or hides columns. | | Column and row edges | Drag a header's right edge to widen a column, or a row's bottom edge to change its height. Double-click resets. | | **More** | Appears when columns run off the right; scrolls to them. A left arrow scrolls back. | | Count and filter icon under a header | "N documents" shows only the rows tagged there; the filter icon sorts and filters. | | Code chip | Opens the code's quotes, best quote first; **+ View more context** shows the document's summary when there is one. Focus-group participant rows show each participant's own codes. | | Long text lists | Show three items and **+ View N more**. With edit rights (Pro), **Add quote to “…”** adds evidence to a row. | With two or more choice columns, a "Cross-tabulation" of them follows the table, then one block per single-choice, multiple-choice, or Yes / No column that found values (below). Counts are rows (documents, participants, or units).

Codebook with Edit, the more-actions menu, Download, the search box and the link to the table of all documents — The Codebook is open when the page loads; the header holds the editing tools.

| # | Codebook control | What it does | Plan | | - | ----------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- | ---- | | 1 | **Edit** | Opens the codebook editor (below). | Pro. | | 2 | **More codebook actions** (⋯) | **Calculate AI accuracy** (below) and **Edit groups to compare**: change segment values and tags, then save; the comparison updates afterwards. | Pro. | | 3 | Download | **CSV: one row per coded quote** (or the themes matching your search) and **ZIP: Excel and charts**. | Pro. | | 4 | Search box | "Search codes, quotes, and elaborations..." | | | 5 | Table link | "Codes counted in documents ·" and a link to the table of all rows. | | Under the search come the themes and sub-themes with counts, **Visualizations** (**Hierarchy Graph** and **Co-occurrence Network / Table**), and a code frequencies chart with **Segment**, **Display** (**Raw** or **Percentage**), **Put segments on y-axis**, bar or pie, and downloads.

Groups to compare tagging dialog on a completed project — Fix a mis-tagged file here instead of creating the project again.

| Section | What it holds | | -------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Themes** and **Questions** | One block per theme or question: **Summary across all files**, **Codes** (opens on its codes table; quotes grouped by code), **Comparison across segments**, and **Summary of individual files**. **Download: Word, Excel, charts**, copy, the instructions, and delete sit behind the block's ⋯; the grip reorders. Quiet **Add Theme** and **Add Question** buttons end each list. A theme with sub-themes opens on **Theme-level analysis**, then one card per sub-theme ("Sub-theme 1.1: …"). **Comparison across segments** says which segment it left out and why ("Gender: not compared. Only one value (Female) is tagged. Tag the other documents in Edit groups to compare."). | | Summary-first cells | A per-document cell opens on its summary, with **Show full analysis** (or **Show full text**) for the rest. Under the summary, its first quote shows as plain text, cut after about 280 characters, and **Show quotes (N)** opens the full quote list (**Hide quotes** folds it). **Expand all cells** in the Contents rail opens them all. | | "Not discussed" cells | A cell that only says the document does not discuss the theme ("Not discussed", "Not mentioned", "N/A") counts as empty and sorts last; the section folds those documents into one line, "Not discussed in N documents: …". | | **Summary of each document** | One summary per file, collapsed. | | **Word Frequency Visualization** | Under "Charts and tables". One word cloud for **All documents**, **Documents in a segment**, or **Documents I pick**; **Compare by** sets clouds side by side per segment value or for **Custom groups**, with the words that set each group apart; **Hide common words**; more options choose the text (full text or coded excerpts only), a theme, words or phrases, minimum count, maximum words, excluded words (saved for everyone on the project), and the word direction. Downloads as PNG, SVG, or a CSV of word counts. | | **Groups to compare** | The segment value of every file in a table, with **Edit groups**, filters, and a download; shown when segments exist. Focus-group participants get their own rows ("Grace · FGD\_01.docx"). |

A theme expanded: its definition, Theme-level analysis, and a card for each sub-theme — Open Theme-level analysis for the summary across all files, or a sub-theme for its codes and quotes; each quote links back to its file.

Role column block under Participant attributes: Values, Overview, a chart counting participants, and the Table of codes — Choice columns get their own block, so a single question such as the participant's role can be counted and charted per row.

| # | Part | What it shows | | - | ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | Column block | The column name and its description. | | 2 | **Values** | Opens with the block and holds the three views below. | | 3 | **Overview** | One row per unit with the value found, plus the documents it came from (left out when each row is a document). Click a value for its quotes. | | 4 | **Visualizations** | A bar or pie chart of how many units have each value, with **Segment**, **Display**, and **Put segments on y-axis**. Shown when two or more values were found. | | 5 | **Table** | Every value with its count of units; click one to read its quotes. | Quotes offer **Edit quote**, **Move this quote to another code**, and **Delete quote**; the pencil next to "Context from AI:" (**Did AI make a mistake?**) explains how to correct a coding; a code's **Edit code definition** can **Update definition only** or **Update definition and re-analyze**.

Edit the codebook and Calculate AI accuracy

* **Edit** in the Codebook header opens an editor with every theme, sub-theme, and code. Drag the grip to reorder, edit names and descriptions, **Move to another theme**, and on each code **Rename code**, **Merge with other codes**, **Split into multiple codes**, or **Move to another theme, sub-theme or question**. Expand a code to drag a quote onto another code. **Undo**, **Reset to Original**, **Download CSV**, and **Save Changes** sit in the toolbar. * **Calculate AI accuracy** (in the Codebook's ⋯) opens **Calculate AI accuracy**. A datapoint is one code and one document. Pick a sample, confirm or reject each coding decision, then click **Complete review**. The page reports **Overall agreement** (Cohen's kappa), **Right when it applied a code**, and **Right when it left a code out**, with a breakdown by code. **Copy summary for a report** and **Download CSV of every check** export the result, and **Fix AI mistakes** lists the datapoints you marked wrong, to edit the definition or re-code. Old reviews can be deleted.

Codebook editor with themes, sub-themes and codes — Changes apply only after Save Changes.

Calculate AI accuracy dialog with sample size choices — A larger sample gives a steadier kappa but takes longer to review.

**Share** **Share** in the action row opens **Share project**, in three parts.

Share project dialog with Invite by email, Links and People with access — Paid Evidano users you invite get edit access; anyone else gets a view-only link by email.

| # | Control | What it does | Plan | | - | ------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---- | | 1 | Email box (**Invite by email**) | One or more addresses. A hint under it says what each person will get ("Will get: Edit", or "Will get: View only (no paid plan)"). | | | 2 | **Add a message** | A custom message for the invitation. | | | 3 | **Invite** | Sends the invitation: a paid user becomes a collaborator who can edit; anyone else gets a view-only link by email. | Pro. | | 4 | **Links** | When offered, **Create link** makes a view-only link that anyone with it can open, with **Expires** (**Never**, **After 7 days**, **After 30 days**, **After 90 days**). Each link has **Copy** and **Revoke**. | Pro. | | 5 | **People with access** | Collaborators ("Can edit") with **Resend invite** and **Remove**; guests ("View only"). | | {% hint style="warning" %} **Public link:** Anyone with a view-only link can view and download the results, without signing in. Set **Expires**, and **Revoke** a link you no longer need. {% endhint %}

The invitation's custom message box, opened with Add a message — The message goes into the invitation email.

## Screen 9: The project list

Project list with Manage files and New document project, search, ownership filters, status and analysis chips — The Analysis column shows which outputs each project has; the status updates on its own.

The page is headed "Document analysis", with **Back to Dashboard**, **Manage files**, and **New document project** at the top. With no projects yet, it shows "No document projects yet" and **See a sample analysis** instead of the table. | # | Control | What it does | | - | ---------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1 | Search box | Finds a project by ID or name. | | 2 | **Search in:** | **Name or ID**, **Project name**, **Project ID**, or **Files used in projects**. To find a project by a file it used, choose **Files used in projects**. | | 3 | **All** / **My Projects** / **Shared with Me** | Filter by ownership. | | 4 | **New document project** | Starts a project. **Refresh** beside the **Status** heading reloads the statuses. | | 5 | **Status** | **Completed**, **In progress** (with a live percentage you can click for stage details; below 30% the row opens them too; from 30%, the details offer **Open partial results**, and a button to open the project once it is done), **Error** with **Retry**, or **Not started** ("This project never started analysing. Delete it and create it again."). | | 6 | **Analysis** | Chips for the outputs, for example **Summary of each document**, **Themes**, **Research questions**, **Compare across segments**. | | | **Date Created** | How long ago the project was created; hover for the date. | | | **Documents** | A count that opens the file list. | | | **Role** | Owner, Admin, Collaborator, or Viewer. | | | **Action** | Open and share (Pro, completed projects); **More actions** (⋯) holds **Delete project** (after a confirmation that names the project) or, for a project shared with you, **Leave project**. Open is greyed out below 30% ("Opens when analysis reaches 30%"). | ## Next steps * [Analyze Interview Transcripts](/ailyze-docs/common-research-workflows/analyze-interview-transcripts.md) — the walkthrough these controls belong to. * [Design Your Analysis Table](/ailyze-docs/analyze-data/design-your-analysis-table.md) — how to choose the unit of analysis and column types. * [How to Read Your Results](/ailyze-docs/analyze-data/how-to-read-your-results.md) — what each section and chart means.