Overview
CAQDAS (computer-assisted qualitative data analysis software) helps you organize transcripts, documents, audio, images and survey responses; code them; write memos; and query patterns such as which codes co-occur or how themes differ by participant group. The three most widely used packages are NVivo (Lumivero), ATLAS.ti and MAXQDA. Each is updated frequently, and all now include AI features for summarizing text and suggesting codes.
Most problems we see are structural: documents imported without case classifications, codebooks that grew into hundreds of overlapping codes, memos scattered in Word files, or team projects merged without a plan. We fix these by designing the project around your methodology — thematic analysis, framework analysis, grounded theory, IPA or content analysis — and your reporting needs.
We also help you decide whether and how to use AI-assisted features. They can speed up orientation and retrieval, but automatically generated codes and summaries are not analysis, and uploading data to a third-party AI service may not be permitted by your consent forms or ethics approval. We help you check settings and document any use.
Who it’s for and the problems it solves
This is for you if you are…
- Students who have been told to use NVivo, ATLAS.ti or MAXQDA and do not know where to start
- Researchers with a messy existing project that needs restructuring before analysis continues
- Research teams that need a shared codebook, team-coding workflow and merged projects
- Anyone deciding whether to adopt the software's AI features in a way their ethics committee will accept
Common problems we help with
- You imported transcripts but cannot compare codes by participant group because attributes are missing
- Your code list has become long, duplicated and unmanageable
- You do not know how to run matrix coding, crosstab or code co-occurrence queries
- Team members' project copies need merging without losing work
- You want to use AI coding but are unsure about data privacy and how to report it
What’s included — and what isn’t
Included
- Software choice advice based on your method, data types, platform (Windows/Mac/web), license and team needs
- Project setup: document import, folders or document groups, cases, classifications/attributes or document variables
- Codebook structure: hierarchy, definitions, color conventions and codebook export for appendices
- Memo, annotation and audit-trail workflows suited to your methodology
- Queries and retrieval: text search, word frequency, coding and matrix queries, code co-occurrence, crosstabs and framework matrices
- Visual tools: mind maps, concept maps, network views, MAXMaps and comparison charts
- Team projects: coding conventions, merging projects, and intercoder agreement features where your approach requires them
- AI feature guidance: what the AI assistants do, privacy settings, and how to document their use
- Export of coded extracts, codebooks and outputs into Word, Excel or report formats
Not included
- Coding your entire dataset so you can present the analysis as your own
- Using AI features on data where your consent or ethics approval does not permit third-party processing
- Selling, installing or troubleshooting software licenses; please use your institution's IT service or the vendor
What you receive
- Configured project file
- An NVivo, ATLAS.ti or MAXQDA project with your documents, cases, attributes and codebook structure set up and ready for coding.
- Codebook document
- An exported, formatted codebook with definitions and examples, suitable for your appendix or team use.
- Query guide
- Step-by-step instructions, with screenshots, for the specific queries and visualizations your analysis needs.
- AI use note
- A short record of which AI features were used, on what data, with which privacy settings, ready to adapt for your methods section.
- Training session
- A live, screen-shared walkthrough so you can run the project independently.
How it works
Method and data review
We learn your methodology, data types and outputs, and confirm which software and version you have.
Project design
We propose a structure for documents, cases, attributes, codes and memos, and agree it with you.
Build and test
We set up or restructure the project and test the key queries on a subset of your data.
Handover and training
You receive the project and guides, followed by a training session and follow-up questions.
Software and features
- NVivo (Lumivero)
- Cases and classifications, framework matrices, matrix coding and crosstab queries, and an AI Assistant for summaries and child-code suggestions. Lumivero has announced a cloud-based generation of NVivo, so check which edition your institution provides.
- ATLAS.ti
- Quotations, codes, document groups, network views and a code co-occurrence table; AI Coding, Intentional AI Coding, Conversational AI and AI summaries, with an AI privacy mode that disables OpenAI integration.
- MAXQDA
- Document variables, Code Matrix Browser, Code Relations Browser, MAXMaps and mixed-methods tools; AI Assist offers summaries, suggested subcodes and segment-level AI coding, activated separately.
- Intercoder agreement
- Built-in agreement tools in NVivo, ATLAS.ti and MAXQDA for codebook or content-analysis designs; not usually appropriate for reflexive thematic analysis.
- Framework analysis
- Case-by-theme matrices (NVivo framework matrices, MAXQDA summary grids) for applied and policy research.
- Exports
- Codebooks, coded segments, query results and visuals to Word, Excel, PDF and image formats for appendices and reporting.
Choosing between NVivo, ATLAS.ti and MAXQDA
All three handle the core tasks well, so the best choice is often the one your institution licenses and your supervisor knows. Differences matter at the edges: MAXQDA has strong mixed-methods and visual tools and works identically on Windows and Mac; ATLAS.ti's quotation-based model and network views suit grounded theory and conceptual mapping, and it offers a web version; NVivo's case classifications and framework matrices suit large, attribute-rich projects. Version names change often, so we check what you have before advising.
Using AI features without losing the analysis
Current versions of all three packages offer AI summarization and code suggestions, and some offer chat-style questioning of your documents. These features typically send text to an external AI provider. Vendors describe safeguards such as zero data retention or regional processing, but your consent forms and ethics approval are what determine whether you may use them.
Even when permitted, AI suggestions are a starting point. In reflexive or interpretive methods, relying on auto-generated codes can undermine the very engagement the method depends on. We help you decide where AI helps (for example, orientation or retrieval), where it does not, and how to report its use transparently.
- Check consent and ethics wording before enabling any cloud AI feature
- Use privacy or offline modes where available and document the setting
- Review every AI-suggested code; keep a record of what you accepted, changed or rejected
- Follow your institution's and target journal's AI disclosure policy
Audit trails that examiners can follow
Software makes rigor visible when you use it deliberately: dated memos for analytic decisions, a codebook with version notes, and saved queries that show how themes were checked across cases. We set these up from the start so the evidence is there when you write your methods section or face a viva.
What the work can look like
Illustrative example — fictional data
Illustrative project structure for a multi-site interview study
A fictional research team interviewed 30 staff across three hospitals and uses NVivo for framework analysis. The design below is invented to show how a project can be structured.
Cases and attributes
One case per participant with attributes for site (A, B, C), role (nurse, doctor, manager) and years in post (banded). This lets the team run matrix coding queries comparing themes by role and site.
Codebook
Twelve parent codes from the framework, each with a definition, inclusion and exclusion notes, plus an 'Emerging' parent for inductive codes reviewed weekly.
AI decision
The consent form did not mention third-party processing, so the team kept AI features switched off and recorded that decision in the project memo.
Illustrative example only; the study and structure are invented and do not describe a real client.
What you’ll need to provide
- The software and version you use, and whether it is a desktop, web or team license
- Anonymized data files and a participant attribute sheet (for example, age group, site, role)
- Your methodology and any existing codebook or project file
- Ethics approval and consent wording relevant to data sharing and AI processing
You don’t need everything ready to ask for a quote — tell us what you have and we’ll say what’s missing.
What affects price and turnaround
Price depends on
- Size and variety of data (text, audio, video, images, survey imports)
- New setup versus restructuring an existing project
- Number and complexity of queries and visualizations required
- Team projects that require merging and agreement testing
- Number of training sessions
Turnaround depends on
- Availability of anonymized files and an attribute sheet
- Software version compatibility and license access
- Whether an existing project needs cleaning before new work begins
Every project is quoted individually. How we price our work.
Questions people ask before starting
Which version of the software do you support?
We work with current and recent versions of NVivo, ATLAS.ti and MAXQDA on Windows and Mac, and with ATLAS.ti Web. Tell us your exact version, because menus and AI features differ between releases.
Can the software do my thematic analysis for me?
No. It organizes and retrieves; you interpret. AI suggestions can help you orient, but in interpretive methods the coding and theme development should be yours, and many examiners will ask how you used any AI output.
Is it safe to use the AI features on interview data?
It depends on your consent forms, ethics approval and institutional policy. Vendors publish privacy information, but they cannot grant you permission. We help you check and document your decision.
Can you merge my team's projects?
Yes. We plan coding conventions first, then merge copies and resolve duplicated codes. If your design calls for intercoder agreement, we set up and interpret the software's agreement tools.
Do you need my raw data?
Only anonymized files, and only the parts needed. For setup and training we can often work with a small subset.