AI is integrated across Marvin to help you with each step of your research workflow. There are 4 features and tools to help you analyze your data. Choose based on the kind of insight you're seeking and the stage of your research.
Ask AI
Deep Research
Analyze
Cross-project analysis
When should I use Ask AI?
Ask AI is a great way to surface specific insights across your repository. You can use it to quickly pinpoint information within a file, project, or across projects. For example, “How did company X discover our product?”
You can also use it as a starting point to develop a hypothesis. For example, you may ask, “Make a list of use cases for our product that participants have mentioned on calls”. Here you may apply a filter for types of users and within a fixed time period. The Ask AI results can help you begin looking in the right places. You can search for notes, quotes, files, and participants with Ask AI.
If you need to analyze files in depth, we recommend you use Deep Research or synthesise notes on the Analyze page.
Analyze page
The Analyze tab in each project is a place for you to synthesize your notes. After you’ve annotated files in your project, you can access all of them together on this tab. You can arrange these by question and file as well.
We recommend using the Analyze page when you’ve already highlighted the key information in each file. Filter notes by project, label, sentiment, etc. to group and locate relevant notes easily.
You’ll also find an Ask AI function on the Analyze page. It works only on the notes you select to help you summarize, surface insights, and spot trends.
Deep Research
Use Deep Research when you need to directly extract insights from raw data. Marvin can analyze files that you haven’t annotated and evaluate them in depth within certain frameworks. This includes thematic analysis, quantified analysis, arranging information in Q&A tables, etc. You’ll get detailed insights from each file you run through Deep Research.
Cross-project analysis
This tab is similar to the Analyze page, but it contains notes across all the projects in Marvin to which you have access. Use filters and Ask AI on this tab to find common themes and trends across projects.
Here's a comparison that may help you decide when to use each feature.
| Ask AI | Deep Research | Analyze |
Purpose | Ask questions to surface insights from data to which you have access | Generate in-depth, report-like analysis from raw data in a project | Granular, controlled analysis on notes you add to files |
Scope | Repository-wide (restricted to files you can access | Project-level only | Project or cross-project level |
Type of analysis | Fast answers, quotes, comparison tables, can expand interactively | Detailed, report-like structure; themes, patterns, hypothesis testing | Thematic, trend, emotional analysis, grouping, affinity mapping |
Speed vs. Depth | Prioritizes speed and provides a starting point for analysis. Not exhaustive. You can expand answers and ask follow-up questions | Prioritizes depth and rigor. Slower but more exhaustive analysis | Allows deep, customizable analysis with the most control |
Applicable filters | Medium level of control. You can filter by file tag, project, creation date, and research area | Limited control (mainly by file tags); less granular | Most granular: select notes, file tags, labels, questions, files, projects |
Interactivity | Highly interactive; you can ask follow-up questions, expand answers | Less interactive during analysis. You can edit reports that are generated | Interactive selection and synthesis of notes (manual as well as AI-assisted) |
Annotation required | No | No | Yes (manual or auto-notes) |
Typical use case | Quick discovery, quotes, trend checks, exploratory research, find files | Proving/disproving hypothesis, deep dives, rigorous analysis on raw data | Detailed, controlled qualitative analysis, grouping, synthesising annotate data and creating reports with it |