Create your first research project

This guide walks you through everything you need to create and launch your first research project with confidence. You’ll learn how to set up a new project, choose the right configuration options, and organize your work from the start so it’s easy to manage and scale.

We've broken it up into the following sections:

Every research effort in HeyMarvin starts with a project. A well-named project with the right permissions in place from day one keeps your work organized, your data protected, and your team aligned on what’s in scope.

Start by creating a new project and giving it a clear, descriptive name that will be easy to identify later — both for you and for stakeholders searching the workspace months from now.

Set permissions so the right collaborators can edit or contribute while sensitive data stays protected. Permissions aren’t one-and-done — you can adjust them as the project matures and access needs shift.

A few minutes spent organizing your knowledge hub upfront pays off every time someone tries to find research later. Marvin gives you project groups, pinning, and filters so your library stays navigable as it scales — for you, your team, and the stakeholders who consume your work.

Group projects in a way that reflects how your team thinks — by year, research area, business unit, product line, or whatever structure fits. One way is to mirror the way your stakeholders naturally search for information.

Pin high-priority projects for instant access, and use filters to sort by ownership, title, or recency. Small workflow choices compound — especially as your hub grows.

The strongest research starts with a clear question. Defining your goals, hypotheses, and target audience at the project level gives your team a shared frame — and gives HeyMarvin’s AI the context it needs to surface stronger themes, summaries, and answers later on.

State what you’re trying to learn in plain language. Specific goals lead to specific answers.
​

Capture what you expect to find and what you’re not yet sure about. This makes it easier to test assumptions against real findings as the research unfolds.

Note who you’re researching and what success looks like. This context grounds Marvin Ask AI and Deep Research in your specific research questions, so the analysis stays relevant.

A clear discussion guide is the foundation of consistent, high-quality data. Whether you’re running live interviews, AI-moderated sessions, or usability tests, a well-built guide keeps every session on track and makes analysis dramatically faster downstream.

Add your questions, tasks, and objectives — and include moderator notes where needed. Research Templates keep methodology consistent across studies and across teams.

Let Marvin Ask AI suggest questions grounded in your project goals, or help you tighten a draft you’ve already started.

Tag questions and tasks separately so HeyMarvin can analyze responses with the right context. Clean structure now means cleaner synthesis later.

A consistent taxonomy is what turns a project into a system. Establishing file tags, labels, and research areas before you start uploading keeps insights findable, sortable, and reusable across your whole organization.

Standardize file tags, labels, and research areas across your workspace. Consistent naming means every team — research, design, product, marketing — can find what they need without translation.

Group files tags by file type, research area, theme, or whatever structure fits your team. Tag groups make application faster and navigation easier as your library grows.

Now bring your data into the workspace. Audio, video, transcripts, documents, surveys, support tickets — Marvin connects to where your customer knowledge already lives, so you don’t have to move it manually.

Add files in batches you can review and tag as you go. Smaller batches make for cleaner annotation.

Pull data directly from Zoom, Gong, Google Drive, Salesforce, Zendesk, Qualtrics, and more. Integrations ingest data continuously, so your workspace stays current without manual uploads.

Add notes, highlights, and tags while the context is fresh. Auto-notes can handle the first pass on transcripts so you can focus on what matters.

Automatic PII redaction and face blur protect participant privacy at the project or file level — no manual step required.

This is where research velocity goes up. Marvin’s AI is purpose-built for research, not retrofitted for it — every theme, summary, and answer traces back to your source material with point-level citations. Use it to surface patterns you’d otherwise miss and to give days of analysis back to your team in minutes.

Let Marvin generate notes and apply tags automatically as files come in. Stay in control of interpretation while skipping the manual coding.

Surface themes, sentiment, and patterns across thousands of data points. Deep Research generates detailed, cited reports in minutes — and every claim links back to source.

Ask questions across your entire workspace

Marvin Ask AI lets anyone on your team ask a question and get an evidence-backed answer in seconds — grounded in your customer data, not the open web.

Research that sits in a repository doesn’t drive decisions. The final step is packaging findings into Insight Reports and surfacing them where your stakeholders already work — Slack, Salesforce, Figma, and email digests.

Create a report with an abstract, goals, methodology, and key findings. Add video clips, quotes, and highlight reels to bring evidence to life.

Turn findings into AI-generated presentations for stakeholder readouts, or set up research digests so the right people get the right insights without asking.

Use Marvin MCP, the Slack integration, and the Salesforce integration to surface insights inside the tools your stakeholders already use. No more digging through folders or waiting for a readout.

Publish reports only when they’re ready for broader audiences. Keep drafts limited to contributors so stakeholders see polished findings, not works-in-progress.