Research operations

Research operations covers the standards and infrastructure that sit around individual studies: how projects are set up, how findings are labeled so they can be compared, how participant records are managed, and how compliance requirements are met without per-project configuration. This page covers building those standards into the objects people use — project templates, workspace tag boards and field groups, and workspace-level retention and redaction settings, so they apply by default rather than depending on each person following a written process.

---

​

Project templates

Most inconsistency originates at project creation. A project template carries context, automation defaults, linked tag boards, and field groups into every project made from it, so someone who has never set a project up gets your setup by default. That is what makes a study repeatable by non-researchers. Build one per study type you repeat — usability test, discovery interviews, win-loss — rather than one generic template nobody fits; the fastest way is to convert a project you were happy with. Doc templates do the same for the write-up, so a PM’s report is structured like a researcher’s.

---

​

Shared tags and fields

Workspace tags and workspace fields are the vocabulary everyone shares. Tag boards are live — change a tag and every linked project updates. Fields defined once in settings keep “segment” meaning the same thing wherever it’s filled in. A shared taxonomy is what makes a question larger than one study answerable at all. Put a label at workspace level if you’d ever use it across two studies: persona, product area, research method, segment. This round’s task names stay in the project. Keep the set small — 200 tags nobody applies consistently are worse than 20 that everyone does. Use single-select fields rather than free text where values need to be comparable; free text drifts and can’t be filtered.

---

​

Participant database

Contacts is one workspace-wide record of the people you talk to, linked to the data and docs they appear in. Get the fields right early — role, company, region, plan — because that’s what everything downstream filters on. Contact automation creates and updates contacts from project transcripts, so records don’t depend on anyone remembering. Update-only enrichment keeps the database from filling with half-identified speakers. Segments turn field filters into reusable groups that maintain themselves, so a team can narrow to one customer group without asking you who to talk to. Connecting your CRM turns volume into weighted priority — see Voice of customer.

---

​

Workspace compliance settings

Set at workspace level, each of these covers every project made afterward, including ones run by people who never read your policy. Compliance becomes a property of the workspace, not a checklist per project.

• Redaction. Turn on automatic blur and redact, tune the instruction prompt, and maintain the always-redact and never-redact term lists. Suggest mode adds a human review step.
• Access. Security settings control web links, who can invite users, who can remove redactions, session age, and where AI processing happens.
• Retention. Workspace data retention deletes video and audio after a period you choose, keeping the transcripts and highlights made from them.

Redaction and retention aren’t retroactive — they apply only to data arriving after you turn them on, so do it before the import you’d regret. Send procurement and legal to Security and trust.

---

​

Self-service answers

Chat gives cited answers to plain-language questions at whatever scope it’s opened, and Search covers the “I know this exists” case. The citation is what gets you out of the loop: nobody needs you to vouch for an answer they can open the quote behind. Chat answers questions in Slack and Teams, and the MCP server opens the workspace to Claude and Cursor. Workspace context docs inject your glossary, ICP, and reporting standards into every Chat and Agent conversation, so answers come back in your language. Three to five focused docs beat one long one.

---

​

Automate maintenance with agents

Agents run on a schedule or an event, which suits the upkeep that otherwise grows with the workspace: stale projects, missing metadata, duplicate docs, contacts to reconcile against your CRM. Workspace hygiene has prompts for each. Keep agents on flagging, not deleting. An agent that reports a problem to a person is safe to leave running; one that merges or archives on its own isn’t.

---

​

Usage reporting

Two capabilities report on how research is being used.

• Doc metrics report, per doc, the unique accounts reached, where people found it, their role, and how far they read — evidence that decision-makers read the finding, not just that you published it.
• Workspace analytics charts projects, data, highlights, tags, docs, and contacts created month over month, so you can see whether the work is spreading beyond your team.

---

​

Where to go next