FAQ

How long does a research run take?

Research typically takes two to five minutes per client, depending on how much public information is available and which LLM tier is configured. Large batches (50+ clients) queue and work through progressively, and you can monitor per-client progress from the run detail page.

Can I research the same client twice?

Yes. Including a client in a new run creates a fresh client research record alongside any existing ones. Both are preserved and accessible from the client's research history. Re-running is useful when significant time has passed or after a company signal (such as a funding round) suggests the prospect's situation has changed.

What sources does the agent use during research?

The agent uses public web sources: company websites, news articles, press releases, LinkedIn, job boards, and similar. It does not access private data, CRM data, or paid data feeds unless you have configured an integration for that.

Why did a client score low when I expected it to score high?

Low scores usually mean insufficient public evidence for the methodology criteria, not that the company is a bad fit. Check the score breakdown on the client research record. It shows which topics had weak evidence. You may want to try re-running after a few months if the company's public profile is expected to grow.

Can I edit a completed research record?

You cannot edit the automatically extracted findings or the score. They are locked to ensure comparability. You can flag individual findings as incorrect (which adjusts their weight in scoring) and you can add notes. The outreach draft is fully editable.

What happens if an automation fires for the wrong client?

If an automation fires incorrectly, reject the task in the inbox. The rejection is logged. If the automation is consistently misfiring, review its condition fields on the automation's edit dialog, or pause it while you refine them.

Who can configure automations?

Editor, admin, and owner roles can create, edit, and enable/pause automations. Admin and owner are additionally required to delete one. Any of those roles can act on tasks in the inbox (approve, reject, snooze).

Is there a limit on how many clients I can add?

There is no hard limit on clients. Research run batches of up to several hundred clients are supported. Very large batches may take several hours.

How is the company news signal source different from running a new research run?

Company news signals are continuous, arriving in near-real time (polling every 15 minutes) and representing specific events. A research run is a comprehensive point-in-time analysis using many sources and producing a full report and score. The two work together: a signal can trigger a re-research run via an automation.

Can I use my own LLM API keys?

Yes. Go to Settings > LLM providers to add your own OpenAI, Anthropic, Ollama, or other provider credentials. You can route specific pipeline roles to specific providers via LLM routing.