Research a business account across sources
Produce a dated evidence table and a draft brief without initiating outreach.
Before the first run
- A supplied business identity and research question
- An approved source scope and requested fields
- Rules for identity matching and unknown values
Define what the agent can do.
Research and draft preparation only. Discovery of a contact does not authorize messaging, calling, or bulk collection.
The implementation sequence
- 01
Resolve the organization
Confirm domain, location, and relevant subsidiary or parent context. Avoid merging facts from businesses that share a name.
- 02
Gather scoped evidence
Visit permitted sources and record only the fields requested. Prefer the organization’s own authoritative pages for claims about its offerings or operations.
- 03
Retain provenance
Attach the direct URL and retrieval date to each factual field. Mark uncertain, conflicting, and absent evidence explicitly.
- 04
Prepare the brief
Separate observed facts from editorial inference. A plausible business hypothesis can be useful if labeled, but should not appear as an established fact.
- 05
Review the destination draft
If a CRM update is requested, show the existing and proposed values with sources. Keep all communication drafts unsent pending separate authorization.
Stop for human approval
Review factual claims and any destination record update. Outreach requires approval of the exact recipient and message.
Approval design principles →Implementation options
Browser Use
An open-source browser agent library with managed agents and cloud browser sessions.
Verified Sep 27, 2026Agent & orchestrationStagehand
A browser SDK that mixes natural-language actions with deterministic browser code.
Verified Sep 27, 2026Control frameworkPlaywright
Deterministic browser automation and testing across Chromium, Firefox, and WebKit.
Verified Sep 27, 2026Browser infrastructureBrowserbase
Managed browser sessions with observability, proxies, and computer-use integrations.
Verified Sep 27, 2026Failure & recovery
- Wrong subsidiary
- Outdated source treated as current
- Unsupported inferred facts
- Prompt injection in researched pages
What drives cost
- Sources per organization
- Identity and evidence conflicts
- Review and CRM handoff
Related use cases
Sources & verification
Reviewed Sep 27, 2026. Architecture recommendations are editorial analysis; linked vendor documentation supports the underlying capability and safety facts.
- developers.openai.com/api/docs/guides/tools-computer-use
- cheatsheetseries.owasp.org/cheatsheets/LLM_Prompt_Injection_Prevention_Cheat_Sheet.html
- docs.stagehand.dev/v4/first-steps/introduction