A lead list can look complete and still be hard to use. Company names do not match the CRM, job titles are stale, the right person is missing, and an email address gives no clue whether it is safe to send. Someone ends up checking rows across LinkedIn, company sites, databases, and spreadsheets before the real sales work can begin.

Strawberry is a browser with built-in AI agents. Give your companion an existing lead or account list and it can work across the web, logged-in sites, your CRM, files, open tabs, and connected enrichment tools to fill the gaps. Just as importantly, it can preserve where useful data came from and keep uncertainty visible when it affects what you do next.

Start with the list you already have

Lead enrichment starts after you know which records you want to work with. The input might be a CSV, spreadsheet, CRM view, account list, event attendees, inbound leads, or a shortlist from earlier research.

Begin with the decision the list should support. Preparing outreach, assigning territories, segmenting accounts, cleaning the CRM, and mapping stakeholders may require different fields and levels of verification.

If you do not have the companies or prospects yet, start with Find New Customers instead. It helps define who fits and builds the initial list before enrichment begins.

Enrich the fields that will actually help

More columns do not automatically make a better list. Choose the fields that improve the next decision, then let your companion find and check those.

What you needUseful enrichment might include
Better company contextWebsite, industry, size, geography, ownership, technology, or another relevant segment
The right peopleCurrent role, seniority, function, location, profile, or relationship to the decision
A reliable way to reach themWork email, phone number, channel, source, and verification status
A clean working listResolved identities, normalized fields, duplicates, stale records, and conflicting values
Relationship awarenessExisting customer status, account owner, open deal, prior contact, or do-not-contact record

People research is optional. A territory or market-analysis list may only need company data; an outreach list may need current decision-makers and contact details.

Look beyond one database

A shared data provider is useful, but no single source has every field fresh and correct. The company site may have the current leadership team, LinkedIn the current role, the CRM the existing relationship, and a specialist enrichment service the best contact data.

Because Strawberry is a browser, your companion can open the real sources, follow links, work in logged-in tools, and investigate ambiguous records rather than returning the first matching profile. It can also connect enrichment and verification tools when they are the best source for a field.

That makes enrichment less like filling blanks and more like resolving the right company or person before adding data to the record.

Balance coverage with trust

A broad pass can fill common company fields quickly. A deeper pass can resolve harder identities, cross-check current roles, and verify the contact details that matter most. For a large or unfamiliar list, review a small, varied sample before spending more time and credits on the rest.

Missing data does not need to be guessed, and useful but unverified data does not always need to be deleted. Keep its source and status clear. An address may be useful for further checking without being safe to treat as send-ready.

  • Preserve the original record so changes are easy to understand.
  • Surface ambiguous matches instead of silently merging them.
  • Keep verification visible when it changes how a field can be used.
  • Review conflicting values before overwriting an external record.

Prepare the list for safe, useful outreach

A few warm messages and a scaled cold-email campaign do not need the same setup. At higher volume, connect a reliable email verification service before sending. It can reduce bounces and help protect the sending account. For ordinary one-to-one outreach, keep the process proportional to the risk.

Once the audience and data are ready, your companion can carry the approved list into personalized outreach without losing the sources, relationship context, or verification decisions behind it. Sending remains a separate approved step.

Make the next list faster

Tell your companion which fields mattered, which sources you trust, what counted as a match, and when you wanted a human review. Share the reviewed list when teammates need the result, or save and share the accepted method as a team skill when they should reuse the process. If they need the same account context across enrichment, outreach, CRM, and other sales work, share the full companion.

When the workflow is reliable, a Routine can check a new list, a recurring import, or agreed CRM records on the right trigger or cadence. A good first version prepares enriched records and stops for ambiguous identities, risky contact data, or external changes that still need review.

Open the Enrich a Lead List skill with one list and the next decision it needs to support. Your companion can prepare a useful first sample, learn from your feedback, and turn the rest into data you can actually use.

Official Strawberry skill

Enrich a Lead List

Turn an existing list into a more complete and trustworthy working set. Enrich the fields the user needs rather than assuming every list requires people or contact details.

Workflow

1. Understand the list and the result

Inspect the list, its current fields, and the intended destination. Use relevant CRM context, connected apps, files, tabs, memory, and prior work before asking the user to repeat information already available.

Understand what each row represents and what would make it useful: company facts, segmentation, current roles, decision-makers, contact information, source links, cleanup, or a combination. Research people only when the result needs them.

2. Agree on fields, depth, and a first sample

Agree on the fields, matching criteria, useful sources, destination, and whether the user wants broader coverage or deeper verification. If decision-makers are needed, use the user's process or strong customer examples rather than assuming one universal title.

For someone new to the workflow, or for a large, unfamiliar, or expensive job, propose a short plan and representative sample. Explain breadth, depth, and credit tradeoffs comparatively; give specific estimates only when reliable. Keep this light for a trusted repeat workflow.

3. Resolve and check the existing records

Preserve the supplied rows, identify duplicates, and surface ambiguous matches instead of silently merging or dropping records. When a CRM or another source of truth is relevant, check customer status, ownership, live deals, prior contact, and do-not-contact records. Flag conflicts or existing relationships rather than presenting them as new leads.

Use enough evidence to resolve each company or person. Never guess a profile URL, employment relationship, email address, phone number, or other missing field.

4. Enrich and calibrate

Use the best available source for each requested field across enrichment services, CRM records, LinkedIn, company sites, connected apps, and other useful web sources. Preserve links and dates where freshness or provenance affects the result.

Do not present an unverified field as verified. A potentially useful email address or phone number may remain when its source is known, but label its status and risk clearly. Never treat risky or unverified contact information as send-ready.

If the list will support scaled cold email and no reliable email verification service is available, suggest connecting one before sending. Explain briefly that verification can reduce bounces and help protect the sending account. Do not make this a prerequisite for ordinary one-to-one outreach.

When calibration would materially improve the work, present a sample and let the user correct the fields, matching, sources, and verification bar before expanding. On LinkedIn, work sequentially at a human pace and stop on warnings, challenges, rate limits, or unexpected states.

5. Expand and prepare updates

Apply the accepted approach without changing field meanings or lowering the verification bar. If a field has poor coverage, explain why and offer a useful choice: try another source, accept clearly labeled unverified data, narrow the field, or leave it missing.

Prepare changes for review before writing to an external destination. Do not overwrite existing or conflicting values without showing what would change and receiving approval.

6. Deliver and make the next run easier

Return the original identifier, resolved identity when needed, requested fields, and any source or verification notes that affect trust or use. Prefer a compact table for a sample and the agreed destination for larger results. Keep missing data missing rather than guessing, and surface stale, conflicting, ambiguous, or unverified values when they could change what the user does.

Check duplicates, identity mismatches, stale roles, formats, broken links, and provenance before delivery. Report coverage or update counts when they help the user judge the result.

Use strawberry/sales/find-new-customers when the user still needs prospects, strawberry/sales/send-personalized-outreach when they want to contact approved leads, and strawberry/sales/keep-crm-updated for approved CRM changes.

After feedback, offer to save the accepted fields, sources, matching rules, verification bar, and destination as a custom skill. If teammates use the same list or data process, offer to share the reviewed result or accepted enrichment method with them; a shared skill can preserve the method without forcing everyone to recreate it. Keep that separate from writing records into external systems. When the team needs the same context and several Sales workflows, offer to share the full companion. When the workflow is trusted and has a real trigger or cadence, suggest a Routine that checks the agreed source, prepares enriched records, and stops for ambiguous matches or reviewable external updates.