---
name: sell-with-fullsy
description: Research a prospect with Fullsy and turn the captured evidence into a calibrated cold-to-sold account plan. Use for prospect research, account planning, outbound angles, discovery preparation, fit assessment, stakeholder hypotheses, or evidence-backed sales messaging. Tie observed prospect claims to Fullsy record ids and clearly label hypotheses.
---

# Sell with Fullsy

Capture a prospect's public web presence, assess fit, and build a cold-to-sold plan where the factual account research has receipts.

## Preconditions

- Fullsy Agent Access must be enabled in the Chrome profile being automated.
- Codex must have a browser integration that can evaluate JavaScript in that same profile.
- Use one trusted approved control page for `window.fullsy` calls.
- If browser access is missing, do not claim that Fullsy captured the prospect.

## Fullsy bridge

```js
await window.fullsy.captureUrl(url, { caseName })
await window.fullsy.getRecord(capId)
```

Poll to completion. Agent captures are limited to one every 15 seconds, so capture sequentially.

## Inputs

Establish, from the user or context:

- what the user sells
- ideal customer / disqualifiers
- prospect company or URL
- desired motion: cold email, call, LinkedIn, discovery, expansion, partnership, etc.

Do not block on minor missing context; make bounded assumptions and label them.

## Workflow

1. Create a case name such as `Prospect — Acme — 2026-08-17`.
2. Read the prospect's primary site before deciding what pages matter.
3. Capture the homepage and only the additional pages that contribute to the sales thesis, for example:
   - relevant product/service page
   - pricing
   - customer/case studies
   - about/company
   - careers when hiring signals matter
   - contact/team only when stakeholder information is relevant
4. Poll every capture until it has a Fullsy record id.
5. Build two columns mentally throughout the analysis:
   - **Observed:** directly supported by a captured page.
   - **Hypothesis:** a reasoned sales inference that still needs discovery.
6. Assess fit against the seller's ICP. Give disqualifiers real weight; do not force a prospect into "good fit" because the user asked for outreach.
7. Identify the strongest trigger or tension visible in the captured evidence.
8. Translate that signal into a sales angle connecting:
   - what the prospect appears to care about
   - what may be costly/difficult now
   - what the seller can credibly improve
9. Identify stakeholder *roles* likely to own the problem. Name specific people only when supported by a reliable current source.
10. Build a cold-to-sold path:
    - opening observation
    - problem hypothesis
    - value bridge
    - proof to use
    - low-friction CTA
    - discovery questions
    - likely objections
    - next-step progression
11. Tie every prospect-specific factual statement in the plan back to one or more Fullsy record ids.

## Guardrails

- Do not present inference as fact.
- Do not claim private information, internal metrics, budget, intent, tech stack, or organizational problems unless evidence supports it.
- Do not treat generic site copy as proof of buying intent.
- A Fullsy record proves what was captured, not that the prospect's public claim is objectively true.
- Prefer a smaller number of sharp, defensible observations over a dossier full of trivia.

## Final output

### Account verdict
Fit: `strong`, `plausible`, `weak`, or `disqualify`, with reasons.

### Evidence that matters
| Observation | Why it matters for the sale | Fullsy record |
|---|---|---:|

### Best angle
One primary angle and, if useful, one fallback angle.

### Who to reach
List likely stakeholder roles and why each matters. Distinguish confirmed names from role hypotheses.

### Cold-to-sold plan
Opening → hypothesis → value → proof → CTA → discovery → objections → next step.

### Claims to avoid
List tempting but unsupported claims the rep should not use.