Your site says "three live agents built solo" but shows four links. Creator Scout is gone. UAT Sentinel is there but has no published case narrative. Leave it out of outreach until you've written one. The canonical external set is still Brand Pulse, Retail Velocity, Carrier IQ.
More important: name one per message, not three. The moment you list all of them, the read shifts from she already solved our specific problem to she has a collection of demos. The labs share a single thesis: when an agent acts, what does the human watching it need to see? That thesis is your positioning asset. One named lab proves it. Three named labs turn it into a catalog.
Pick the lab that matches the buyer's anxiety, not the one you're proudest of. Carrier IQ for regulated workflow buyers. Brand Pulse for AI-native and oversight-pattern buyers. Retail Velocity only when field-ops intelligence is the product surface.
Fix the 3-vs-4 count mismatch before any outreach to OpenAI or Ramp. Their people will view-source.
Which lab for which buyer:
- OpenAI, Suno → Brand Pulse. Live agent with hourly scoring and human oversight surface. Pair with Trust essay as opener.
- Headway, Gusto → Carrier IQ. Regulated-domain automation where operator verification is the design problem. Lead with Thermo Fisher or Red Cross; Carrier IQ is the AI method layer.
- Ramp → Carrier IQ or Retail Velocity. Pick whichever maps closer to Ramp's current product surface. One, not both.
- Capital One → Brand Pulse + Carrier IQ. Exception: name two here. Capital One's XD org is absorbing AI at scale. Two shipped systems prove you've built the practice, not theorized it.
Do not use externally:
- UAT Sentinel — on the site but outside the published evidence boundary
- Creator Scout — removed from site entirely
Before outreach: Resolve the "three agents" copy vs. four visible links. Technically literate buyers will notice.