An AI sales agent emailed a prospect named Chris to ask if his fish tank was still thriving, and it booked a meeting. No one on the sales team had mentioned a fish tank. The agent found it in a year-old meeting transcript stored in the company's own CRM, the system where sales teams log contacts and conversations.
Why this stood out
Aurasell, the company behind the agent, showed it at a SaaStr AI Day session. The pitch was a problem with most AI prospecting tools. They run almost entirely on third-party data, bought from the same vendors. So every team ends up sending similar emails to the same people.
Aurasell's agent also reads first-party data: the company's own CRM history, past meeting transcripts and internal conversations. Chris had met with Aurasell more than a year earlier, and the timing never worked out. The intro small talk in that transcript was about his large home fish tank. No data vendor has that detail.
A hook is not enough
Steve, Aurasell's CMO, was clear that a personal detail alone does not get a meeting. Saying you saw someone likes golf gets a shrug. The hook has to lead into a problem the prospect actually has. Age matters too. A signal from six months ago is usually weaker than one from last week.
How the agent works
- There is no fixed sequence. The agent builds a plan for each contact and changes it as new signals arrive.
- If a contact opens three emails and clicks links but never replies, the next step becomes a phone call.
- It tracks job changes and reaches out at the new company, and also finds the right person at the old one.
- If a profile looks stale, it searches news and press coverage about the person.
Aurasell also uses the agent outside sales. A recruiting agent messaged candidates, answered questions in their time zones and booked 20 meetings in three days while its owner was on vacation. Time zones are a known weak spot, since agents can default to UTC and email someone at 3 a.m.
Where humans still win
Aurasell's human SDRs, the reps who make first contact with prospects, book 14 to 16 meetings a week. About 60% come from cold calls. Most of the rest come from LinkedIn. Email produces the occasional meeting, because their buyers are sales leaders flooded with it. Steve tests candidates by having them write a prospecting email by hand, without AI.
The takeaway from the session was that most outbound needs both first-party and third-party signals, chosen to fit the goal of the campaign. Amelia plans to benchmark the agent against the ones SaaStr already runs. Results are not yet known.
The takeawayOld meeting notes and CRM history can give sales emails a personal, relevant opener that bought contact data cannot, as long as the hook leads to a real problem.