Results | Haines GTM & Revenue Systems

AI implementation at a seed-stage home services company

Result: Post-call work, 20 minutes to under 5. Two more appointments a day on each rep's calendar.
Situation: A concierge business where responsiveness is the product, and reps spent the time after every call typing instead of dialing.
Blocker: Nobody had written down what a good call log contained. I graded a hundred real calls to write the spec, then let the model do the typing.

Deal desk at a public residential solar company

Result: Deal rejections, 25% to 5%. Cancellations down 11%, and the gate still processes tens of thousands of deals a quarter, ten years on.
Situation: Reps sent incomplete deal packages, ops kicked them back, and customers cancelled in the gap.
Blocker: Nothing in the workflow depended on the paperwork being right. Commission paid on signature, so the package was an afterthought.

Proposal generation at a private roofing and solar manufacturer

Result: Proposal turnaround, 48 hours to 20 minutes. Pitch rate up 20% on a $48M partner channel.
Situation: Two hundred roofing partners sold the product, and every proposal round-tripped through an internal team.
Blocker: The pricing and product data already existed. The handoff between the partner's CRM and ours did not.

Appointment routing at a public residential solar company

Result: Appointment-to-sit conversion, up 50% across 800 retail reps.
Situation: Thousands of retail appointments a month, and conversion looked like a rep skill problem.
Blocker: Appointments went out with no allocation logic, so the wrong rep got the wrong appointment at the wrong time.

Shared inbox at a seed-stage home services company

Result: SLA compliance, none to the low 90s. The urgent tickets surface the day they arrive.
Situation: A shared inbox took 3,000 tickets a month with no categories and no SLAs, so the urgent ones waited behind the rest.
Blocker: Most tickets were sortable by rules and the rest needed judgment, and nobody had separated the two. Rules for the obvious ones, a model for the ambiguous ones, a priority score on every ticket.

Contractor referrals at a seed-stage home services company

Result: Speed-to-referral, up 50%. Reps match and book a contractor while the homeowner is still on the call.
Situation: Matching a homeowner to a vetted contractor was the product, and it happened after the call, by hand, across several people.
Blocker: The matching data was in the CRM the whole time. Nothing put it in front of the rep during the call, so I prototyped the widget in a day and handed the dev team a working version.

Twenty minutes of notes after every call.