AI in the sale. What owner-led companies are doing with it, and what it's worth. September 28, 2026 · 8 stories
This week

AI wins where the buyer is already leaning in

SaaStr's AI agents have closed millions, all from buyers who'd raised a hand. Plus a CPA firm, a freight broker, a harness maker and more.

Collage: an old brass service bell on a shop counter beside a paper order ticket, with small green squares rising from the bell.
Editorial illustration / AI-generated

Hey,

Issue two, and a wider spread of businesses this week: a software company, a CPA firm, a freight brokerage, a wire-harness manufacturer, a building-products distributor, an insurance agency and a law firm. Same rules as last week. Every item has a label, and a lot of this week's numbers come from the companies selling the tool, so you'll see the vendor label more often than I'd like.

One pattern showed up in almost all of them. The AI results were strongest where the buyer had already raised a hand, whether that was a quote request, a reorder or an inbound call. The one firm that put a bot in front of prospects in a stressful moment ended up taking it back out.

Software & events

SaaStr runs 21 AI agents. Every dollar they've closed came from a buyer already leaning in.

Collage: a vintage telephone switchboard with hundreds of patch cords, only a handful of plugs lit green.
Editorial illustration / AI-generated
442K

website chats handled by one inbound agent

614

booked meetings from those chats

$1M+

in closed sponsorship revenue from that agent

1.5

people on the go-to-market team, doing what 6+ used to

SaaStr, self-reported, September 2026

SaaStr, the B2B software community and events business, now has 21 AI agents in production, built on tools like Qualified, Agentforce and PandaDoc. Founder Jason Lemkin says a go-to-market team of about 1.5 people now does as much as 6+ used to. The Qualified inbound agent handled 442,000 website chats, which turned into 614 booked meetings and more than $1M in closed sponsorship revenue. A win-back campaign to about 1,000 sponsor leads who'd gone quiet got 72% open rates and 10%+ response rates.

Measured result SaaStr's own numbers from its own post. Worth remembering that SaaStr is deep in the AI business itself.

Source: SaaStr, Sep 14, 2026

My take

Read his line twice: “Every dollar our agents have closed came from a motion where the buyer was already leaning in.” 442,000 chats for 614 meetings probably also tells you most of those chats were never going to buy. The AI seems very good at catching a buyer who's ready, and by his own account it hasn't talked a hesitant one into anything yet. So discovery and the business case are still a human job, and I think they're worth doing well.

Research

Half of B2B software buyers now start their search in an AI chatbot instead of Google

G2 surveyed 1,076 B2B software buyers in March. 51% now start their buying process with an AI chatbot, 71% use chatbots for research (up from 60%), and 69% ended up choosing a different vendor than they'd first planned. Separately, BrightEdge reported on September 24 that ChatGPT referral traffic more than doubled between January and August, though AI is still a small share of total referral traffic.

B2B software buyers and AI chatbots 0%25%50%75%100%Start their buying process with a chatbotStart their buying process with a chatbot: 51%51%Use chatbots for research (60% a year earlier)Use chatbots for research (60% a year earlier): 71%71%Chose a different vendor than they first plannedChose a different vendor than they first planned: 69%69%
B2B software buyers and AI chatbots
Start their buying process with a chatbot51%
Use chatbots for research (60% a year earlier)71%
Chose a different vendor than they first planned69%
G2, The Answer Economy, 1,076 B2B software buyers, March 2026

Directional A vendor survey (G2 runs a software review site) of software buyers specifically. From March.

Source: Demand Gen Report, on G2's survey, May 4, 2026 · BrightEdge on AI referrals

My take

If your buyers behave anything like software buyers, the first round of research might now happen without you, inside someone else's chatbot. Ask a couple of AI assistants what they'd recommend to a company like your best customer, and see whether you come up and whether they describe you correctly. If they don't, that's a website and content problem, and it's fixable.

Accounting

A six-person CPA firm lifted its conversion rate about 25% with a smarter intake form

Collage: a vintage typewriter with an index-card intake form, and a brass sorting tray letting only a few green cards through.
Editorial illustration / AI-generated

Agate CPA, with offices in Atlanta and Rogers, Arkansas, used Lovable (an AI app builder) to build a prospect intake form and Claude to get it onto the firm's website. The form screens prospects before anyone gets on a call: do they own a business, what industry, how many people, what revenue, startup or established, and why they're leaving their current CPA.

+25%

Conversion rate after the new intake form

Firm-reported, via Journal of Accountancy

The firm says its conversion rate rose about 25%, and its people now spend more time with prospects who are a potential good fit.

Measured result Trade press, figure reported by the firm, with no baseline or time period given. The article credits the redesigned questions as much as the AI.

Source: Journal of Accountancy, Aug 1, 2026

My take

I think the questions did most of the work here. They're the firm's definition of a good-fit client, written down. Most owners carry that definition around in their head, and that's exactly why they end up qualifying every lead themselves.

Freight brokerage

A freight broker cut quote replies from 35 minutes to 12 and won 18% more quotes

Collage: a vintage freight truck beside a stopwatch and a stack of shipping manifests, one marked with a green check.
Editorial illustration / AI-generated

ShipNova, a freight brokerage in Denton, Texas, uses Vooma's AI to handle incoming quote requests and build loads.

35 min12min

Time to reply to a quote request

Vendor-reported

Its quote win rate went up 18%, and building a load dropped from 20 minutes to 3.

Vendor-reported Vooma's case study, written about 18 months after the brokerage launched. No publication date is shown and it likely predates 2026. Company size isn't stated.

Source: Vooma, undated

My take

Most owners know their close rate. Very few know how long their team takes to answer a request that comes in, and it's often the cheaper number to move. I'd measure it for two weeks before buying anything that promises to shrink it.

Manufacturing

A wire-harness maker's president says quoting software saved him about 18 hours on one quote

Collage: coils of wire harness beside a technical drawing and a vintage adding machine, with a green clock.
Editorial illustration / AI-generated

KCM Cable, a custom cable and wire-harness manufacturer in Tooele, Utah, moved its quoting into Cableteque's software. President Kory Ewell says one quote alone saved him about 18 hours of work, and that customers have told him the company's pricing is more competitive. For wider context, Deloitte's June survey of 140+ manufacturers found 84% already get measurable value from AI, but only 20% of their use cases are scaled.

~18hrs

Saved on one complex quote, in the president's words

Vendor-reported

Vendor-reported Cableteque's own case study, undated, and one quote rather than an average. The Deloitte figures are a survey published by Deloitte Switzerland.

Source: Cableteque, undated · Deloitte, AI in Manufacturing 2026

My take

It sounds like the president was the one doing those 18 hours. In a lot of owner-led manufacturers, quoting stays with the owner because the pricing logic lives in their head. The software can't write that logic down for you. That part's still on you, and it's the part that frees you up.

Distribution

A building-products distributor turned reorder reminders into sales calls

Collage: stacked lumber and millwork trim beside an open order ledger and a reorder tag with a green dot.
Editorial illustration / AI-generated

Building Products Inc., a millwork and building-materials distributor that supplies retail lumberyards, rolled out Proton.ai's AI CRM, which prompts reps when a customer is due to reorder. CEO Kyle Schull says the “due to reorder” feature made the sales team much more proactive, rather than just reacting to incoming orders. For scale, MSC Industrial, a far bigger distributor, reports a 13% conversion rate on its AI recommendations.

27%

Conversion rate on the AI's recommendations

Vendor-reported

Vendor-reported Proton's case study from March, older than the rest of this issue. The MSC figure comes from Distribution Strategy Group's August roundup, which credits it to a Proton case study too.

Source: Proton, Mar 16, 2026 · Distribution Strategy Group, AI 25

My take

Honestly, this is AI remembering things, which isn't glamorous. But reorders are money that's already yours, and a lot of reps only find out about them when the order shows up. A reminder at the right time turns that into a call.

Insurance

An insurance agency put AI on its phones and bound 41% more policies

Collage: a vintage telephone operator headset on a desk beside insurance policy folders, one sealed in green.
Editorial illustration / AI-generated

bolt Agency, the in-house property and casualty agency of insurance technology company bolt, writes about $30M in annual premium. Over six months it used AI to answer inbound calls, qualify and route leads, and follow up on quotes. AI now handles more than 90% of its inbound calls.

+41%

Bound policies over six months

Vendor-reported

It also reports 1.6x lead conversion and a 57% cut in handling and quote time.

Vendor-reported bolt sells insurance technology, so this is its own product running in its own agency. Baseline metrics were tracked, but the before numbers aren't published. Published by the American Agents Alliance.

Source: American Agents Alliance, Sep 5, 2026

My take

Take this one with the most salt in the issue, since the agency and the software come from the same company. What I'd keep from it is where they aimed the AI: the phones and the follow-up, which are the two places leads go quiet without anyone noticing.

Law

A 20-lawyer firm tried an AI chatbot for intake, then took it back out

Collage: a vintage law office desk with an open appointment book and a telephone, a small tin robot set aside in a drawer.
Editorial illustration / AI-generated

Damian Turco runs Turco Legal, a 20-attorney divorce and family law firm. It tried a website chatbot to greet prospective clients, didn't see conversion improve, and went back to letting people book a free consultation directly with a senior attorney. His reasoning: “I'm not going to have a prospective client in that important moment… communicate with a bot instead of a person.”

Anecdotal One firm's experience, based on what he observed rather than a formal measurement. Independent legal press.

Source: Wisconsin Law Journal, Sep 7, 2026

My take

This fits the rest of the issue. The AI wins came from buyers who were already leaning in. A family-law prospect in a hard moment wants a person, and Turco decided that first conversation was too important to hand off. Worth figuring out which of your conversations are like that before you automate anything.

The takeaway

What this means if you're still the one selling

Across a software company, a CPA firm, a freight broker, a manufacturer, a distributor and an insurance agency, the AI wins came from the same place: the moment a buyer had already raised a hand, whether that was a quote request, a reorder, a call or a chat. AI answered faster, followed up without forgetting and filtered out the wrong fits.

AI hasn't taken over the middle of the sale yet: discovery, the plan, the business case. That's still where a hesitant buyer becomes a customer, and in most owner-led companies it still runs through the owner. Get that part written down so someone else can run it, and let the AI handle the edges.

Cheers,
TeeJay Johnson
Owner, Foundry Revenue Partners

Collage: a stone arch on a blueprint. Machine arms set green blocks at both ends while a brass drafting tool lowers the carved keystone into the top.
Editorial illustration / AI-generated
Free pre-assessment

Is your sales process ready for AI?

The free pre-assessment looks at how your sales engine is running today against your goals, before anything gets recommended.

Book a call

All issues