AI in the sale. What owner-led companies are doing with it, and what it's worth.October 6, 2026 · 7 stories
This week
The jobs only the owner could do
A shop owner quoting in 10 minutes instead of two-plus hours, an agency's statement of work out within minutes of a yes, and a £4,200 lesson.
Editorial illustration / AI-generated
Hey,
A lot of owner-led companies run on a sentence that starts “only I can”: only I can price it, only I can write the proposal, only I know what this client actually said on the call. I think it's often true, and it's often the thing that keeps the business from growing past the owner.
So most of this week's stories are one of those jobs, and what happened when a company tried to move it off one person with AI. I've sorted them by what the job really needed from that person: their time, their memory or their judgment. Two didn't go to plan, and those are the ones I'd read first. Most of the numbers come from the companies selling the tools, so the labels matter again this week.
01
Your time: pricing
A machine shop owner used to spend two hours or more quoting a job. Now it takes about 10 minutes.
Editorial illustration / AI-generated
Preston Murray co-founded Bullray MFG, a machine shop in San Clemente, California, with a friend after working as a fabricator and model maker at Rivian. He quotes with Uptool, AI quoting software built for small, high-mix shops. “Pricing out and quoting a job used to take me two hours or longer. Using Uptool, it now takes about 10 minutes,” Murray told IMTS. “It also keeps our pricing competitive, so we are not under- or overestimating.” Nathan Dillon, general manager and owner of VelocityCNC, described the same job in a February product announcement: “I would spend hours a day quoting – time we don't charge for.”
2+ hrs10min
To price and quote a job at Bullray MFG
The owner's words, in an IMTS feature on Uptool
Dillon says cutting his quoting time 10x in 2025 contributed to his biggest sales year to date, “more than double the previous year.”
The workflow
Uptool, AI quoting software for small machine shops
Builds the price and the quote for a job the owner used to work out by hand.
What it changed About 10 minutes per quote, down from two hours or longer, in the owner's words.
Vendor-reported Both accounts come from pieces that promote Uptool: an undated IMTS feature on the exhibitor (around IMTS, September 2026) and a February product announcement. The feature doesn't give Bullray's size; it describes Murray as an owner-operator, so it's likely well below most readers'. Dillon's sales year is his own claim, and the tool is one of the things he credits.
Two hours a quote is a cost nobody invoices. If you're the one pricing every job, time it for a week and write down what you actually look at to land on a number. That list is what any quoting tool, or any new estimator, will need from you.
02
Your time: the proposal
An agency that spent weeks per deal on scoping, pricing and proposals now has the statement of work out within minutes of a yes
Editorial illustration / AI-generated
KANE, a creative and growth agency in Princeton, New Jersey, used to spend weeks per deal on scoping, pricing, proposal writing and contracting. Co-founders Michael and Joni Kazantzis built a Claude pipeline that, in Anthropic's words, “scores fit from the meeting notes, prices from their own past projects, drafts the proposal, and has the statement of work out for signature within minutes of a yes.” Michael Kazantzis: “We've taken an entire process that used to take us weeks and it's done almost within a few minutes in a day, and even the clients recognize that.”
The workflow
Claude, set up as a scoping-to-contract pipeline
Scores fit from the meeting notes, prices from the agency's own past projects, drafts the proposal and prepares the statement of work.
What it changed The statement of work goes out within minutes of a yes, where the process used to take weeks per deal.
Vendor-reported Told by Anthropic, which sells Claude. Claude is also the AI I use to help research this newsletter, so take that into account. The agency's size isn't given, and no win rate or revenue figure is reported.
The phrase I'd notice is “prices from their own past projects.” My guess is a lot of owners price proposals from memory, and part of why this works, I'd think, is that the past projects were there to learn from. If yours live in old emails and your head, gathering them into one folder is the first step, with or without AI.
03
Your memory: what the client said
A consultant let AI write her proposals for six months. She says it cost her a £4,200 deal.
Editorial illustration / AI-generated
£4,200
deal lost to a rival whose proposal “mentioned things we'd talked about”
~40
proposals drafted with AI over six months
40–50 min
per proposal now, down from about three hours
~1 in 3
close rate on proposals over £1,500, up from roughly 1 in 4
Lilach Bullock, her own account, July 2026
Lilach Bullock, a UK marketing and AI consultant, used ChatGPT to draft roughly 40 proposals over six months. Each one used to take her about three hours. A mid-sized accountancy firm in Reading had told her on the call that its team “hated posting on LinkedIn because a director had been trolled there in 2019 and everyone was still nervous about it.” Her AI draft turned that into “addressing internal reluctance to post,” which she calls “technically accurate and completely dead.” She lost the £4,200 job to a competitor charging more, and the prospect told her the other proposal “mentioned things we'd talked about” and hers felt like it could have gone to anyone. A £900 workshop client had received three near-identical proposal structures from three consultants that same week, and said it made her feel like she was “buying a template, not hiring a person.”
She kept the AI draft but added a 15 to 20 minute manual rewrite to put the client's specifics back in. Proposals now take her about 40 to 50 minutes, and she says her close rate on proposals over £1,500 has gone from roughly 1 in 4 to closer to 1 in 3, across 25-plus proposals.
The workflow
ChatGPT for the first draft, plus a manual rewrite
Drafts the proposal from her notes; she then spends 15 to 20 minutes writing the client-specific detail back in.
What it changed About 40 to 50 minutes per proposal instead of three hours, and a close rate she puts at closer to 1 in 3 on larger proposals.
Anecdotal One consultant's account of her own pipeline, which she calls “not a scientific study.” Published in July, outside our usual 60-day window. She sells AI consulting, and a one-person business is far smaller than most readers'.
This is the most useful failure I've read in a while. The AI had her notes and still sanded the specifics off, and the specifics were the part only she knew. If you hand proposals to a tool, have the person who ran the call put those details back in and do the final read.
04
Your memory: the follow-up
A small trailer maker says deals that took 76 touches now close in 13
Editorial illustration / AI-generated
SylvanSport makes adventure trailers and camping gear in Brevard, North Carolina, and was founded in 2004; Tom Dempsey is its founding partner and CEO. It now uses HubSpot's Prospecting Agent along with new workflows. Erika Holland, who leads customer experience and AI strategy there, says that last year a deal took an average of 76 touches to close, and that “thanks to the Prospecting Agent and our new workflows,” the company now closes in 13. Dempsey says it gave him “something that I have wanted and needed for our 20+ years in business but no one has been able to deliver.”
7613
Average touches to close a deal
Vendor-reported, via HubSpot
The workflow
HubSpot's Prospecting Agent, plus new workflows
Runs the company's prospecting follow-up inside HubSpot alongside new workflows.
What it changed Average touches to close a deal down from 76 to 13, in the company's words.
Vendor-reported Quotes on HubSpot's own case-study page, undated, with no detail on what counts as a touch (her word is touchpoints), who the buyers are or how many deals. The company credits its new workflows as well as the agent, and its size isn't given.
The quote credits the new workflows alongside the agent, and my read is that someone finally decided what the follow-up should be. When follow-up depends on who remembers, it usually depends on the owner. I'd write the sequence down first, then let a tool run it.
05
Your judgment: estimating
A $40–60 million printer looked at handing estimating to AI, decided it wasn't ready, and pointed AI at freight quotes instead
Editorial illustration / AI-generated
Printing Impressions profiled a commercial printer with annual sales of $40 million to $60 million that began implementing AI in early 2024; its executives asked not to be named. Estimating, a job for its team of estimators, looked like a strong candidate, but projects spanning print, mail, fulfillment and logistics “proved challenging for AI to fully grasp.” “We realized we weren't ready for AI to take that on end-to-end,” one executive said. “So instead, we focused on tools that support estimators, like shipping and reporting, rather than trying to replace them.” Freight quotes used to need several handoffs, so the company built an AI agent trained on its negotiated carrier rates and shipping options.
Staff can now get shipping costs and transit times in real time and respond to customers almost instantly.
The workflow
An in-house AI agent trained on negotiated carrier rates and shipping options
Answers staff questions on shipping costs and transit times, replacing a chain of handoffs.
What it changed Freight answers in real time, so staff can respond to customers almost instantly. No figures given.
Anecdotal The company is unnamed at its executives' request, so the account can't be checked. Published in February, outside our usual window, with no numbers beyond revenue.
I like this one because they said no to the obvious target. Estimating at a printer takes judgment across print, mail and logistics, so they pointed AI at a slow step around it instead. If your hardest job won't hand off yet, look for the slow step beside it.
06
What it's worth
21% of owners got a written offer last year. 1% were ready to sell.
Editorial illustration / AI-generated
21%
of owners got a written offer last year
1%
were ready to sell
70%
have only early-stage or no formal succession plans (Chase)
8%
say they're fully prepared to hand over ownership (Chase)
Value Builder Exit Readiness Index, Q2 2026 (1,032 assessments); Chase survey of about 1,000 small business owners, March 2026
Value Builder, which scores companies on how sellable they are, says 21% of owners in its latest quarterly index got a written offer last year, and only 1% were ready to sell. Unsolicited offers rose from 13% to 21% of owners in 18 months, the median owner is now 54, and 58% plan to exit within five years. A separate Chase survey of about 1,000 small business owners in March found 40% plan to retire within the next decade, 70% are in early-stage planning or have no formal succession plan, and 8% report being fully prepared to transition ownership.
Directional Value Builder sells its assessment to advisors who work with owners, and the owners in the index took it, so they aren't a random sample. The public page doesn't define “ready to sell,” and Chase doesn't give company sizes. Neither source ties readiness to who does the selling; that link is my view, in the take.
As I understand it, buyers pay more for a business that keeps selling after the owner steps back. When the pricing, the proposals and the key relationships all run through you, a buyer will see it, and I'd expect it to show up in the offer. My guess is every job in this issue that moved off one person makes a business a little easier to buy.
07
Where AI goes
Fast-growing companies put more AI on their sales teams this year, and fewer of their CEOs think it will make sellers more productive
Inc.'s survey of the leaders of America's fastest-growing private companies, run with Salesforce, found AI deployment among sales teams rose from 49% last year to 62% in 2026. Over the same year, the share who think AI will increase seller productivity shrank from 27% to 22%, and the average sales team went from 41 people to 35. Even so, 78% of the CEOs expect to expand their sales team in the next year.
Inc. 5000 CEOs on AI in sales
Inc. 5000 CEOs on AI in sales
Sales teams with AI deployed, 2025
49%
Sales teams with AI deployed, 2026
62%
Think AI will increase seller productivity, 2025
27%
Think AI will increase seller productivity, 2026
22%
Inc. CEO Survey, run with Salesforce, 2025 and 2026. Sample size not published.
Directional Inc. 5000 companies are faster-growing than most readers', and the respondents aren't broken out by size. Inc. doesn't publish the sample size or the exact questions, and Salesforce, which co-ran the survey, sells AI for sales.
More AI and less faith in it is roughly what I'd expect when tools land on jobs nobody has written down. The companies in this issue that got results picked one job, often one the owner used to carry, and moved it on purpose. I'd start there rather than with a rollout across the whole sales team.
The workflows behind the results
Every tool in this issue, what it does, and what it changed.
Look at what the tools leaned on. KANE's proposals worked from its real past projects. Lilach Bullock's proposals only started winning again when the details from the call went back in. SylvanSport credits its new workflows alongside the agent. And the printer's estimating, which I'd guess ran on judgment nobody had written down, stayed with people.
So here's a 20-minute exercise for this week. List every job in your sale that only you do. Next to each one, write whether it needs your time, your memory or your judgment. My rough rule: time is the easiest to hand off, memory has to be written down first, and judgment moves last. Pick one from the time column and start there.
Cheers, TeeJay Johnson Owner, Foundry Revenue Partners
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.