Stu Willson (@stuartawillson) header
Stu Willson (@stuartawillson) profile picture

Founder, Pluris (https://xitter.gq/t.co/9LpUqioi3N) + @aireceiptswork. Vetted independent AI teams: what’s worth doing, what it costs, who can build it. Host of Proof of Work

Los Angeles, CA
Joined September 2008
Photo by Stu Willson (@stuartawillson)Photo by Stu Willson (@stuartawillson)Photo by Stu Willson (@stuartawillson)Photo by Stu Willson (@stuartawillson)Photo by Stu Willson (@stuartawillson)Photo by Stu Willson (@stuartawillson)Photo by Stu Willson (@stuartawillson)Photo by Stu Willson (@stuartawillson)Photo by Stu Willson (@stuartawillson)Photo by Stu Willson (@stuartawillson)
If you’ve fixed something real with AI, I want to put it in front of the people who have the same problem. Most people and companies don’t go looking for “an AI consultant.” They go looking for someone who’s already solved the thing that’s eating their week: invoices matched by hand, quotes that take three days, a support queue nobody can keep up with. So I built ai receipts. A business person finishes the sentence “It sucks that...” and gets back real fixes, each with the name of the person who built it. If they want to talk to you, they ask for an intro and we make it. We’ve seeded it with hundreds of fixes from the Pluris AI Use Case Library. We know there are hundreds of thousands more out there, and we want a place where they can be found, no matter how big or small the fix. Sharing one takes a few minutes. Say what sucked, how you fixed it, roughly what it cost and how long it took. I review it, and once it’s up you get a link to share. Link below. If you’ve shipped something you’re proud of, share it. We built a home for it. aireceipts.work/share?utm_so…
4
5
1,613
👇
a good example of our AI project at work: - reach out to a company in Oct '24 - receive response in April '26 after about 15 touches; "no thanks" - put in Garden - receive ping around a spiking signal in Sept; reach back out; get a call; currently pursuing LOI 🤘 how it works: padraicmcconville.com/writin…
1
3
798
Stu Willson retweeted
A mortgage brokerage was re-typing every client and deal between Finsure and HubSpot. Each new enquiry and every deal stage update meant entering the same info twice, by hand. They now have a two-way sync that checks both systems four times a day, matches fields, cleans up duplicates and keeps contacts and deals up to date in both places. Now, contacts and deals move across loan stages without anyone re-typing details. The team can update either system and see it reflected in the other. Just shared on ai receipts by Ben Ron at Workvolve. aireceipts.work/r/recpOoxZY5…
1
3
82
A behavioral health provider was paying $849 for every lead from Google Ads, and a third of those leads never got a call back. Every lead, patient and payment lived in one Excel sheet. Cost per lead is now under $150, and revenue went from about $2.2M to $3.3M. It took 4 to 8 weeks and cost under $25K, and they’re opening a second location on the same setup. What OpsFixers built is in the comments, and you can ask Andy about it there. There are 49 more AI solutions in healthcare on file, from 36 other builders. Built by Andy Le at OpsFixers. This one’s on @aireceiptswork: What OpsFixers built: aireceipts.work/r/reckC2OOCF… AI solutions in healthcare: aireceipts.work/industries/h…
2
3
295
A lot of people saved this, so here's why I think that’s case: Lots of people sell AI automation. Very few can walk into a $60M specialty distributor running a 15-year-old ERP and build something people actually use that scales reliably. A mid-market buyer is screening for specific things. Has this firm done this exact workflow in my industry? Can they work with the systems we already have? Can they handle data that lives in spreadsheets and people's heads? Will they own adoption, or just hand over a login? For PE-backed companies, it also has to show up in EBITDA within the hold period. Every jump in model capability makes the easy stuff self-serve and makes custom software economical for problems that were never worth solving before. Some of it is agentic, some of it is ordinary software that used to cost too much to build. All of it is new demand, and it needs someone who knows the workflow, the systems and the people. If you count everyone who can build an n8n flow, the market looks crowded. If you count the people who've done computer vision QC on an extrusion line, or prior auth automation for a multi-site specialty practice, the list gets short fast, and finding them is the real problem. If you've shipped AI inside a real operating business, show the work on @aireceiptswork . If you're a buyer, you can browse what builders have shipped there, reach out to them, and check whether they're a fit without revealing who you are (aireceipts.work). If you'd rather have white glove service, Pluris will match you with the right specialist from our vetted network (checkpluris.com).
People talk about the AI implementation market being crowded. It is not.
24
77
6,039
👇
We’re looking for AI implementers and consultants who work with small businesses and teams. Demand for Littlebird from small teams is outpacing what we can staff, so we’re ramping an implementation partner program fast. If this is you, email [email protected] or DM us.
194
An industrial company that designs custom projects for large clients had a ceiling on how much work it could take. Every project needed drawings and a bill of materials, and a team of drafters rebuilt each one by hand, copying from one tool to the next. It ate most of their hours. The same team now turns out twice as many projects, which frees up about 7 FTE of drafting. Engineers review and approve each design instead of drawing it. It took 4 to 8 weeks to build and cost under $25K. The hard part was turning what the design team knew into something a system could apply. How Tauler Group did that is on AI Receipts, and you can ask Alvaro about it directly. There are 25 more AI solutions for manufacturers on file, from 20 other builders, linked there too. Built by @toledotauler at Tauler Group. This one’s on @aireceiptswork. 🟩 How Tauler Group built it: aireceipts.work/r/recbbEUjnf… 🟩 All AI solutions for manufacturers: aireceipts.work/industries/m…
206
People talk about the AI implementation market being crowded. It is not.
we need 10x more AI deployco’s - especially for the mid-market. no amount of AGI will magically create utopia while I’m excited that physical AI is having its moment, software folks should think twice before over-rotating into a 10x tougher problem space, when there’s untapped greenfield in plain sight
18
4
2
157
23,081
💯 👇
we need 10x more AI deployco’s - especially for the mid-market. no amount of AGI will magically create utopia while I’m excited that physical AI is having its moment, software folks should think twice before over-rotating into a 10x tougher problem space, when there’s untapped greenfield in plain sight
1
267
A growing construction company wanted to bid on more work, but experienced estimators are hard to hire and the ones it had were buried in prep. Preparing a 200-line tender took around 40 hours. Now it takes about 8, and the estimators still make every pricing call. The hard part was everything the company already knew about prices, suppliers and past jobs, scattered across spreadsheets and people’s heads. How AI Point 8 put that to work is in the comments, and you can ask Stipe about it there. There are 13 more solutions for quotes and bids like this on file, from manufacturers, contractors and distributors. Built by @StipePoint8 at AI Point 8. This one’s on @aireceiptswork. How AI Point 8 built it: aireceipts.work/r/recU5ASE28… All solutions for quotes and bids: aireceipts.work/problems/quo…
1
1
4
211
Stu Willson retweeted
A two-location restaurant group in South Florida handled invoices and payroll by hand, with no way to check vendor prices or keep point-of-sale data clean. Each GM did things their own way, and there was no employee handbook or training. Now, an AI tool reads invoices and codes every line item to the right category, checks prices, and pulls in sales data. Staff training, handbooks, quizzes and GM checklists are all in one place. They can now track prices, keep payroll straight, and train staff the same way at both locations. Just shared on ai receipts by Andy Le at OpsFixers. aireceipts.work/r/recItr3gFa…
1
1
58
Stu Willson retweeted
A video production company for physicians ran projects across four tools by hand. Task status, file review and payments were scattered, so someone had to copy updates between them and things got missed. Now, an automated system connects ClickUp, Frame.io, Stripe and Google Drive. Projects, tasks and folders are created automatically, review feedback and payments sync, errors trigger alerts, and AI drafts scripts for approval. The client can now run projects without copying updates between tools, and hand over the system with full documentation. Just shared on ai receipts by Ahmer Shehzad at Evaluix AI. aireceipts.work/r/recwixhZIO…
1
1
58
Stu Willson retweeted
A multi-service wellness studio used booking software that didn’t match its needs. Staff schedules, timesheets, time off, memberships and credits were all tracked separately or by hand. They built a custom operations platform that ties bookings to staff shifts, includes a time clock, manages timesheets and time off, tracks memberships and credits, stores payments, and offers one-click onboarding for clients. All 5,360 client records were migrated with no downtime. The studio now runs everything from one place. Just shared on ai receipts by Andy Le at OpsFixers. aireceipts.work/r/reckYhlgbJ…
1
1
86
Stu Willson retweeted
A national car repair business was manually collecting evidence and fixing fleet billing every month. This delayed invoicing and kept finance staff busy with repetitive work. AI agents now handle the entire process, running in a custom system built for finance tasks. Billing can happen right at month end, and staff spend 30 percent less time on this job. Just shared on ai receipts by Richard Hurley at Wearecuria.com. aireceipts.work/r/rect5QL871…
1
1
86
Stu Willson retweeted
A commercial real estate broker spent hours prospecting by hand, writing every cold email from scratch, and paying a virtual assistant who wasn’t producing results. They built an AI prospecting system that finds companies showing hiring and fundraising signals, drafts outreach in the broker’s own voice for approval, and syncs with his CRM through a custom MCP server. The junior broker can now run the entire prospecting process, from lead identification to email drafting and CRM updates, without outside help. Just shared on ai receipts by Andy Le at OpsFixers. aireceipts.work/r/rec9Jx8Ydi…
2
1
2
122
Stu Willson retweeted
A US B2B wholesaler with over 200 orders a month built every invoice by hand in QuickBooks from Pipedrive deals, taking two to four minutes per invoice. A bot now builds each invoice in QuickBooks straight from Pipedrive, flags problems for review, and reads purchase order numbers from attachments. In three and a half months the bot prepared 597 invoices, of which 530 were confirmed sent. It read and placed purchase order numbers from attachments on 26 invoices. Just shared on ai receipts by Wesley de Nooijer at Pretrain (@pretrainwes). aireceipts.work/r/rec6uk1qco…
1
1
103
Stu Willson retweeted
A security training company’s certification platform launched with open admin access, exposed passwords, broken checkout and missing student records. The team audited the platform, fixed the security issues, rebuilt payments on Stripe and recovered historical student data. They now have a secure platform, working payments, restored student records and an ongoing tech partner. Just shared on ai receipts by Andy Le at OpsFixers. aireceipts.work/r/reckbDZbIi…
1
1
61
Stu Willson retweeted
A construction company wanted to bid on more projects, but every estimate took about 40 hours. Their team spent that time piecing together prices from old spreadsheets, emails and subcontractor quotes. Now, an AI-assisted platform brings all their historical pricing, resource costs and subcontractor quotes into one workspace. It structures new tenders, suggests prices with sources, and lets estimators review, adjust and compare everything in one place. For a typical 200-item estimate, prep time dropped from 40 hours to 8 hours. That is 32 hours saved, or an 80 percent reduction. Just shared on ai receipts by Stipe Renic at AI Point 8 (@StipePoint8). aireceipts.work/r/recU5ASE28…
1
1
66
Stu Willson retweeted
A hospitality operator with 400 seasonal workers had HR teams repeating the same answers about arrival, uniforms and procedures. Information was scattered across manuals and conversations, with different details for each property. They built a multilingual assistant that answers staff questions from approved documents and FAQs, tailored to each property. HR controls access, updates answers and reviews common questions through a central console. Now, workers get answers in their language through a web app or WhatsApp. HR keeps company knowledge up to date in one place. Just shared on ai receipts by Stipe Renic at AI Point 8 (@StipePoint8). aireceipts.work/r/recgdqdupB…
1
1
65
Stu Willson retweeted
A Dutch solar and battery installer with 10 staff used to write every quote by hand after each sales call. Drafting a quote took about 9 hours on average, so quotes sometimes went out late. Now, every sales call is transcribed and turned into a draft quote in Moneybird. The salesperson checks or tweaks the draft in Telegram on their phone. A draft is ready in about 3 minutes. The team now sends 85 quotes a month, up from 53, and saves 32 hours of quote work each month. Just shared on ai receipts by Wesley de Nooijer at Pretrain (@pretrainwes). aireceipts.work/r/recAJX8Oiu…
1
1
75