On 5 August, @Data Swami posted in the General Business Forum with a question that ought to have a short answer.
He had spent six months building out an AI agent setup for his own business, working through the errors, the wasted usage and the blind alleys until it did what he needed. People had started asking how he had done it. So he was thinking about selling it as a done-for-you service: he installs and configures the whole environment, then provides hosting, maintenance, fixes and, importantly, coaching on how to actually use the thing.
His proposed pricing was £1,000 to £2,500 for setup, then £500 to £750 a month. His question was narrow. Where in those ranges should he land?
Thirty-two replies later, nobody had given him a number.
That is not the forum being awkward. That is the forum doing the thing it does best, which is declining to answer the question you asked because the question underneath it matters more. And the underneath question, in almost every pricing thread I have ever read on UKBF, turns out to be the same one.
Data Swami answered them properly. Target: solopreneurs, one-person businesses, coaches and consultants, probably not a one-man-band plumber. Saving: time, and possibly money if it lets people drop subscriptions to software that half-fits. Competitors: basic ChatGPT and Claude agent setups, nothing quite comparable. Goal: value.
Those are reasonable answers. They did not survive contact with the rest of the thread.
Then Paul Kelly made the point that, for my money, is the most valuable thing in the whole thread: "But this is the issue - they do not know what they need."
He was not being rhetorical. He described his own position, which is that he wants to find a daily use for AI in his business and cannot. He set up OpenClaw, called it Jeeves, and it worked fine. Then he realised there was no specific operation he could hand it that would genuinely change his week. He might get round to building a marketing assistant at some point. His business will survive if he does not.
That is not scepticism from the sidelines. That is a demand report from exactly the customer profile being described, delivered for free, in the thread. If you are selling to small business owners and one of them tells you they want to buy what you sell and still cannot work out why, that is the most useful sales research you will get all year.
Value-based: define your ideal customer, map the economic benefit precisely, then price as a share of the value captured. If it saves or makes them a meaningful amount, the price reflects the outcome, not the hours you put in.
Cost-plus: model every direct and indirect cost, including API usage, infrastructure, software and your own support time, add your target margin, project your client volume, and that gives you a floor.
Neither number is the price. The gap between them is where the price lives. Run both and the answer usually becomes obvious.
EnterprisePro also raised the segment question, and it stung a bit. Solopreneurs are solopreneurs partly because they want to keep costs down and do things themselves. For that group, a self-service product with a free trial may fit better than a consultative setup. The growing firms Data Swami already works with looked like the easier market.
How much time does it save, whose time is it, and what is that time worth to them? What could they be doing instead with those hours? And the question most people building a product never ask: is there a chance you are taking away work your customer actually enjoys doing?
He also offered a selling line that I suspect a few people will steal. Ask a prospect what they would pay to free up three hours a week. If they say a hundred quid, you have learned they value their own time at £33 an hour. If they answer inside your range, you have a deal. If they answer above it, you were underpriced. He was candid enough to add that this is much easier to pull off on a forum than in a real conversation, which is fair.
Data Swami had listed what the setup could build: invoices, social scheduling, landing pages, small apps, automated routing of email leads. fisicx went through them one by one and pointed out that almost all of it is one-off work. You build the app once. You build the landing page occasionally. You set up the lead routing once and then it runs. His conclusion was that the setup fee made sense to him and the subscription did not.
Data Swami's answer was that the monthly fee is mostly coaching and one-to-one time, with maintenance as the smaller part, and that clients can stop paying once they are running on their own.
fisicx pushed once more: for how long does someone need coaching, and how many clients can one person support at that intensity before the model breaks?
That one is not resolved, and I am not going to pretend it is. Both positions are defensible. But the test underneath it applies to everyone reading this: for every element of your monthly fee, name the work that actually recurs every month. Anything that does not recur belongs in the setup fee. Customers work this out eventually, and they work it out at renewal.
That reframing matters. Buyers do not experience £500 a month as £500 a month. At some point in the sales conversation, usually the point at which they are about to sign, they multiply it by twelve. If your pitch is not built to survive that multiplication, it will not.
VectorNavGRP's other point closed the loop on the whole thread: a consultant billing £1,000 a day who recovers a day a month sees this as obviously cheap. Another one-person business will not justify it at any price, however good the product is. Same product, same price, completely different answer, and the only variable is who you sold it to.
It starts at solopreneurs, consultants, one-person businesses and very small teams. By reply five it excludes plumbers. By reply eleven it has moved towards small professional services firms. By the end, Data Swami is describing something specific: financial advisers with annual review cycles, drowning in admin, wanting to take on more clients and unable to because of the paperwork. That is a customer you can find, phone and quantify.
He came in asking for a price. He left with a market. On the evidence of the thread, that was the better trade.
The thread is still open if you want to add to it: How do you price a new product/service?
AI disclosure: I used AI assistance to help structure and draft this piece. Every member contribution referenced above was taken from the live thread and attributed to the member who posted it. No quotes or examples were invented.
Related reading
He had spent six months building out an AI agent setup for his own business, working through the errors, the wasted usage and the blind alleys until it did what he needed. People had started asking how he had done it. So he was thinking about selling it as a done-for-you service: he installs and configures the whole environment, then provides hosting, maintenance, fixes and, importantly, coaching on how to actually use the thing.
His proposed pricing was £1,000 to £2,500 for setup, then £500 to £750 a month. His question was narrow. Where in those ranges should he land?
Thirty-two replies later, nobody had given him a number.
That is not the forum being awkward. That is the forum doing the thing it does best, which is declining to answer the question you asked because the question underneath it matters more. And the underneath question, in almost every pricing thread I have ever read on UKBF, turns out to be the same one.
The four questions that came back first
@Paul Kelly ICHYB got in first, and did not offer an opinion on price at all. He asked four things instead: who is the target customer, what does the product save in time or money, where are competitors on price, and is the goal volume or value?Data Swami answered them properly. Target: solopreneurs, one-person businesses, coaches and consultants, probably not a one-man-band plumber. Saving: time, and possibly money if it lets people drop subscriptions to software that half-fits. Competitors: basic ChatGPT and Claude agent setups, nothing quite comparable. Goal: value.
Those are reasonable answers. They did not survive contact with the rest of the thread.
"They do not know what they need"
@fisicx came at it from the other end, asking what the product actually is, why a bloke cleaning windows would need it, and whether the target customer can afford it at all. He put the number back on the table: £500 a month without visible, tangible benefit is a lot of money for a small business.Then Paul Kelly made the point that, for my money, is the most valuable thing in the whole thread: "But this is the issue - they do not know what they need."
He was not being rhetorical. He described his own position, which is that he wants to find a daily use for AI in his business and cannot. He set up OpenClaw, called it Jeeves, and it worked fine. Then he realised there was no specific operation he could hand it that would genuinely change his week. He might get round to building a marketing assistant at some point. His business will survive if he does not.
That is not scepticism from the sidelines. That is a demand report from exactly the customer profile being described, delivered for free, in the thread. If you are selling to small business owners and one of them tells you they want to buy what you sell and still cannot work out why, that is the most useful sales research you will get all year.
Cost-plus, value-based, and the floor
@EnterprisePro set out the two models side by side, which is worth repeating because it is the cleanest summary of pricing method anyone has posted on the forum recently.Value-based: define your ideal customer, map the economic benefit precisely, then price as a share of the value captured. If it saves or makes them a meaningful amount, the price reflects the outcome, not the hours you put in.
Cost-plus: model every direct and indirect cost, including API usage, infrastructure, software and your own support time, add your target margin, project your client volume, and that gives you a floor.
Neither number is the price. The gap between them is where the price lives. Run both and the answer usually becomes obvious.
EnterprisePro also raised the segment question, and it stung a bit. Solopreneurs are solopreneurs partly because they want to keep costs down and do things themselves. For that group, a self-service product with a free trial may fit better than a consultative setup. The growing firms Data Swami already works with looked like the easier market.
Start with what it is worth
@ZipserSir put it most directly: you started with the wrong question. Not "what do I charge", but "what is it worth".How much time does it save, whose time is it, and what is that time worth to them? What could they be doing instead with those hours? And the question most people building a product never ask: is there a chance you are taking away work your customer actually enjoys doing?
He also offered a selling line that I suspect a few people will steal. Ask a prospect what they would pay to free up three hours a week. If they say a hundred quid, you have learned they value their own time at £33 an hour. If they answer inside your range, you have a deal. If they answer above it, you were underpriced. He was candid enough to add that this is much easier to pull off on a forum than in a real conversation, which is fair.
The bit about subscriptions that most people get wrong
The sharpest structural point came from fisicx, and it is portable to any business considering a monthly fee.Data Swami had listed what the setup could build: invoices, social scheduling, landing pages, small apps, automated routing of email leads. fisicx went through them one by one and pointed out that almost all of it is one-off work. You build the app once. You build the landing page occasionally. You set up the lead routing once and then it runs. His conclusion was that the setup fee made sense to him and the subscription did not.
Data Swami's answer was that the monthly fee is mostly coaching and one-to-one time, with maintenance as the smaller part, and that clients can stop paying once they are running on their own.
fisicx pushed once more: for how long does someone need coaching, and how many clients can one person support at that intensity before the model breaks?
That one is not resolved, and I am not going to pretend it is. Both positions are defensible. But the test underneath it applies to everyone reading this: for every element of your monthly fee, name the work that actually recurs every month. Anything that does not recur belongs in the setup fee. Customers work this out eventually, and they work it out at renewal.
Nobody was looking at the year
@VectorNavGRP arrived late and did the arithmetic nobody else had done out loud. Setup plus twelve months is a £7,000 to £11,000 first-year decision.That reframing matters. Buyers do not experience £500 a month as £500 a month. At some point in the sales conversation, usually the point at which they are about to sign, they multiply it by twelve. If your pitch is not built to survive that multiplication, it will not.
VectorNavGRP's other point closed the loop on the whole thread: a consultant billing £1,000 a day who recovers a day a month sees this as obviously cheap. Another one-person business will not justify it at any price, however good the product is. Same product, same price, completely different answer, and the only variable is who you sold it to.
Watch the target move
Read the thread from top to bottom and the interesting thing is not the pricing advice. It is the customer definition shifting in real time.It starts at solopreneurs, consultants, one-person businesses and very small teams. By reply five it excludes plumbers. By reply eleven it has moved towards small professional services firms. By the end, Data Swami is describing something specific: financial advisers with annual review cycles, drowning in admin, wanting to take on more clients and unable to because of the paperwork. That is a customer you can find, phone and quantify.
He came in asking for a price. He left with a market. On the evidence of the thread, that was the better trade.
What to take from it
- The price question is nearly always the last question, not the first. If you cannot name the customer and the problem, no number will be right.
- Quantify the value in their numbers, not yours. Six months of your effort is not a benefit to a buyer.
- Separate the one-off from the recurring. Charge setup for setup and subscription for things that genuinely repeat.
- Show the first-year total before they work it out themselves.
- Sell the outcome, not the machinery. As fisicx put it, echoing Paul Kelly's phrase: sell the sizzle, not the sausage. People do not want AI. They want their week to be easier.
- Check your own capacity. A price that only works if you personally coach fifty people is not a price, it is a trap.
The thread is still open if you want to add to it: How do you price a new product/service?
AI disclosure: I used AI assistance to help structure and draft this piece. Every member contribution referenced above was taken from the live thread and attributed to the member who posted it. No quotes or examples were invented.
Related reading
