The short version: ChatGPT is good at pricing research, tier structuring, and pressure-testing your numbers against competitors, but it's useless at the one thing that determines your price: how much your customers believe you're worth. Use it for the maths and the market scan, not for the confidence.
What ChatGPT is good at when it comes to pricing
I've been pricing things badly for most of my career. My first ever speaking fee was £250 for a full day, back when I thought getting booked at all was the win. It took me nearly a decade to get to four figures for the same day, and most of that gap wasn't market conditions changing. It was me finally believing I could ask for more.
ChatGPT won't fix that belief problem for you. But it's a decent research assistant for the parts of pricing that are about data: what competitors charge, how to structure tiers, what language makes a price feel fair, and where your numbers have obvious gaps. Think of it as a very fast intern who has read every pricing page on the internet but has never once had to sit across from a client and hold their nerve while asking for £3,000.
The prompts I use
These are the ones I go back to, not a theoretical list. I've tested each of these more than once, on real services I've priced this year.
1. The competitor scan prompt
"I offer [specific service, e.g. a 90-minute AI strategy session for small business owners]. List 8 to 10 UK and US based providers who offer something comparable, and estimate their pricing based on publicly available information. Group them by price tier and note what's included at each level."
ChatGPT can't browse live pricing pages with certainty (prices change, and it will sometimes guess), so treat the output as a starting list, not gospel. I've had it name three consultants I then had to Google to confirm, and one of the three numbers it gave me was about 18 percent off. It's a shortcut to a shortlist, not a finished market report.
2. The tier-building prompt
"I sell [product/service] at a single price of [£X]. Help me split this into three tiers: a stripped-back version, my current offer as the middle tier, and a premium version. For each tier, suggest what to add or remove and a price gap that feels logical rather than arbitrary."
This one is useful for services people, because most of us undercharge by only having one option. When I did this for a half-day workshop I'd been selling flat at £1,200, ChatGPT suggested a £750 "lite" version (no follow-up materials, group delivery only) and a £2,000 version with a recorded session and a 30-day email follow-up. I built both. The £2,000 tier now sells more often than the original middle price did on its own, because buyers anchor against the top and the middle suddenly looks reasonable.
3. The value translation prompt
"My service costs the client £X. In plain, non-salesy language, describe what that £X buys them in terms of time saved, money made, or risk avoided. Give me 5 versions ranging from very literal to more emotional."
This is the prompt that earns its keep on a sales page or in a proposal. It stops you writing "£1,500 for a full day consulting" and gets you closer to "£1,500 for a day that saves you roughly 40 hours of trial and error, based on what it took me to figure this out the hard way." If you're building any kind of digital product alongside your services, this same prompt works well when you're figuring out how to price something in the digital products a solopreneur can sell in 2026, because value language matters more for a £47 template than a £5,000 retainer.
4. The objection stress-test prompt
"Play the role of a potential client who thinks my price of £X for [service] is too high. Give me their three strongest objections, then help me write a calm response to each that doesn't involve discounting."
I use this before every price increase I make. It's uncomfortable in a useful way, it forces you to write the answer to "why does this cost so much" before someone asks you in a call and you freeze.
5. The margin sanity-check prompt
"Here are my costs for delivering [product/service]: [list them, including your own time at a reasonable hourly rate]. My current price is £X. Calculate my margin and tell me what price I'd need to hit a 60 percent margin."
Most people pricing their own services never cost their own time, so their "margin" is fictional. This prompt forces the maths onto the page. When I ran it on a done-for-you social content package I was selling at £600 a month, it showed my real margin was closer to 22 percent once I priced my hours at £75 an hour, not the 70 percent I'd assumed. I put the price up to £950 within a month.
A real example: repricing my own workshop
Last year I had a half-day AI workshop that I'd been selling at £1,200 flat for about 18 months. It was fine. It wasn't wrong. But I kept feeling that slight cringe when I quoted it, which is usually a sign the number is too low, not too high.
I ran it through three of the prompts above in one sitting. The competitor scan showed similar half-day sessions from UK based consultants sitting anywhere from £900 to £3,500, with the higher end almost always including some kind of follow-up support. The tier-building prompt gave me the structure I mentioned earlier. The value translation prompt gave me a line I still use: "a day that would take you three months of trial and error to work out on your own."
The new pricing went live at £1,500 for the base version and £2,200 for the version with 30 days of email support. I lost precisely nobody over the price change. One client said the new tiered structure made the decision easier, because she could see exactly what she was and wasn't getting. That's the part ChatGPT helped with, structure and language. It did not, and could not, give me the nerve to hit publish on the new price. That bit was still mine.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
The uncomfortable part nobody puts in these lists
Here's the bit that gets glossed over in most "AI pricing prompts" posts. ChatGPT's training data skews toward what people already charge, which is to say it skews toward average, and average is usually too low if you're any good at what you do. Ask it to suggest a price for almost any service and it will drift toward the middle of whatever range it's seen, because that's the statistically safe answer. It is not designed to tell you that you're worth more than the market average. It's designed to give you a plausible, defensible, unremarkable number.
I've watched this happen with clients who ask ChatGPT to price a service and then treat the output as a ceiling rather than a starting point. If you're already nervous about charging what you're worth, an AI tool that hands you a "reasonable" middle-of-the-road figure will feel like permission to stay small. Use its numbers as a floor to check you're not underpricing wildly, not as the top of what's possible.
The second uncomfortable truth is that no prompt fixes a weak offer. If your service is interchangeable with ten others, no amount of clever tier language will let you charge premium rates. Pricing prompts work best on things that already have some differentiation, a track record, a specific outcome, or a niche. If you don't have that yet, the more useful work is building the thing worth paying more for, and something like the ChatGPT prompt pack for small business owners is a better starting point than a pricing prompt on its own.
A simple workflow, start to finish
If you want to reprice something this week rather than just read about it, here's the order I'd do it in:
- Run the competitor scan prompt and manually check three of the named competitors' actual current prices, because ChatGPT will sometimes invent or misremember specific figures.
- Run the margin sanity-check prompt with your real hourly rate included, not a nominal figure.
- Build three tiers using the tier-building prompt, even if you only launch two of them at first.
- Write your value translation lines and put the strongest one on your sales page or proposal template.
- Run the objection stress-test prompt and rehearse your answers out loud before your next sales call, not just in your head.
None of this takes more than about 90 minutes, most of it in a single ChatGPT session. The slower part is sitting with the new number for a day or two before you send it to anyone, because your gut will lobby hard to lower it again.
If pricing conversations are part of a bigger pattern of you avoiding admin and repetitive research tasks, it's worth reading about using ChatGPT agents to save 10+ hours a week, because a lot of small business owners who underprice are also the ones drowning in tasks that a bit of automation would clear, leaving more headspace for the harder strategic calls like pricing.
Where this fits with everything else you're using AI for
Pricing prompts sit alongside the other AI habits worth building into a small business, things like using AI for sales page images covered in AI images for business, or keeping across weekly shifts in the tools themselves via something like AI news roundups for small business. Pricing isn't a one-off task you do once and forget. Prices should move at least once a year for most services, and having a five-prompt routine you can rerun each January makes that a 90 minute job instead of the multi-week agonising it usually turns into.
And if you're weighing up whether a ready-made set of prompts is worth paying for versus building your own list like this one, that's a fair question, and I've answered it directly in what AI prompt packs are and whether they're worth buying.
If you want the full set, start at the ChatGPT Prompts and Limits: 30 Guides to Fun Prompts, Photo Edits and Daily Caps.
For a quick check, run it through my prompt generator.
Frequently asked questions
Can ChatGPT tell me exactly what to charge?
No. It can give you a market range, help you structure tiers, and stress-test your reasoning, but the final number has to account for your specific reputation, results, and demand, none of which it knows unless you tell it, and even then it will lean toward safe, average figures rather than what you could command.
What's the single best ChatGPT prompt for pricing a new service?
Start with the margin sanity-check prompt: list every cost including your own time at a fair hourly rate, give ChatGPT your proposed price, and ask it to calculate your real margin. Most people discover their margin is far thinner than they assumed, which is usually the push they need to raise the number.
Should I trust ChatGPT's competitor pricing research?
Treat it as a starting list, not a finished report. ChatGPT will sometimes give you plausible-sounding but outdated or slightly wrong prices for named competitors, so always check the top two or three it mentions directly on their own websites before you use those numbers to justify your own.
Why does ChatGPT tend to suggest prices that are too low?
Because its answers reflect the average of everything it has seen, and average pricing tends to sit below what a skilled or well-positioned provider could charge. Use its suggestions as a floor to check you're not badly underpriced, not as a ceiling on what you're allowed to ask for.
