Straight answer: most AI tools give pricing advice based on scraped freelancer marketplace data, which means they default to bargain-bin numbers unless you feed them your own results first. If you are using ChatGPT, Claude or Gemini to work out what to charge, you are probably being nudged toward undercharging, and the fix is a five-minute prompt change, not a new tool.
Worth reading next: Is an AI Assistant Better Than a Chatbot for Business Tasks?.
The £500 email that stopped me in my tracks
A client of mine, a marketing consultant in Leeds who has been running her own show for about three years, asked ChatGPT to help her price a new package: a monthly social content and strategy service for small ecommerce brands. She fed it the deliverables, the hours, roughly what a "reasonable" freelancer might charge.
It came back with £500 a month.
She sent it to me before pitching it, half convinced. I told her flatly not to send that number anywhere near a client. For that scope of work, hours included, strategy calls, content calendar, reporting, £500 a month in the UK is what someone charges when they're new, scared, or both. I've seen consultants doing a fraction of that work charge £1,500 to £2,500 a month and get paid without a fight.
That gap between what the AI suggested and what the market pays isn't a rounding error. It's structural, and it's worth understanding before you let a chatbot anywhere near your rate card.
Why AI defaults to bargain-bin numbers
Large language models learn from what's on the internet, and a huge chunk of publicly available pricing text comes from places like freelance marketplaces, forum threads where people ask "am I charging too much", and old blog posts written for people just starting out. That corpus skews low. It's full of people racing each other to the bottom, not people who've built a six-figure consultancy and stopped apologising for their rates.
So when you ask an AI model "what should I charge for X", it's not reasoning from first principles about value delivered. It's pattern-matching against a pile of text where most of the visible numbers are the cautious, competitive, entry-level ones. Confident, expensive pricing tends to live behind closed doors, in proposals and contracts nobody publishes. The model never sees it, so it can't suggest it.
This is the uncomfortable bit nobody selling AI tools wants to say out loud: the tool isn't wrong because it's broken, it's wrong because the data it learned from is full of people who were also underpricing themselves. Ask it to check your Instagram management rate and it'll happily anchor you to Fiverr logic. Feed it your own outcomes first and the answer changes completely.
What changes the answer
I tested this with three different consultants over the past couple of months. Same question, two different ways of asking it.
First version: "What should I charge for social media management for a small business?" Every single time, across ChatGPT, Claude and Gemini, the answer landed somewhere between £300 and £700 a month.
Second version: I gave it context first. Specific results from past work, specific client size, specific time saved or revenue generated, then asked the pricing question. The numbers jumped straight to £1,200 to £2,000 a month, sometimes higher, with the model itself explaining why the value justified it.
Same tool. Same question underneath. Completely different answer, because I stopped asking it to guess and started giving it something to reason from.
Here's the practical version of that, step by step:
- Write down three real results you've delivered for past clients, in numbers where possible (hours saved, leads generated, revenue lifted, cost reduced).
- Tell the AI those results before you ask anything about price.
- Ask it to price based on the value of the outcome, not the hours the task takes.
- Then separately ask it to find you three comparable UK service providers and their published rates, so you're checking its answer against something real, not just accepting the first number.
- Round up, not down. If the AI gives you a range, quote the top of it, not the middle.
That last step matters more than people think. Every consultant I know who's raised their prices in the past year did it in one uncomfortable jump, not a series of gentle 10% increases. Gentle increases just move you slowly through the same low band the AI already anchored you to.
Build the calculator yourself instead of asking the bot every time
There's a better long-term fix than re-prompting an AI every time you quote a new client, and it's one most small business owners overlook completely. Instead of asking a chatbot to price you fresh each time, build your own pricing logic once and put it on your website as a tool your visitors use themselves. I've written before about using interactive calculators to engage and convert visitors, and pricing is one of the strongest uses for it. You set the value logic once, based on your own results and your own floor price, and every visitor who plays with it gets pushed toward a number you've already decided is fair to you. Nobody's chatbot gets a vote.
It also does something psychological that a private chat with an AI can't. A visible calculator on your site signals confidence. You're not hiding your pricing logic behind a "get in touch for a quote" form because you're worried the number will scare people off. You're showing your working, which is exactly what buyers trust.
Brands that never let the market talk them down
It's worth looking at how bigger brands handle this, because the pattern holds at any size. Calm's marketing strategy never competed with free meditation apps on price. It built enough perceived value around sleep and calm that a premium subscription felt obvious rather than optional. GoPro's marketing strategy did the same thing with cameras, selling an identity and a lifestyle rather than specs, which is exactly why people paid £300 to £400 for a camera that competitors matched on paper for less. And Duolingo's marketing strategy shows the flip side, using a free tier to build enormous trust and then converting a slice of that audience to a premium price that most users never blink at, because the value case was built long before the price was ever mentioned.
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.
None of those companies asked an algorithm what the "going rate" was and copied it. They decided what they were worth, built the story around it, and let the market catch up. A one-person consultancy can do the same thing on a smaller scale. The pricing conversation always goes better when the value case comes first and the number comes second.
Automate the admin, not the judgement
Where AI earns its keep in this whole process is the admin around pricing, not the pricing decision itself. Once you've settled on a rate you're confident in, tools like IFTTT's approach to simple automation show the model to copy: connect small, boring, repetitive tasks so they run themselves. A new enquiry lands in your inbox, it triggers a template proposal with your rate already filled in, it logs the lead in a spreadsheet, done. That's the right place for automation. Automating the price itself, letting a chatbot set the number fresh every time based on whatever it scraped that day, is how you end up quietly discounting yourself without ever deciding to.
I keep my own rate card in a single document now, updated maybe twice a year, and I refuse to let any tool re-negotiate it downward mid-conversation with a prospect. AI drafts the proposal. It doesn't set the number.
When to bring in help instead of guessing alone
If you're running a small business and you keep finding that every pricing conversation, whether it's with a chatbot or with yourself at 11pm, ends up landing lower than you meant it to, that's usually a sign the problem isn't the tool. It's that nobody's ever sat with you and worked out what your time and results are worth against the current market. That's a fairly quick fix with the right outside eyes on it, and it's one of the more common things I get asked about when people look into what an AI consultant costs and whether the spend pays for itself. In most cases I've seen, fixing the pricing conversation alone covers the cost of that advice within the first new client you sign at the corrected rate.
The wider point stands regardless of who you ask, though. AI is a brilliant drafting partner and a useful research assistant. It is not a confident business owner who knows what your work is worth. That part still has to come from you, and no amount of clever prompting replaces the decision to simply charge what you're worth and hold the line.
Related reading: 4 Ways That Your Business Can Reduce The Time It Spends On Admin.
Frequently asked questions
Is AI pricing advice always too low?
Not always, but it defaults low far more often than it defaults high, because it's trained on a lot of publicly visible freelance and marketplace pricing, which skews toward cautious, competitive numbers. Feed it your own results and outcomes first and the suggested price usually jumps significantly.
What's the fastest way to check if an AI-suggested price is fair?
Ask the same tool to find three UK competitors offering a similar service and show you their published rates, then compare that against its original suggestion. In most cases the gap alone tells you the first number was too cautious.
Should small businesses use a pricing calculator instead of quoting individually?
A calculator built on your own value logic works well for standard packages and helps convert website visitors without a back-and-forth quote process. For bespoke or high-value work, a direct conversation still tends to close better, because pricing conversations are also value conversations.
Is it worth paying an AI consultant just to help with pricing?
If underpricing is costing you more than one client's worth of revenue a year, which it usually is, then a short pricing and positioning session with someone who knows the market usually pays for itself within the first repriced project.
Related reading: Why Your AI Meeting Notetaker Might Be Breaking the Law (And Killing Your Sales Calls) and AI Chatbots Are Answering Your Customers Before They Reach Your Website.
If you want the full breakdown, here is everything I know about AI marketing.
Related: ai notetaker bots client calls small business.
Prefer to hand this over? This is write for us about small business.
