The rumour that won’t stop whispering
You’ve probably heard it over coffee, in a board meeting or scrolling LinkedIn on a Monday morning: the AI bubble is about to burst.
It’s a bold claim about a technology that has slipped into almost every corner of working life in under three years. So is there anything in it? And more importantly, what would it mean for your business if there was?
What does “the AI bubble bursting” actually mean?
A bubble forms when money pours into something faster than that thing can earn it back. Valuations climb on expectation rather than results, and investors keep buying in because everyone else is. Eventually reality catches up, the money slows, and prices correct, often sharply.
We’ve been here before. When the dotcom bubble burst in 2000, the internet didn’t disappear. Hundreds of companies did, though, and the ones that survived had to start charging what their services genuinely cost to run.
The AI theory follows the same pattern. Billions are being poured into data centres, chips and model training, while the tools themselves are sold to us at prices that don’t come close to covering the bill. If investor patience runs out, providers have three options: fold, merge, or pass the true cost on to their customers.
So a burst wouldn’t mean AI vanishes overnight. It would mean the era of cheap, heavily subsidised AI comes to an end, and that’s the part businesses should be paying attention to.
Our perspective: the bill is coming
The numbers behind the rumour are hard to ignore. In the first three months of 2026, OpenAI reportedly burned through $3.7bn against $5.7bn of revenue, with both figures tripling year on year (The Next Web, citing The Information). The company doesn’t expect to turn a profit until the end of the decade.
The reason is simple. Every question you ask ChatGPT, Claude or Perplexity costs real computing power to answer, so the more people use these tools, the more they cost to run. Growth doesn’t automatically bring profit here; it can bring a bigger bill. Today’s monthly subscription is, in many cases, being propped up by investor money.
We’ve seen this play before. Uber spent years subsidising cheap rides, got everyone hooked, then raised its prices once the habit was formed. If AI follows suit, prices could rise sharply and quickly.
That matters because AI is no longer a novelty. It’s composing client emails, running calculations and helping plan the next quarter. Imagine your licence costs doubling across fifty seats overnight. Imagine workflows built around a tool you can no longer justify, and a team that has quietly forgotten how to do the task without it. How would your business manage?
Should we rely on AI, or use it in small doses?
This is the real question the bubble raises. Using a tool and relying on it are two very different things. One gives you an edge; the other hands someone else control of your costs, your processes and your continuity.
The answer isn’t to ditch AI. It’s to be honest about where it genuinely earns its keep, and where it has crept in simply because it was there, it was cheap and it was exciting.
Where we stand at bots for that
AI is a brilliant thinking tool. As a sounding board for ideas, a first draft generator or a way to make sense of a messy problem, it’s hard to beat. But for a lot of everyday business tasks, it isn’t better, and it certainly isn’t cheaper, than good old automation.
Drowning in manual, repetitive work? Tell us the task and we’ll show you what to automate.
We say that as a company born and bred from automation. Since 2012, we’ve been building and running automation for businesses, and we know it like the back of our hands. As AI has unfolded, we’ve watched plenty of firms reach for it to fix problems automation was already solving: data entry, reconciliations, moving information between systems, producing the same report every month. Those tasks follow rules. When the rules don’t change, neither should the cost, and neither should the result.
That doesn’t make us anti AI. We use it ourselves, on the side, where it genuinely adds something. It’s a fantastic enhancer. It just isn’t needed for everything.
Automation vs AI: what’s the difference?
As the rumours swirl, it’s worth being clear on the difference, because the two are constantly confused and lumped together. In short, automation does exactly what you tell it to. AI works out what it thinks you want.
| Automation | AI | |
|---|---|---|
| How it works | Follows fixed rules and steps you define | Predicts an answer from patterns in its training data |
| Output | The same result every time | Can vary from one attempt to the next |
| Best for | Repetitive tasks with clear rules: data entry, reconciliations, reporting | Open questions: drafting, summarising, brainstorming |
| Cost | Predictable and largely fixed once built | Tied to usage and provider pricing, which can change |
| Accuracy | Precise when the rules and data are right | Can sound confident and still be wrong |
| Audit trail | Every step is traceable | Harder to explain how it reached an answer |
| Human oversight | Light once tested | Ongoing review needed |
| If the provider raised prices tomorrow | Little to no impact | Your costs rise with them |
The best results often come from using both: automation doing the heavy lifting reliably in the background, with AI layered on top only where judgement or language genuinely adds value.
Before you jump into a purely AI product
Before signing up to the next shiny tool, ask whether AI is what the problem actually needs. If it is, make sure four things are in place first.
Governance. Know what data is going into the tool, where it ends up and who is accountable for the outputs. For firms handling client financials, this isn’t optional.
A backup plan. If AI were switched off tomorrow, or priced out of reach, could the work still get done? Map the processes that depend on it and decide now what the fallback would be.
Integration. A tool that doesn’t talk to your practice management software, CRM or ledger creates more manual work, not less. Check how it will slot into the way your business really runs.
Training. Your team needs to know when to trust the output, when to check it and when not to use it at all. The tool is only as good as the people steering it.
The bubble may burst. Your business doesn’t have to.
Nobody can say for certain when, or if, the AI bubble will pop. But the economics are clear enough to plan for. The smart move isn’t to stop using AI; it’s to know exactly where you need it, build the rest on something you control, and have a plan ready if the price tag changes.
That’s where we come in. Whether it’s automating the processes that don’t need AI at all, or putting the governance, training and fallback plans in place around the ones that do, bots for that can help you build a setup that holds firm whatever happens to the market.
Not sure where AI is genuinely earning its keep in your business? Get in touch with the team at bots for that and let’s find out together.
