AI is Not Free

Contents

2026/27 Payroll Legislation Guide

Payroll Legislation Guide 2627

The facts, figures, thresholds and allowances for 2026/27, in one handy guide.

OpenAI is not just an AI company. It’s now an investor in specific industries with the implication being that OpenAI believes it can entirely restructure how those industries work. 

Our view is simple. AI is not replacing skilled people in finance and HR. Investors are buying firms and using AI to make those people more valuable. That value has a real cost in development and tokens. Used well, it is worth paying for. 

OpenAI took an ownership stake in Thrive Holdings in December 2025. Thrive recently raised $2 billion and has pulled together more than 50 accounting firms and 2,000 professionals. According to Thrive, its AI tax agents have processed more than 7,000 returns at 98% accuracy and cut preparation time by more than 30%. 

It is not alone: 

Is AI replacing accountants?

For two years, the story was that AI would replace accountants. However, these movements in the market suggest the opposite is true: investors are buying accountancy firms and using AI to make them more valuable, not less. 

The implication is that AI multiplies the value of skilled people rather than replacing them. But it comes with a catch. 

AI is not free

It takes development and it takes tokens. Every natural-language query and every automated task has a cost. That is true for us as a provider, and it shapes what we build and how. 

Our CTO, James Rowell, built Capture Expense as an AI-native product, not an AI-enabled one. An employee talks to the platform in plain language. AI processes the expense. An administrator can request a report in plain language from inside Teams, and what used to take hours comes back on request. AI runs through the product rather than sitting on top of it. 

James is now bringing the same approach into Cintra People, our combined HR and payroll platform. We’re building a conversational AI interface that allows employees and administrators to work with their payroll and people data in natural language, turning complex systems and workflows into something anyone can simply talk to. 

This does not make the product cheaper. It costs us tokens and development time. What it does is make the work better, faster, and more accurate, and it frees your most valuable resource, your people, to spend their time creating value rather than processing it. 

How many UK SMEs are using AI?

This is already happening. According to the British Chambers of Commerce, 54% of UK firms were actively using AI in 2026 (around 94% of those surveyed were SMEs). That is up from 35% in 2025 and 25% in 2024. So far, it is not costing jobs: 95% of AI-using SMEs report no change to their workforce over the past year. The point is not replacement. It is human-in-the-loop work, with AI handling volume and people handling judgement. 

Adoption is uneven. Larger SMEs and B2B professional services firms lead. Smaller firms, consumer-facing businesses, and manufacturers move more slowly. Some enterprises invest early; others wait. Some SMEs hold back through caution; others use their size to move quickly. 

Microsoft and WPI Strategy estimate that generative AI adoption could add up to £78 billion to the UK economy through its SMEs over the next decade. When that report was published in 2025, fewer than one in five SMEs had adopted AI. It is a projection, not a measured result, but it suggests many SMEs could be missing out on value by being slower to change. 

The OpenAI and Thrive deal is simply the clearest evidence yet of the value the market sees in applying AI to finance and human-capital work. 

We are focused on SME and mid-market. Our job is to guide our customers through this transition, to demystify it, and to provide a service that is AI-native from the ground up. AI is not free. Used well, it is worth paying for. 

Frequently asked questions

A: The market suggests the opposite. Investors are buying accountancy firms and using AI to make them more valuable. AI multiplies the value of skilled people rather than replacing them. 

A: Every natural-language query and every automated task has a cost in tokens. Building AI into a product also takes development time. 

A: An AI-native product has AI running through it rather than sitting on top of it. In Capture Expense an employee talks to the platform in plain language and AI processes the expense. 

A: The British Chambers of Commerce found that 54% of UK firms were actively using AI in 2026 compared with 25% in 2024. Among SMEs using AI 95% reported no change to their workforce size over the past year.