UK software firms struggle to bill AI products accurately

Research published by PwC UK and m3ter shows many UK software companies are struggling to bill accurately for AI products and services, with 63% of executives reporting a lack of full confidence in their billing operations.

The findings point to a widening gap between the rapid adoption of AI pricing models and the finance systems needed to convert them into booked revenue. Among 350 UK software executives surveyed, 44% said they struggle to capture and measure customer usage in pay-per-use models, while 62% expect some exposure to revenue leakage.

Usage-based pricing is becoming more common as software companies seek to charge for AI tools by consumption rather than through fixed subscriptions alone. Just over a third of respondents said they had introduced pay-per-use pricing alongside traditional charging for AI, and half had changed pricing at least twice in the past year.

Many companies have introduced AI through premium products or by adding it to standard products and raising prices. That has increased pressure on billing, finance and commercial systems, which were often built for simpler recurring-revenue models.

System Gaps

The research found that 87% of respondents reported a lack of integration between billing platforms and ERP or general ledger systems. Another 48% said billing and customer relationship management systems were not integrated.

These gaps can force finance and operations teams to rely on spreadsheets and manual processes to set prices, track usage, and reconcile revenue. In practice, that increases the risk that billed amounts do not fully reflect product consumption, especially when contracts include multiple pricing tiers, credits, discounts or usage thresholds.

The study defines revenue leakage as unrealised value in products and services. This includes usage above contracted limits that is not captured or monetised, pricing clauses that are not properly reflected in billing workflows, and discounts or credits that are applied incorrectly or missed altogether.

Revenue leakage is estimated at between 4% and 7% of annual recurring revenue. Concern is higher among companies that have already adopted usage-based pricing, with 72% saying they lack confidence in their exposure to leakage.

This commercial challenge comes as companies continue to seek returns on their AI investments. PwC's 29th Global CEO Survey found that 30% of respondents saw revenue increase from AI in the past year, suggesting many businesses are still working out how to turn AI development into dependable income.

Weak infrastructure is also contributing to billing disputes and compliance concerns. Respondents cited reduced customer trust, less room to experiment with pricing, and greater concern about audit and regulatory scrutiny when billing and finance processes fail to keep pace.

"AI is transforming what companies sell but outdated billing and finance means they aren't getting paid properly. For businesses to capitalise on AI, they need the operational infrastructure to accurately measure, track and recoup income. Companies that lack such infrastructure won't realise the full value of the products and services they sell," said Griffin Parry, Chief Executive Officer and Co-Founder of m3ter.

Pricing Pressure

The pace of change in software charging models is adding to the strain. AI products often involve variable usage, feature-based charging or hybrid structures that combine subscriptions with pay-per-use elements. Each adjustment can create new data, contract and accounting requirements, particularly when sales, finance and product teams move at different speeds.

That makes accurate metering and billing more important for software groups trying to protect margins while refining how they charge for AI. Without stronger links between product usage data and the quote-to-cash process, companies risk undercharging, overcharging or delaying invoicing.

The results indicate a need to modernise core finance and billing systems as pricing models grow more complex. PwC UK and m3ter have been working together since 2025 on efforts to reduce commercial risk and improve revenue integrity in software businesses.

For software providers, the issue is no longer limited to product design or market demand. The survey suggests that billing, reconciliation and financial control are becoming direct factors in whether AI revenue is actually realised.

"Packaging and pricing are changing faster than ever with AI, making it harder to ensure every penny is captured. Modern pricing depends on accurate billing but fragmented systems are widening the gap between pricing strategy and realised revenue. Without stronger foundations, software companies risk amplifying existing revenue leakage as they move toward more complex AI-driven pricing models," said Jonny Donnelly, PwC, UK.