Is AI helping or hindering R&D tax credits?

The Bloomberg Tax analysis on whether soaring tech salaries qualify for R&D tax credits raises an interesting question.

As AI drives remuneration for engineers into the millions, tax authorities are questioning whether those individuals were genuinely conducting qualifying research as opposed to dedicating most of their time to strategy, management and external commitments.

There’s an undeniable tension growing between how AI is actively accelerating innovation in practice, and how that same innovation is currently interpreted by HMRC for tax purposes. If AI is doing more of the work, what exactly counts as R&D?

We find ourselves at the intersection of two realities pulling in opposite directions. On one side, AI is genuinely accelerating innovation by compressing development cycles, unlocking new capabilities and helping ambitious businesses move faster than was previously considered possible.

But across the way, HMRC’s definition of qualifying R&D was not written in the knowledge of what AI’s capabilities would ultimately become.

But here we are, and the gap between how innovation is unfolding in practice and how it is being interpreted for tax purposes is widening. There’s a real financial risk involved, so how can businesses ensure AI continues to help them secure R&D tax credits, rather than slipping into a reality where AI becomes a significant hindrance?

The new pace of innovation

Innovation is the lifeblood of the UK scale-up scene. Investors increasingly recognize that innovation is the driving force behind business performance, and that’s why nine in 10 consider R&D tax relief essential to business success, particularly for maintaining growth momentum and extending operational runway.

AI is a significant part of what is driving that investment, as businesses across sectors continue to integrate AI tools into their development processes. These tools are being used to test hypotheses faster, automate complex workflows and push into technical territory that would previously have required far greater time and resource.

From an innovation standpoint, this is exactly the kind of ambition the R&D relief scheme was introduced to support. The challenge is that the scheme’s eligibility criteria have not kept pace with how that innovation is now being done.

Where AI creates a problem for HMRC claims

HMRC defines R&D as activities that seek to resolve scientific or technological uncertainty, so the scheme was designed to reward genuine experimentation. When AI is deployed to help carry out genuine R&D, such as using it to interpret large data sets or automate processes, no issues arise.

Likewise, if a business develops proprietary AI tools to help create logic systems and algorithms from scratch, HMRC also considers R&D.

However, when businesses start using off-the-shelf AI tools or agents, putting them into their own products and services under the pretense of ‘innovation’, it doesn’t constitute as R&D and so therefore cannot be claimed against.

Against a backdrop of rising enquiry rates, businesses recognize that HMRC checks are detailed investigations into both the technical and financial aspects of a claim. Those that once filed claims with minimal supporting evidence face greater scrutiny, and blurred lines around AI-assisted innovation is another consideration.

Other businesses are underclaiming or choosing not to claim at all due to fear of triggering an enquiry, or because of the significant time and effort involved in preparing a robust submission. Both outcomes carry a real cost to the businesses involved and to the wider innovation economy.

How R&D relief acts as a growth strategy

Founders should continue to treat R&D as a major growth driver. From securing funding to shaping operational priorities, prioritizing R&D is one of the most effective ways for founders to build trust across the board.

Research found that 97% of UK investors are now actively shaping the R&D strategies of the companies they fund, and investors are becoming increasingly diligent in looking at how a company prepares its R&D claim, as poor R&D governance can delay growth or destabilize the runway.

The R&D relief scheme remains one of the most valuable tools available to UK innovators, so transparency is critical for qualifying companies.

For businesses working closely with AI, that means documenting everything and building a technically credible claim that reflects the true ambition of what teams are developing, not just the speed at which AI is helping them develop it.

We've featured the best IT automation software.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro