In a significant decision for software companies in the UK, the First-Tier Tribunal has ruled that Get Onbord Limited’s development of an AI-enabled “know your customer” (KYC) system constituted qualifying research and development (R&D) activities under UK tax law.
Get Onbord Limited (GOL) was an innovative fintech startup that sought to develop a new AI system for automating the KYC verification and risk profiling process for financial services firms. Their goal was to create an AI-enabled holistic analysis system that could achieve superior results compared to manual human analysis while meeting all regulatory requirements.
GOL claimed R&D tax credits for the costs of their project, but HMRC initially rejected the claim, arguing that GOL’s work did not advance overall scientific or technological knowledge and capability as required under the Guidelines.
However, the tribunal firmly rejected HMRC’s arguments after careful examination of the technical evidence. This decision contradicts the narrow interpretation often taken by HMRC caseworkers who have denied R&D claims for software projects that combine existing AI components into advanced integrated systems.
GOL’s expert witness, Mr Edward Cahill, provided compelling testimony about his company’s R&D project. According to the witness, GOL’s AI platform performs the complex identity verification, risk profiling and customer due diligence processes that banks and financial firms have traditionally relied on teams of human analysts to conduct. He explained that rather than employ large teams to manually review data from multiple sources, GOL has developed algorithms that can quickly ingest and process that information to make determinations.
A key innovation is the use of intelligent online forms that validate user inputs in real-time by automatically cross-checking against external data sources. Essentially, it validates what someone is writing as they complete the form. This allows catching issues or inconsistencies immediately during the data entry process, rather than after lengthy manual reviews.
While GOL utilised some existing open-source technologies like the ChatGPT language model, Cahill emphasised their core innovation lies in combining such AI components in a novel way to achieve new capabilities. He described the project as creating new knowledge and capability using open-source material, where the element of innovation is putting all the existing tools and information together and writing the code that links and integrates the component parts.
In responding to scepticism that GOL was simply doing routine data analysis, Cahill stressed that one could say that about all sorts of artificial intelligence. At the core of what they are doing is a lot of data, but they are using it to try to provide an intelligent answer to a question.
HMRC argued that GOL placed too much weight on the presence of an appreciable improvement, but the simple test is whether the process advanced overall knowledge in the field of technology.
HMRC contended the core “problem” GOL aimed to solve related to know-your-customer (KYC) and anti-money laundering (AML) compliance processes, not an underlying technological advancement. They claimed GOL merely automated an existing manual process using conventional technologies and artificial intelligence tools. HMRC asserted GOL needed to show their work represented “an advance in the underlying science or technology to be able to achieve it.”
HMRC also criticised GOL for not clearly explaining the technological uncertainties they were attempting to resolve, stating GOL had not explained the technological uncertainty they were seeking to overcome in clear language, having instead chosen to focus on the functionality and uniqueness of their product. HMRC suggested that because all software has unique functionality and coding, accepting GOL’s project as R&D would extend relief to all coding activity, which is clearly not what was intended.
The Tribunal found that GOL faced significant technological uncertainties in developing algorithms for tasks like data normalisation, integration of diverse data sources, speeding up processes, and building robust APIs. The evidence showed these uncertainties could not be readily resolved by a competent professional in the field. While GOL utilised some existing technologies like APIs and open-source code libraries (which is common practice), the Tribunal accepted that GOL wrote massive amounts of new code to integrate these components into a novel AI system with superior capabilities not available in existing products or processes.
Based on the evidence, including testimony from GOL’s expert witness deemed a “competent professional”, the Tribunal concluded that GOL had carried its burden of proving the project met the R&D criteria under the Guidelines. GOL’s appeal was allowed, establishing its entitlement to the claimed R&D tax credits.
The decision provides an instructive counter-example to HMRC’s often overly-restrictive stance toward software R&D claims related to AI and emerging technologies. The ruling reflects the principle that integrating existing technologies in an innovative way to create new capabilities can constitute an advance in overall technological knowledge.
The decision is likely to be welcomed by the UK tech community as supporting tax incentives for cutting-edge work in emerging fields like AI.