Get Onbord: What Proving An R&D Advance Actually Takes

The refusal letter probably says your project "was not an advance in science or technology". It may add that the solution was readily deducible, or that no competent professional has been identified. And it was quite possibly written by an officer who has never seen your code.

In 2024, a company beat exactly that refusal from the weakest position imaginable. Get Onbord Ltd was in liquidation, and its case was argued by its former director, who holds no computer science qualification. The tribunal found for the company—not on a technicality, but because the evidence showed a genuine technological advance.

This analysis covers what counted as proof, who counts as a competent professional, how the burden really works, and why an almost identical claim failed two days later. For the rest of the battlefield—working out which letter you are holding, the claim notification and additional information forms that can invalidate a claim on their own, and any penalty exposure—start with our R&D tax credit appeals guide.

What The Tribunal Actually Had To Decide

Get Onbord Ltd (in liquidation) v HMRC [2024] UKFTT 617 (TC) was heard by Tribunal Judge Mark Baldwin and Mr Mohammed Farooq at a remote video hearing on 3 January 2024, and the decision was released on 9 July 2024. Get Onbord ("GOL") had claimed a payable R&D tax credit for a novel, automated artificial intelligence process for "know your client" (KYC) verification and risk profiling—the customer checks banks were still running manually, with large teams of people. HMRC refused: on its analysis, the project did not advance overall knowledge or capability, so was not research and development at all.

Neither side brought a barrister. GOL was represented by Barrie Dowsett of Myriad Associates (the consultancy that prepared the claim) and Edward Cahill, its former director, who argued the technical case himself; HMRC by David Lewis, one of its own litigators. The appeal was allowed—and for anyone facing the tribunal unrepresented, that combination is the most encouraging fact in the case.

The tribunal walked the statutory chain once, quickly. A payable credit under section 1054 CTA 2009 requires a "Chapter 2 surrenderable loss" (s.1055(2)(a)), which requires the additional deduction in s.1044, which requires "qualifying Chapter 2 expenditure" (s.1051). Condition B in section 1052(3), in the version the tribunal applied, required that expenditure to be "attributable to relevant research and development undertaken by the company" (since substituted for merged-scheme periods, so the current text reads differently).

"Research and development" takes its meaning from s.1041 CTA 2009 and s.1138 CTA 2010, which defer to Treasury regulations under s.1006 ITA 2007—and for the periods in this case, regulation 2 of SI 2004/712 gave the force of law to the guidelines on the meaning of R&D for tax purposes (the "Guidelines"). SI 2023/293 has since done the same for the current 2023 version, and every Guidelines paragraph cited in this article is numbered identically in both.

Every link in that chain collapsed into one question: did GOL's project seek an advance in overall knowledge or capability through the resolution of scientific or technological uncertainty? "Overall" is doing real work there: the advance must be to the knowledge or capability of the field, not just your company's own (Guidelines, paragraph 6). "We had never built anything like this before" is not, by itself, an advance.

Disputes like this travel through the enquiry and closure notice machinery in Schedule 18 FA 1998—see our corporation tax appeals and enquiries and closure notices guides.

Which regime are you in? Get Onbord was decided under the pre-merger SME scheme. For accounting periods beginning on or after 1 April 2024, Schedule 1 to the Finance Act 2024 (commenced by SI 2024/286) replaced the two old schemes—the SME scheme and the large-company RDEC—with a single merged scheme plus enhanced relief for R&D-intensive companies (ERIS). The definition of R&D is unchanged, so everything this case says about advances, competent professionals and evidence applies identically to merged-scheme claims—and pre-merger periods will keep reaching the tribunal for years. Not sure which regime your claim sits in? The R&D hub decodes it.

Who Counts As A Competent Professional

Under the Guidelines, technological uncertainty exists when knowledge of whether something is feasible, or how to achieve it in practice, is not readily available or deducible by a "competent professional working in the field" (paragraphs 13-14). So who your competent professional is—and how their evidence lands—can carry the whole claim.

Mr Cahill's paper credentials were unpromising: a University College Dublin degree in International Commerce and French, and no software qualification. He accepted in cross-examination (live questioning by HMRC's representative) that he was not a software developer, but said he writes code because it is often faster to do it himself. The tribunal's assessment, at [82], is the passage practitioners now quote:

"Although Mr Cahill does not have any formal qualifications in this area (nor, we note in passing, does Sam Altman, who dropped out of his computer science course at Stanford), Mr Cahill was a very impressive witness, who spoke with complete fluency about the technical way in which GOL's project worked and answered Mr Lewis' questions with assurance. ... we are completely satisfied that Mr Cahill has experience (including in coding) and up-to-date knowledge of software capabilities, albeit perhaps only in the area he works in, to be a 'competent professional' for our purposes."

The contrast is Flame Tree Publishing Ltd v HMRC [2024] UKFTT 349 (TC), decided earlier the same year. There, a publisher with some familiarity with computing and a computer user who was not a programming or software professional were held not to be competent professionals, and the claim failed. HMRC's own manual, CIRD81300, takes the same line: an intelligent interest in a field does not by itself make someone a competent professional.

The test, in short, is not certificates but demonstrated experience, hands-on capability and up-to-date knowledge of the field—proved live, under cross-examination. A tribunal applies the Guidelines, not HMRC's gloss, but nothing in Get Onbord suggests a job title will do.

One question was left open. Flame Tree had accepted HMRC's submission that a claim needs evidence from a competent professional. HMRC did not run that argument here, and the tribunal, doubting the concession, declined to decide the point: "it comes as a relief not to have to express a conclusion on this issue" ([84]). It elsewhere called such evidence "clearly highly desirable, whether or not it is strictly necessary". Treat it as essential.

Using Existing Technology Can Still Be An Advance

HMRC's core objection was that GOL's platform was assembled from existing technologies, so any advance was routine or readily deducible. The tribunal disagreed.

GOL had a "project"—"a number of activities conducted to a method or plan in order to achieve an advance in science or technology" (Guidelines, paragraph 19)—an overall process with a defined objective. That the work evolved over time did not mean there was no method or plan.

The tribunal recorded Mr Cahill's evidence that a massive amount of new code was written, going "far beyond 'routine' adaptation of existing technologies" and establishing "a capability that was not there before" ([92]). It accepted that a new function solving a real-world problem in a new and creative way "is at least an indication that there has been an appreciable technological advance"—on Mr Cahill's evidence, no bank had developed a system like this that he knew of, the market solution being to "throw bodies at the task" ([93]). And on the use of open source and other existing components ([94]):

"We do not consider that, of itself, the use of open source, or other existing, materials is an indication that a particular development is routine or readily discernible from publicly available materials. ... it is not necessary for that to be the case that each component part of the solution must itself be novel or bespoke to the project in question. As Mr Cahill observed, given the amount of open-source software/AI material available, if complete novelty were the test, no software project would ever amount to R&D."

That is the heart of the decision, and the Guidelines point the same way: an appreciable improvement to an existing process can be an advance (paragraph 9(c)), and combining standard technologies can be R&D where a competent professional cannot readily deduce how they should be combined (paragraph 30). A project that fails can still be R&D (paragraph 10): what matters is the uncertainty you tackled, not whether the venture succeeded commercially.

The tribunal's conclusion, at [95], doubles as a checklist:

"we are satisfied, on the balance of probabilities, that: (1) GOL had a project with a defined aim; (2) the technology GOL sought to develop and incorporate in the project was not already publicly available or readily deducible; (3) the technology it sought to develop to achieve the project's aims amounted to more than 'routine' copying or adaptation of an existing product or process; and (4) the project required the resolution of technological uncertainties which a competent professional working in the field could not have readily resolved."

Those four findings are the four things your evidence has to establish. A refusal letter asserting the opposite of any of them is telling you where the fight will be.

Who Has To Prove What, And When It Starts To Shift

Be clear about this first: the legal burden of proof was on the company throughout. The tribunal agreed with HMRC that it was for GOL "to substantiate its entitlement, not for HMRC to justify its decision to deny the credits claimed" ([74]).

But it then discussed what lawyers call the shifting evidential burden, drawing on the Court of Appeal's decision in Wood v Holden [2006] EWCA Civ 26, which held it wrong to rest a decision on the burden of proof where the state of the evidence points the other way.

Applied here ([77]): there may come a point where the company can say, "We have done enough to raise a case that our project comprised an overall advance in science and technology. What more can the Tribunal expect from us? The burden must now pass to the Revenue to produce some material to show that, despite what appears from everything we have produced, our project was a routine advance."

The honest footnote is at [96]: the tribunal said it had reached that point, but did not decide the case that way—it simply asked whether, on all the evidence, it was "more likely than not" that the four statements in [95] were correct, and answered yes. The burden shift is guidance, not the reasoning that decided the case, and not a tactic that wins an appeal by itself.

A credible, detailed technical case does not win by default, but it forces HMRC to respond with something better than assertion. In Get Onbord, it had nothing better.

Why HMRC's Evidence Did Not Help HMRC

HMRC's only witness was the caseworker, Mr Umar. He had no technology experience or expertise, did not know the credentials of the in-house software team members whose advice he relied on, and this was the first software claim he had dealt with ([64]). The tribunal found him honest and straightforward—but "his lack of scientific knowledge or experience meant that his evidence was of no real help to us" ([65]). HMRC called no expert evidence at all.

The failure was institutional, not personal: since 2023 HMRC has been running a volume compliance approach to R&D claims in which caseworkers assess technical claims at scale. Get Onbord shows what happens when a decision built that way meets a fluent expert in a hearing room.

The practical move for your own dispute: ask HMRC, in writing and early, what technical evidence supports the assertion that your advance was routine or readily deducible, and who provided it. The answer—or its absence—tells you what case you actually have to meet.

The Cautionary Twin: Tills Plus

Two days after Get Onbord, the First-tier Tribunal released Tills Plus Ltd v HMRC [2024] UKFTT 614 (TC) (Judge Robin Vos). Same genre: an AI/software claim under the SME scheme, prepared with consultants—and this appellant even had a barrister. It lost.

What makes the pairing instructive is how much Tills Plus had going for it. The expert, Dr Zade, was accepted as a competent professional on his CV ([141]), and his report was accepted as describing a project which, if accurate, would constitute R&D ([142]-[148]). The claim still failed.

Dr Zade never gave evidence and could not be cross-examined, which limited the weight his report could carry ([146]). And the report did not describe the work actually done: the enquiry explanations, the services agreement and the invoices all described a more mundane project—a "virtual hospitality manager" built by combining existing modules ([149]-[160]). On that real project the tribunal found it was more likely a "routine adaptation of existing products or services" ([164]), with "no evidence from a competent professional as to whether there would in fact have been any uncertainties in combining the various modules" ([166]). The appeal was dismissed on the R&D issue.

The company whose technical story was consistent, and whose expert turned up and answered questions, won; the company whose polished report contradicted its own contemporaneous record, and whose expert never testified, lost. Remember, too, that First-tier Tribunal decisions do not bind other tribunals: Get Onbord is persuasive, not a guarantee, and Tills Plus is the proof.

Putting Your Scientific Cards Face Up On The Table

The tribunal closed with advice for both sides of every future R&D dispute ([99]):

"We consider that these proceedings would have been much more straightforward (and possibly could have been avoided) if, at an early stage, both parties had 'put their scientific cards face up on the table'. Ideally, GOL would have produced a single document in which it marshalled all its scientific/technological evidence, including evidence from a competent professional (which is clearly highly desirable, whether or not it is strictly necessary), and HMRC would then have replied to that document with details of its own scientific analysis and evidence."

That is a playbook you can follow now. One document, structured project by project: who the competent professional is and why (experience, hands-on capability, current knowledge); what the baseline of publicly available knowledge was; what uncertainties a competent professional could not readily resolve; and how you set about resolving them. HMRC's own GfC3 guidelines for compliance ask for essentially the same things.

Before you write a word of it, do what Tills Plus never did: read the claim report side by side with your contemporaneous record—the enquiry correspondence, the contracts and invoices, the commit history, the design notes. If they tell different stories, resolve that before HMRC finds the contradiction. Then decide who can stand up and answer questions on it all.

Do not mistake this for a lowered bar. The tribunal in Hadee Engineering Co Ltd v HMRC [2020] UKFTT 497 (TC) expected the company to identify the uncertainties and the advance project by project (though no particular form of record is required), and Tills Plus applied the same standard. Get Onbord cleared that bar with live, cross-examinable evidence.

If you appeal, frame your grounds of appeal around the four findings at [95] and prepare the person who can prove them to give evidence—our hearing preparation guide covers witness statements and bundles.

One caution: the tribunal decided Get Onbord on the evidence in front of it. Whether the same reasoning helps you depends on your own project and what you can prove.

Before You Commit To The Tribunal

Still mid-enquiry, with no closure notice and no end in sight? You are not powerless: the company can apply to the tribunal under paragraph 33A of Schedule 18 for a direction requiring HMRC to close the enquiry—see our corporation tax appeals guide.

Once the closure notice amendment or assessment lands, the deadline comes first: 30 days to give notice of appeal (paragraphs 34(4) and 48(2) of Schedule 18). For corporation tax the notice goes to HMRC, not the tribunal—you then notify the appeal to the tribunal when you want it decided; our step-by-step appeal guide covers the mechanics. Miss the deadline and you are into discretionary late appeal territory. And do not let it pass in silence: an unappealed closure notice becomes final, the tax and any clawed-back credit collectible whatever the technical merits.

Before the tribunal, you can ask for a free statutory review by an officer not previously involved—HMRC then has 45 days to complete it, and you have 30 days from its conclusion to notify the tribunal. Be realistic: review is better at discrete legal or procedural errors than at a fact-heavy Guidelines fight—which is what a Get Onbord-shaped case is. Nor does appealing mean fighting to the end: you can settle at any stage, including through HMRC's alternative dispute resolution (ADR) scheme.

Costs should not be what stops you. There is no fee to appeal, and the First-tier Tribunal default is that each side bears its own costs. The exceptions: a Complex-category allocation (you can opt out of costs-shifting within 28 days) and awards for unreasonable conduct—see our tracks and costs guide.

R&D appeals are usually Standard or Complex, and the no-costs default cuts both ways: what you spend on advisers or an expert is not recoverable even if you win. That is a trade-off to make knowingly—though Get Onbord shows a well-prepared company can manage without a barrister.

If HMRC has paid a credit and wants it back, appealing does not by itself suspend collection—you can apply to postpone the disputed tax while the appeal runs. Interest accrues at 7.75% on tax that turns out to be due, from its original due date, whatever the appeal's outcome—see interest on unpaid tax. And if a Schedule 24 penalty rides on the refused claim, our reducing HMRC penalties guide covers the behaviour and disclosure arguments.

Budget for time as well as money: tribunal cases take typically 6-12 months from appeal to decision, and evidence-heavy R&D disputes can run longer—Get Onbord's tribunal reference dates from 2022 (TC/2022/13281), and the decision arrived in July 2024.

Finally, aim to win first time. Whether a project is R&D is an evaluative conclusion drawn from found facts, so an onward appeal needs an error of law in the Edwards v Bairstow sense, not a different view of the evidence. Get Onbord itself has no reported onward appeal (checked 5 August 2026). Permission to appeal to the Upper Tribunal has to be applied for within 56 days of the tribunal's full written reasons, a right the tribunal noted at [100].

A Footnote On Liquidation

The last feature of the case is a warning, not a precedent. GOL went into liquidation in October 2023; Mr Cahill conducted the January 2024 hearing without the Joint Liquidators' knowledge; HMRC applied for a re-hearing. The tribunal refused: the Joint Liquidators had since endorsed his conduct of the appeal, had power to do so under Schedule 4, paragraphs 4 and 13, of the Insolvency Act 1986, and a re-hearing would have served nobody.

But the tribunal stressed how exceptional this was: "We would be very surprised if this had happened in any other case; if it has, it must surely be very rare" ([15]). Do not plan around that indulgence—if your company is in liquidation, get the liquidators' written authority before taking any step in an appeal.

HMRC also raised the going-concern condition in section 1057 CTA 2009, then withdrew it: the credit had already been set against GOL's unpaid PAYE in June 2021, so section 1057(3) applied ([68]). (The section is omitted for merged-scheme periods.) None of this means an insolvent company is generally free to claim—the position is technical, and needs advice.

Key Legislation And Resources

The Judgment

• Get Onbord Ltd (in liquidation) v HMRC [2024] UKFTT 617 (TC) — First-tier Tribunal (Tax Chamber), Tribunal Judge Mark Baldwin and Mr Mohammed Farooq; appeal reference TC/2022/13281; heard 3 January 2024, decision released 9 July 2024

Legislation

• Section 1054 CTA 2009 — entitlement to the payable R&D tax credit on a Chapter 2 surrenderable loss
• Section 1055 CTA 2009 — Chapter 2 surrenderable loss
• Section 1044 CTA 2009 — the additional deduction for qualifying Chapter 2 expenditure
• Section 1051 CTA 2009 — qualifying Chapter 2 expenditure
• Section 1052 CTA 2009 — Condition B, the provision the case turned on (since substituted for merged-scheme periods)
• Section 1041 CTA 2009 — "research and development" takes its meaning from s.1138 CTA 2010
• Section 1138 CTA 2010 — the definition, subject to regulations under s.1006 ITA 2007
• Section 1006 ITA 2007 — Treasury power to specify activities by reference to guidelines
• SI 2004/712, regulation 2 — gave the Guidelines the force of law for the periods in this case (replaced by SI 2023/293 from 2023)
• Schedule 18 FA 1998 — enquiries (para 24), closure notices (para 32), appeals (paras 34 and 48)
• Schedule 1 FA 2024 and SI 2024/286 — the merged scheme, for accounting periods beginning on or after 1 April 2024

Key Cases

• Get Onbord Ltd (in liquidation) v HMRC [2024] UKFTT 617 (TC) — a director with no formal qualifications accepted as the competent professional; novel integration of existing technology held an advance; appeal allowed
• Tills Plus Ltd v HMRC [2024] UKFTT 614 (TC) — the foil, released two days later: accepted expert report contradicted by the contemporaneous record, expert never testified, appeal dismissed on the R&D issue
• Flame Tree Publishing Ltd v HMRC [2024] UKFTT 349 (TC) — witnesses with familiarity but no professional standing held not competent professionals; the baseline Get Onbord distinguishes
• Hadee Engineering Co Ltd v HMRC [2020] UKFTT 497 (TC) — identify the uncertainties and the advance project by project, though no particular form of record is required; applied in Tills Plus
• Wood v Holden [2006] EWCA Civ 26 — the Court of Appeal authority on the shifting evidential burden that Get Onbord applied

HMRC Guidance

• Guidelines on the meaning of R&D for tax purposes — the DSIT Guidelines with force of law: advances (paras 6-9), uncertainty and the competent professional (paras 13-14), projects (para 19), combining standard technologies (para 30)
• CIRD81300 — competent professional — HMRC's gloss on who qualifies
• CIRD81960 — software projects — HMRC's approach to R&D claims for software
• GfC3: help to see if your work qualifies as R&D — HMRC's best-practice expectations: the competent professional, the baseline, the uncertainties
• HMRC's compliance approach to R&D tax reliefs — the volume-compliance context behind the refusal letters

On This Site

• R&D tax credit appeals — the hub: scheme mechanics, the Schedule 18 route, penalties, and what to do when the consultancy goes quiet
• Corporation tax appeals — the Schedule 18 FA 1998 machinery in depth
• HMRC enquiries and closure notices — the enquiry lifecycle your dispute is inside
• Writing grounds of appeal — structuring grounds around the four [95] findings
• Preparing for your tax tribunal hearing — witness statements, bundles, and getting your competent professional ready to testify
• HMRC internal review — the free statutory review before you commit to the tribunal
• Settling your tax tribunal case — settlement agreements and HMRC's ADR scheme, available at any stage
• Tribunal tracks and costs — the no-costs default, the Complex category and the 28-day opt-out
• Postponing payment during appeal — whether the disputed amount must be paid while the case runs
• Interest on unpaid tax — what accrues on a clawed-back credit whatever the outcome
• How to appeal to the tax tribunal — the filing mechanics, step by step
• Late appeal to the tax tribunal — if the 30 days have already passed
• Edwards v Bairstow: error of law — why evidence-based findings like these are hard to disturb on appeal
• Upper Tribunal appeal — the onward route; permission must be sought within 56 days
• Reducing HMRC penalties — the Schedule 24 penalty that can ride on a refused claim
• After your tax tribunal decision — what the release of a decision means for both sides
• Tax dispute timeline — where an R&D enquiry sits in the wider journey

This article is for informational purposes only and does not constitute legal or tax advice. For advice specific to your situation, consult a qualified tax adviser, accountant, or solicitor.