Choosing a Specialist R&D Tax Adviser

Questions to ask a prospective R&D adviser

A question bank for those early conversations and meetings when you are choosing a suitable R&D adviser, organised by what each question is for, with the answers worth hearing and the answers that should worry you.

Question marks

There is no shortage of question lists for choosing an R&D adviser. Most of them are fine and few of them help much, because the difficulty was never knowing what to ask. It is knowing what a good answer sounds like when you get one, and telling it apart from an answer that is merely confident.

This guide is the question bank, organised by what each question is actually for, with the answers you are listening for and the answers that should worry you. If you are still screening firms out, the warning signs come first. If you have the answers already and need to decide, our future guide on how to compare advisers will help with that.

Before you ask anyone anything

Four questions to answer for yourself. Turning up without these makes every conversation vaguer than it needs to be, and vagueness is the condition in which a weak firm sounds identical to a strong one.

Which of our projects do we think involved technological uncertainty?

Not which were expensive, complex, difficult or new to us. Which involved something your own competent people did not know how to do and could not readily work out. You do not need the right answer, you need a starting position that a firm can push back on.

Who is our competent professional?

The person who led the technical work and meets HMRC's definition, and who can explain what was uncertain and why. If nobody obvious comes to mind, that is worth knowing before you engage anyone.

What are our accounting periods, and have we claimed before?

If you have not claimed recently, there is a notification requirement that falls six months after the end of the period of account, and missing it can potentially end a claim before it starts. Any firm should confirm your position on this in the first conversation, unprompted.

What happened last time?

If you have claimed before, what was submitted, who prepared it and did HMRC ask anything? A new adviser inherits that history whether they mention it or not.

Then check the record on each firm before you meet it. Professional body membership, filing discipline, trading history and adverse media are all public, so those are not questions worth spending a meeting on. Anything you can verify without asking should be verified without asking, because an answer you are given is an answer someone chose to give. We have the record assembled for every adviser we list, and the detail of how our assessment works is in the FAQ.

Questions about the people who would do the work

"Who would actually do this work, and can I meet them?"

Firms do not prepare claims, people do. You are trying to establish whether the person in front of you is the person who would be preparing your claim, or a different function of the same business. This is not a trick question and a good firm answers it in one sentence with names. Hesitation, or a general answer about the strength of the team, is the answer you should note.

"What are they qualified in, and what have they worked on in our field?"

Listen for whether the sector experience is real or adjacent. "We do a lot of manufacturing" is not the same as being able to describe the technological baseline in your part of manufacturing.

"You mention former HMRC people. Who are they?"

The claim is common and rarely tested. Genuine ex-inspectors are a valuable commodity, some work across several firms, and a firm may keep names off its website to stop competitors poaching them. That is fair enough. But once you are close to engaging, a firm with the real thing will name them in conversation and say whether they will actually work on your claim. A firm that will not, at any stage, is asking you to take a marketing line on trust.

"If you have ex-HMRC people, what did they do at HMRC, and at what grade?"

HMRC has tens of thousands of staff, most of whom have never worked a compliance case, let alone an R&D one, and most sit below inspector grade, as Officers or Compliance Caseworkers. A few years on a helpline is ex-HMRC but it is not the same as having experience running R&D enquiries. Ask what grade they held, which team they sat in and whether they dealt with R&D claims and enquiries from the HMRC side. LinkedIn will usually tell you whether the answers hold up.

"Who can talk to our engineers without an interpreter?"

If nobody at the firm, or available to it, can hold the technical conversation directly, your narrative will be written by someone paraphrasing what they were told. That reads exactly as it sounds when HMRC opens it.

"Do you have tax people? If we would be better off surrendering or carrying back losses rather than taking the credit, who on your team does that analysis?"

R&D relief is tax legislation before it is anything else, and the claim does not end at the qualifying expenditure. Whether to take the credit, surrender enhanced losses as group relief or carry them back to an earlier year is a corporation tax decision, and getting it wrong can cost more than the fee. A firm with a sales team and a writing team but no tax capability cannot do this analysis, and will usually default to the cash credit because it is the simplest thing to sell. You are listening for a name, and for some recognition that the question exists.

Questions about eligibility and judgement

These are the questions that reveal whether a firm has judgement or a process for saying yes.

"How do you decide what qualifies?"

You are listening for the statutory test and the guidelines, for the language of technological advance and technological uncertainty and for the competent professional as the person who determines both. You are listening against a description of a workflow. "We take you through a structured process" describes their internal admin. It does not answer the question.

"What in our work do you think might not qualify?"

The single most useful question in this list, and the one firms are least prepared for. Everything you've done is unlikely to be entirely R&D qualifying. A firm that has looked properly will name something and explain why it falls outside. A firm that says it all looks claimable to them has told you either that they have not looked or that they were never going to say otherwise.

"Would you ever tell a company it does not have a claim?"

Everyone says yes. The follow-up is the question: when did you last do it, and what was the work? A firm that can describe the case has done it. A firm that answers in principle has answered in principle. The same applies to a repeat claim. Ask whether they reassess each year against a baseline that has potentially moved, or assume a company will always qualify and simply roll the last narrative forward.

"What would you need from us to be confident this qualifies?"

A good answer is a list of people, documents and time. A bad answer is that they can handle it from next to no information, such as simply looking at your financial accounts.

Questions about the process and your expenditure

"How would you establish our qualifying expenditure, and how long does it take?"

This is where the risk lives. The answer should involve your finance people, the people who did the technical work and be more involved than just a few minutes. It needs to consider in detail how staff time is attributed, how subcontracted and externally provided work is treated and where the boundaries fall. A firm that makes this sound quick and painless has described either a process you have not understood or one you would not want.

"Will you come and see the work?"

For desk-based R&D such as software, remote is usually fine. For anything practical or site-based (a production factory, a workshop, a lab, a testing range), an adviser who has seen the work first-hand tends to describe it better and ask sharper questions, and that shows under enquiry. Some firms are entirely remote and a visit is simply not part of how they work. That is not wrong, but if your work needs to be seen, the answer you want is yes, or a clear reason why not, and you want it before you sign rather than when HMRC asks something the adviser cannot answer first-hand.

"Who reviews the claim before it goes to HMRC, and were they involved in preparing it?"

You are ideally looking for an independent quality assurance step. A claim signed off by the same person who prepared it is not necessarily wrong, but it is a potentially weaker process than a second pair of eyes, and a small firm should at least be able to tell you how it compensates for that.

"How much of this runs through a portal or questionnaire, and how much is conversation?"

Portals are not the problem. Plenty of good firms use one to collect costs and documents, and it saves everyone time. The absence of one is usually fine too. The question is what happens after any initial form is filled in. If the answer is that a report appears, with no call to the people who did the work and nobody testing what was actually uncertain, then the questionnaire is the process and the narrative will be written from it. You are listening for where the conversation sits in the process, who has it and with whom.

"Will I see and approve everything before submission?"

The only acceptable answer is yes, in writing. It is your company's return and your company's liability. A firm that submits without giving you plenty of time to review and agree the draft submission has misunderstood whose risk this is.

"Who submits, and how do you work with our accountant?"

Claims fall over in the handoff more often than people expect. Ask who produces and submits the Additional Information Form and who is responsible for the later tax return (CT600) submission itself.

Questions about fees

"What is the basis, and what exactly is included?"

Ask them to mark off the components: eligibility assessment, technical interviews, establishing qualifying expenditure, the narrative, the cost analysis, the Additional Information Form, claim notification where relevant, submission or liaison with your accountant, enquiry defence and whether this is full or partial. The cheapest quote is usually cheapest because some of those are missing.

"What triggers the fee?"

This could be on submission or when HMRC pay out, with part payment up front, or some other model. It is good to know which.

"What do I pay if the claim is reduced to nil?"

Ask whether anything already paid is refundable. For a genuinely contingent firm the answer may be nothing, which is a real point in its favour. For a fixed fee firm you have paid the fee. Neither is wrong. Discovering which one you bought after the event is.

"Does the fee include enquiry support?"

Percentage or contingency fees are a common model for R&D claims. They often include enquiry defence, but do not assume that yours does. Check. If defence is excluded, expect a materially lower fee.

"What stops sales staff inflating my claim?"

Percentage fees reward larger claims, and many firms pay sales commission on the claims each person converts. Ask how the firm prevents overclaiming, particularly where commission-paid sales staff are involved.

"What happens to commission paid on my claim if the salesperson leaves before HMRC rejects it?"

Ask any firm whether commission is clawed back on a rejected claim and most will say yes. You have no way of testing that, or whether the controls they think they have actually work in practice. Asking it this way makes them think about the process. Enquiries can open up to a year after submission and take two or three years to resolve, so the salesperson having left is a real possibility. If the answer is nothing, or waffle, the person who sold you the claim had every incentive to oversell it and no exposure to being wrong.

"Is my fee refunded in full if HMRC rejects the claim?"

Ask whether the firm refunds its fee, in full, on a rejected claim. Not a partial credit against next year.

"How long is the contract, and am I free to leave?"

Multi-year lock-ins and automatic renewal clauses do happen. Check you are not signed up for years without knowing it.

Questions about HMRC enquiries

"Tell me about an enquiry you defended and lost."

Ask this one. Any firm with real volume has lost one, or has had a claim reduced, and the way they handle the question is worth more than any statistic they can give you. You are listening for a specific case, what the disagreement was about and what they changed afterwards. A firm that has never lost one is telling you the sample is small, or the firm is new, or it may be that the answer is not quite straight.

"What is your enquiry rate?"

Useful as a conversation, potentially useless as a metric. It is self-reported, unaudited and varies legitimately with sector, claim size and how long a firm has been filing. Zero across hundreds of claims is not always the boast a firm thinks it is in the current compliance environment.

"Who handles an enquiry, by name, and is it done in house?"

Defence can be outsourced, bought in or provided through an insurance-backed arrangement. None of that is improper. What matters is who will actually handle it, whether the cost is included in your fee and whether there are any limits on the support provided.

"Is defence included, and if not, what does it cost?"

The firm has been paid to produce the claim. Being asked to pay again to defend the work it produced is the moment most people wish they had asked this.

"Is enquiry defence limited?"

For example, does it stop after they help with the first or second letter, or are they with you all the way to tribunal if needed, at no extra cost to you?

"If enquiry defence is paid separately, how?"

Where HMRC enquiry defence is not included in your fee, some firms may charge for it separately or by the hour, or in blocks, or pass you on to a third party to agree a fee with them, on top of any fee you have already paid the adviser. You need to be clear on this because R&D enquiries are a real possibility.

Questions about data, insurance and risk

"What is your professional indemnity insurance (PII) cover?"

Confirm it exists and that the level is proportionate to the claims they prepare. A firm submitting six-figure claims on a minimal policy may not be insured in a meaningful way that helps you.

"Where does our information go, and does any of it go into AI tools?"

You are handing over financial records, project documentation and possibly source code or lab notes. It is a fair question and a firm that treats it as an odd one has answered it.

Better versions of the questions everyone asks

Four common questions, each of which invites the answer you do not want, and the version that does not.

Instead of "are you HMRC approved?"

ask nothing, because there is no such thing. HMRC does not approve, accredit or endorse advisers or their methodologies. The question only gives a firm the chance to say something misleading.

Instead of "how much will we get?"

ask "what would you need to see before you could tell me?" The first question invites a number that nobody can responsibly give you. The second tests whether they know what the work involves.

Instead of "have you worked with companies like us?"

ask "name one, and tell me what was technically uncertain about their work." The first is answered yes by everybody. The second cannot be answered by a firm that has not done it.

Instead of "how quickly can you submit?"

ask "how long does it take to do this properly?" Speed is a feature of their process, not of your outcome, and it is not the thing you want them optimising.

How to read the answers

Across all of it, the same pattern separates the firms.

Good answers are specific

They contain names, dates, project descriptions and numbers that are not your claim value. They cite the test rather than the process.

Good answers admit limits

A firm that tells you what it does not know, what might not qualify or what it would need before committing is a firm that expects to be held to what it says.

Bad answers are fluent and general

Not wrong, exactly. Just untestable. Every sentence could have been said to any company in any sector.

Bad answers return to the benefit

Ask about quality assurance, get an answer about how much companies like yours typically recover. The redirection is the tell.

The rule of thumb: if an answer would work word for word for a different company in a different industry, it was not an answer to your question.

The list, in one page

Ask Listen for
Who would actually do this work, and can I meet them? Names, in one sentence
Who writes the technical narrative? A name, and whether it is the same person who will defend it
What have they worked on in our field? The baseline in your field, not a sector label
Your former HMRC people, who are they? Names, grade, team and what they did there
Do you have tax people who handle loss surrender and carry back? A name, and recognition the question exists
How do you decide what qualifies? The statutory test, not a workflow
What in our work might not qualify? Something specific, with a reason
When did you last tell a company it had no claim? A case, not a principle
How would you establish our qualifying expenditure? Our people, their people, weeks
Will you come and see the work? Yes for site-based R&D, or a clear reason why not
Who reviews it, and did they prepare it? An independent name
How much runs through a portal, and how much is conversation? Where the conversation sits and who has it
Will I see and approve it before submission? Yes, in writing
Who submits, and who owns the Additional Information Form? A clear division with your accountant
What is the fee basis, and what is included? Components marked off individually
What triggers the fee? Submission, payment or later
What do I pay if the claim is reduced to nil? A number or percentage and what is refundable
What stops sales staff inflating my claim? A named control, not "trust us"
What happens to commission if the salesperson leaves before HMRC rejects it? Clawback, with a period attached
How long is the contract, and am I free to leave? One or two claim periods, or more, and does engagement auto-renew
Tell me about an enquiry you defended and lost A specific case, and what changed after
Who handles an enquiry, and is it in house? A name and an honest answer about outsourcing
Is defence included? Yes, or a price
Is defence limited, and if so where does it stop? All the way, or a clear stopping point and price
What is your PII cover? A level proportionate to the claims
Where does our information go? A straight answer, without surprise at the question

The question to fall back on

If you only ask one thing, ask what in your work might not qualify.

It cannot be answered with marketing. It cannot be answered without having looked. It forces a firm to show you its judgement rather than its enthusiasm, and judgement is the entire product. A firm that will not draw a line anywhere in your work is not going to draw one when it matters either, and the person who finds out where the line actually was will be an HMRC inspector, a year from now, talking to you and not to them.

Frequently asked questions

What should I ask an R&D tax adviser before appointing them?
Ask who would actually do the work and get names. Ask how they decide what qualifies and listen for the statutory test rather than a description of their workflow. Ask what in your work might not qualify. Ask how they would establish your qualifying expenditure and how long it takes. Ask what the fee covers, what triggers it and what you pay if the claim is reduced to nil. Ask who handles an HMRC enquiry, by name, and whether it is included.
What is the single best question to ask an R&D tax adviser?
Ask what in your work might not qualify. It cannot be answered with marketing and it cannot be answered without having looked, so it forces a firm to show its judgement rather than its enthusiasm. A close second is asking about an enquiry the firm defended and lost, because any firm with real volume has one and the way they answer it is more informative than any statistic they can offer.
Should I ask an R&D adviser how much my claim will be worth?
Not if you are expecting a precise answer. Nobody can tell you the exact final claim size before your qualifying expenditure has been established, which is a joint exercise that takes time. The question invites a number that cannot responsibly be given, and a firm willing to give one anyway has told you something about itself. Ask what they would need to see before they could tell you instead.
How do I know whether an R&D adviser's answer is any good?
Good answers are specific and contain names, cases and stated limits. Bad answers are fluent, general and steer back to the benefit. The rule of thumb is that if an answer would work word for word for a different company in a different industry, it was not an answer to your question.
Is it reasonable to ask an R&D adviser about their enquiry rate?
It is reasonable to ask, but treat the answer as a conversation rather than a definitive metric you can use to compare advisers. It is self-reported, unaudited and varies legitimately with sector, claim size, how it is measured and how long the adviser firm has been filing. How a firm discusses its enquiries tells you considerably more than an unsubstantiated percentage does.

Question about the platform or the Radar Assessment itself? See the FAQ.

Ready to look at specific advisers?

Search advisers