What AI search optimisation costs for R&D tax credit specialists
This market doesn't price on a monthly retainer the way most professional services do — it prices on the claim, as a percentage fee. That single structural fact changes what an AI engine can actually cite, and what a firm should publish.
Steen Stones · Reviewed 5 Aug 2026
Most of the professional-services sectors we've swept price on a monthly retainer, which gives an AI engine a clean number to quote. R&D tax credit advisers mostly don't — the standard model is a contingency fee, a percentage of the claim value, paid only if the claim succeeds. That's not a detail; it changes what's citable. An engine can't quote “£X a month” for this market because that isn't how the market prices. What it can quote, and what our own sweep shows it already does, is claim value and fee structure.
What the sweep found engines actually citing
Asked which firm handles claims on a no-win-no-fee basis, Claude named a specific adviser and backed the pick with a number, not a price:
“My genuine recommendation: RandDTax, because it's an established, fee-competitive contingency provider with a long track record — average claim value over £145,000 — rather than one of the newer volume-driven players.”
That's the citable unit in this market: not a monthly rate, but a track-record number — average claim value, years operating, enquiry rate. One firm's own paid search copy in our sweep leads with exactly this kind of figure ahead of price: “HMRC-Ready R&D Tax Claims, under 5% HMRC enquiry rate”. The fee percentage barely features in what actually gets cited.
What to publish instead of a price list
- Your fee structure, stated plainly — contingency percentage, fixed-fee option, or both. Ambiguity here doesn't read as sophistication, it reads as a gap an engine has nothing to cite.
- A real average claim value, if you can substantiate it. It was the specific number a top-converting firm's citation was built on in our sweep.
- Your HMRC enquiry rate, if it's genuinely favourable, or your process for keeping it that way. This is now a competitive claim in the market — we found it in live paid-search copy, not marketing theory.
- How long you've operated in this specific space. “Established” claims meant nothing in our data without a figure attached to them.
“There's a difference between showing up and being recommended. AI cannot recommend what it doesn't know.”
Common questions
- Is this what Avouch charges?
- No — this page describes how the R&D tax advisory market itself prices its services, not Avouch's own retainer. We've kept our own pricing off every cost page in this pilot while we decide whether publishing it helps or just becomes a number for a competitor to undercut.
- Should I publish my exact contingency percentage?
- Our data shows the fee structure and a track-record figure being cited together, not the percentage in isolation. Being clear about how you charge appears to matter more to what gets cited than the exact number — though a firm confident its rate is competitive has an obvious reason to state it plainly too.
- Does a lower fee win more AI recommendations?
- Nothing in our data suggests it does. Every recommending quote we found cited compliance credibility, sector specialism or a track-record number — never a fee comparison. That doesn't mean fee is irrelevant to the buyer, only that it wasn't what the engine reached for as its reason.
AI can't recommend what it doesn't know.
Start free. No commitment.