How UK tax research tools differ (2026)

In short: UK tax research tools in 2026 fall into three groups — the commentary libraries (LexisNexis Tolley, Croner-i, Thomson Reuters Checkpoint, Wolters Kluwer CCH), the AI-native research tools (Blue J, Fiscari), and general-purpose AI such as ChatGPT. The differences that matter are underneath the marketing: what the AI is grounded in, whether its citations land on the source, whether it refuses when it cannot ground an answer, and whether you can see the price before a sales call.

What “tax research” is made of

A typical piece of UK tax research blends four layers. Every tool handles them differently, so the first question about any product is which layers it actually draws on, and whether it shows you the trail.

1. Primary law — the Taxes Acts, VATA 1994, TCGA 1992, ITTOIA 2005, statutory instruments and case law.
2. HMRC’s own material — internal manuals, VAT notices, Revenue & Customs Briefs and agent guidance.
3. Editorial commentary — Simon’s Taxes, De Voil, Tolley guidance, Croner-i practice notes and similar.
4. The client’s facts — which no tool knows unless you tell it.

The three groups of tool

Commentary libraries

These are the long-established platforms. Their strength is editorial depth: commentary written and maintained by specialists, which is worth paying for on complex reorganisations, share schemes and structuring work. Their AI features summarise or search that commentary. Pricing is usually quote on request and sold on annual terms. Thomson Reuters’ current Checkpoint marketing in the UK centres on international coverage (Checkpoint World) and its integration with Digita and ONESOURCE; its CoCounsel AI announcements to date have been US-focused.

AI-native research tools

Built around retrieval and generation from the start. Blue J is the most established, with multi-jurisdiction coverage, a published Individual tier at £1,500 per user per year, and a reported disagree rate of around 1 in 700; in the UK it has an exclusive Big Four partnership with KPMG announced in February 2024. Fiscari, built by Glasscubes, answers only from UK legislation, HMRC manuals, VAT notices and tribunal decisions, cites the paragraph, refuses when it cannot ground an answer, and publishes its pricing from £19.50 per user per month with a free tier.

General-purpose AI

Excellent for drafting, summarising and explaining, and the wrong shape for tax research: it is not restricted to a fixed corpus, it can produce citations that look right but do not exist, and it answers every question with roughly the same confidence whether or not the sources support it.

Four questions to ask any vendor

1. What corpus is the AI grounded in? A defined, auditable source set — a commentary library or the public UK primary sources — rather than the open web.
2. Can you click through to the source? A citation should land on the section or paragraph the answer rests on. If it does not, the grounding is marketing rather than engineering.
3. Does it refuse when it cannot ground an answer? The single biggest quality signal. A tool that always answers will sometimes invent one.
4. Can you see the price and trial it first? Being able to put a week of real client questions through a tool before committing is the fairest test there is.

Which kind of tool fits which firm

• Large firms and in-house teams doing structuring work — keep a commentary library for the matters that need it; consider an AI-native tool alongside it for the daily research load.
• Mid-tier firms at renewal — run the incumbent and an AI-native tool against twenty real questions from recent file notes, and compare what each returns and what each refuses.
• Cross-border practices — a multi-jurisdiction tool such as Blue J or Checkpoint World does something a UK-only tool cannot.
• Sole practitioners and small firms — published pricing, a free trial and cited answers from primary sources usually matter more than a commentary library you will rarely open.
• Software companies — look for an API. Fiscari offers a REST API with JSON responses and citations in the payload.

Head-to-head comparisons

• Fiscari, Tolley+ AI, Croner-i, Blue J and ChatGPT compared
• Fiscari vs Tolley
• Fiscari vs Croner-i
• Fiscari vs Blue J
• Fiscari vs Thomson Reuters Checkpoint
• Fiscari vs ChatGPT

Compiled from each vendor’s own published material, October 2026. Third-party prices and accuracy figures are those vendors’ own figures; confirm current terms with the vendor. Fiscari is a Glasscubes product.