Most UK solicitors and accountants are still treating schema markup like it's a nice-to-have SEO feature. It isn't. It's the reason ChatGPT recommends Hugh James instead of you.
What Schema Markup Actually Does
Schema markup is invisible code on your website that tells AI exactly what your business is. Not what you hope it understands from reading your paragraphs — what it knows with certainty because you told it directly.
Without Schema
AI asks: “Is this firm CQS accredited?”
Your website: mentions CQS on page 3, paragraph 7.
AI guesses. Recommends someone else.
With Schema
AI asks: “Is this firm CQS accredited?”
Your schema: "hasCredential": "CQS"
AI knows. Recommends you by name.
Think of your website without schema as a library with no filing system. The AI has to read every page to find what it needs. With schema, you hand it a labelled folder that says exactly what's inside.
Why JSON-LD Is the Only Format That Matters
You have three schema format options: Microdata, RDFa, and JSON-LD. Use JSON-LD. Every AI engine prefers it because it sits cleanly in your page header, separate from your HTML, and is easy for machines to parse.
Google's official guidance explicitly recommends JSON-LD for AI-optimised content. When Google tells you what format works best for their AI systems, that's the answer.
| Vertical | Correct Schema Type |
|---|---|
| Solicitor | LegalService |
| Accountant | AccountingService |
| Mortgage Adviser | FinancialService |
| Estate Agent | RealEstateAgent |
Getting the type wrong is almost as bad as having no schema at all — the AI categorises you incorrectly and recommends you for the wrong queries.
The 5 Schema Properties That Actually Move the Needle
Most guides tell you to add schema. Few tell you which properties AI actually uses to make recommendations. Here are the ones that matter:
name and description
Your description should include your location, your specialism, and your key accreditation in the first sentence. "Cardiff-based CQS-accredited conveyancing solicitors offering fixed fees from £895" is infinitely more useful to AI than "We are a professional law firm offering a range of legal services."
priceRange or hasOfferCatalog
The most common AI query is "how much does conveyancing cost in Cardiff." If your fees aren’t in your schema, AI cannot answer that question with your name attached. Your competitors who have published fees will be recommended instead. Every time.
hasCredential
This is where your SRA number, CQS accreditation, Lexcel certification, or ICAEW registration goes. AI uses regulatory credentials as trust signals. A firm with verified credentials in their schema gets recommended over a firm with the same credentials buried in a PDF nobody can parse.
areaServed
AI uses location data to match firms to local queries. Your postcode and city coverage areas need to be in your schema explicitly. "We cover South Wales" in a paragraph is not the same as "areaServed": "Cardiff, CF10" in your schema.
aggregateRating
Your Google reviews need to be reflected in your schema. 52 reviews averaging 4.8 stars means nothing to AI if it isn’t structured. Once it is, AI can confidently say "Lucas Law Solicitors in Penarth has 52 reviews averaging 4.8 stars" — and that becomes a recommendation.
The Problem With Schema You Set and Forget
Schema drift
Where markup falls out of sync with actual page content — is one of the most common reasons AI systems stop citing previously trusted content.
You update your fees in January. You forget to update your schema. AI is still recommending you based on your old prices. A client arrives expecting £895 and you quote £1,100. That's a trust problem that starts with stale schema.
This is the core problem with manual schema implementation. Someone installs it once, forgets about it, and six months later it's actively working against them.
What This Means for UK Professional Services Firms
The firms that will dominate AI recommendations in 2026 are not necessarily the biggest or the best known. They are the ones with the most accurate, most complete, most current structured data.
A Cardiff conveyancing solicitor with properly implemented schema telling AI their CQS accreditation, fixed fees, postcode coverage, and 52 Google reviews will be recommended over a larger firm whose website has no schema at all.
In March 2025, both Google and Microsoft publicly stated they use schema markup for their generative AI features. ChatGPT then confirmed it uses structured data to determine which products and services appear in its results. This is not speculation. It is confirmed by the platforms themselves.
The TendorAI Approach
Most tools tell you what schema to add. TendorAI installs it for you and keeps it current automatically.
Fees updated?
Schema updates on your website within minutes.
New accreditation?
Appears in your structured data the same day.
Review count up?
Schema reflects it automatically.
No developer. No quarterly audits. No schema drift.
Frequently Asked Questions
What is schema markup in simple terms?
Do I need a developer to add schema markup?
How quickly does schema markup affect AI recommendations?
What schema type should a solicitor use?
What is schema drift?
Is your schema helping or hurting you?
Run a free AI visibility report. We'll check what AI platforms currently know about your firm — and what's missing.
Published 19 March 2026 by TendorAI
← Back to Blog