The £2 Letter: What the World's First AI-Only Law Firm Means for US Small Practices

The £2 Letter: What the World's First AI-Only Law Firm Means for US Small Practices

A law firm charges £2 to send a legally valid letter before action. It is authorized by a national legal regulator. It is already handling cases.

This is not a pilot program. It is not a hypothetical. Garfield.Law Ltd is the world's first fully AI-powered law firm to receive authorization from the UK Solicitors Regulation Authority (SRA) — the equivalent of a state bar granting a law license — to provide regulated legal services. The firm targets debt recovery cases in the UK small claims court.

If you practice in small claims, debt collection, landlord-tenant, or any other high-volume, routine legal area, this is worth understanding.

What Garfield.Law Actually Does

Garfield.Law is built on a hybrid AI and expert system modeled to the UK Civil Procedure Rules. It handles the litigation assistance workflow for small debt claims (up to £10,000) from initial contact through pre-action stage.

Pricing:

• £2 for a "polite chaser" letter
• £7.50 for a formal letter before action

The firm explicitly targets two audiences: small businesses that would otherwise handle these claims without legal counsel, and high-street (small) law practices that want to offer AI-powered service delivery to their own clients.

It is actively expanding into additional practice areas.

The SRA authorization is the key fact. Garfield.Law is not a "legal tech tool" that helps lawyers draft letters. It is an authorized law firm that provides regulated legal services — the same classification as any firm with a licensed solicitor. The regulator reviewed the model, assessed the risk controls, and granted authorization.

Why the UK Model Matters for US Small Firms

Garfield.Law operates under UK law and UK regulatory authority. It does not currently practice in the United States.

But the SRA has just answered a question every US state bar is watching: can an AI-only firm provide regulated legal services and meet professional responsibility standards?

The answer, as of 2026, is yes — at least in one major jurisdiction.

US state bar experiments are already moving the same direction. The Utah sandbox and the Arizona Alternative Business Structure (ABS) framework both permit non-traditional law firm ownership structures that could enable a Garfield-style model under US law. The regulatory question in the US is not whether — it is which state bar moves first, and when.

Three US AI-native law firms — Manifest OS ($60M Series A, immigration law), General Legal (YC W26, commercial contracts at $500/contract, one-hour turnaround), and Superlegal (Utah sandbox, $117/contract) — are already building comparable models under existing alternative frameworks. The Garfield authorization adds a new data point: a traditional regulator has now vetted the model.

That is a signal to other regulators, not just competitors.

Which Practice Areas Are Most Exposed

Not all legal work faces the same AI-only law firm risk. The exposure is concentrated in a specific type of work.

High-risk (routine, document-driven, high volume):

• Small claims debt recovery
• Standard letters before action and demand letters
• Residential landlord-tenant notices
• Basic commercial contract templates (standard NDAs, simple vendor agreements)
• Routine compliance filings
• Boilerplate employment agreements

Lower-risk (judgment-intensive, relationship-dependent):

• Complex commercial litigation
• Multi-party transactions with negotiated terms
• Regulatory risk advisory
• Criminal defense
• Family law (custody, contested divorce)
• High-value real estate with complex contingencies

The Garfield.Law model at £2/letter is not competing with your complex commercial litigation practice. It is competing with the part of your practice where you send three letters and hope the client pays.

The question for any small law firm is: what percentage of your annual revenue comes from the first category? If the answer is more than 20%, the Garfield model is relevant to you — even before it arrives in the United States.

Three AI-Native Competitors, Three Price Points

The competitive picture for small US law firms now has three distinct vectors operating simultaneously:

From the UK (regulatory precedent): Garfield.Law — authorized AI-only firm, £2 letters, actively expanding practice areas. Sets the bar for what a regulator will accept.

From below (US AI-native firms): Manifest OS, Superlegal, General Legal, LegalOS, Arcline — five firms in different stages. Manifest raised $750M in implied value and explicitly refuses to sell AI tools to existing law firms. General Legal charges $500 for a commercial contract with a one-hour turnaround.

From inside your clients (in-house AI): Wordsmith ($70M, 14x revenue growth in 12 months) and Sandstone AI ($30M, 40x revenue growth in 90 days) are selling AI systems to the corporate clients you currently advise as outside counsel. Clients are not waiting for outside counsel to offer lower prices. They are building internal capability to handle routine legal work themselves.

What Garfield.Law adds is the regulatory validation. When a skeptical client says "an AI firm could never be authorized to practice law," the answer, as of today, is: one already is.

What Small Law Firms Should Do

This is not an immediate crisis. Garfield.Law is in the UK. US regulatory change moves slowly. The AI-native US firms are still building.

But the time to think about this is before a competitor opens in your market, not after.

Step 1: Audit your revenue by work type. Pull your billing records from the last 12 months. Sort by matter type and estimate what percentage of revenue came from routine, repeatable work versus judgment-intensive, relationship-driven work. If you have never done this, now is the time.

Step 2: Calculate your small claims and debt recovery exposure. If you handle creditor-side debt collection, landlord-tenant notices, or similar high-volume, low-complexity matters: what is the total revenue? What would happen to your firm if a competitor offered those services at one-quarter the price?

Step 3: Build a moat before you need one. The defensible law firm in an AI-native competitive environment has two characteristics: (a) institutional knowledge and client relationships that cannot be replicated by a new entrant, and (b) AI tools that reduce its own delivery costs so it can compete on price when it needs to. The second part is accessible today through Harvey, CoCounsel, or Claude for Legal — all of which are available without an enterprise implementation project.

Step 4: Watch the regulatory calendar. Utah, Arizona, and California are the US bar associations most likely to move first on alternative structures. A Garfield-style authorization in any of these states changes the timeline significantly. Subscribe to your state bar ethics alerts.

The Bottom Line

The significance of Garfield.Law is not that it offers cheap legal letters in the UK.

It is that a national legal regulator looked at an AI-only law firm model, reviewed it against professional responsibility standards, and said yes.

That decision will echo. Other regulators are watching. US AI-native firms are lobbying. The question is not whether the US will see Garfield-style authorized AI-only firms. It is when, and which practice areas they target first.

Small law firm owners who understand their exposure now — and who use that window to strengthen their relationships, document their institutional knowledge, and reduce their own delivery costs with AI — will have options when the model arrives.

The ones who don't will face a £2 letter with no response ready.