TwinLadder Weekly

Issue #12 | July 2025

---

Editor's Note

You draft a motion that traditionally takes four hours. Using Claude or GPT-4, you complete it in 45 minutes. The quality is identical — perhaps better, since the AI caught a jurisdiction-specific procedural requirement you might have missed.

Do you bill four hours? Forty-five minutes? Something in between?

I have been practising long enough to remember when this exact question arose with Westlaw. Before electronic research, a thorough case law review might take a full day in the library. Westlaw compressed that to hours. Nobody seriously argued you should bill for a day's work. But nobody seriously argued you should bill only for the minutes at the terminal, either. The profession muddled through, and billing practices quietly adjusted.

This time is different. The efficiency gains from AI are larger, more visible, and arriving at a moment when clients are already sceptical of hourly billing. And now Oregon has drawn the line with unusual clarity. Every lawyer using AI needs to understand where that line falls.

I raised this with colleagues at a roundtable in Berlin last month. The American lawyers in the room focused on the ethics opinions. The European lawyers focused on something different: client expectations. As one general counsel from a Dutch manufacturing company put it, "I do not care about your state bar rules. I care that I am paying for value, not for your learning curve." European clients, particularly in-house teams accustomed to fixed-fee arrangements, are less patient with the hourly billing model than their American counterparts — and AI is accelerating that impatience.

---

Oregon Draws the Line: What Ethical Billing Requires

The Core Rule

Oregon State Bar Formal Opinion 2025-205, issued March 2025, provides the clearest guidance yet on AI billing. I will state the core rule directly, because it deserves to be stated directly:

"If the use of AI results in significant time savings, lawyers may not engage in billing practices that duplicate charges or falsely inflate billable hours."

That is unambiguous. If AI cut your work from four hours to 45 minutes, billing four hours violates the fee reasonableness standard. Oregon is not alone — 23 states have now issued formal AI guidance and billing consistency is a recurring theme. Florida's Opinion 24-1 says the same thing differently: efficiency gains "must not result in falsely inflated claims of time." The ABA's Formal Opinion 512 frames it as a reasonableness question under Model Rule 1.5.

The Billing Scenarios

Oregon's framework addresses multiple billing scenarios, and they do not all point in the same direction.

Scenario	Oregon's Position	Practical Impact
Time actually spent	Bill only for actual attorney time	Clearest rule; reduces revenue for hourly billers
AI tool costs	Must be disclosed in writing before charging	Similar to Westlaw/Lexis pass-throughs
Verification time	Full review time is billable	20 min AI draft + 90 min review = 110 min billable
Learning time	Cannot charge to clients without consent	Firm overhead, like CLE attendance
Alternative fees	Investment in custom AI workflows is compensable (Virginia)	Supports transition to value-based pricing

The verification time nuance is important. If AI drafts a document in 20 minutes and you spend 90 minutes verifying accuracy, checking citations, and ensuring jurisdictional compliance, the full 110 minutes is billable. Oregon explicitly recognises verification as real attorney work. The AI generation time is what you cannot inflate — the human review time is legitimate.

The Economic Reality

The economic reality is stark. For hourly billers, AI creates a structural revenue problem. If you could bill eight hours for research that now takes two, you have lost six billable hours. Doing more work does not fully compensate unless you have unlimited demand. The solo practitioner who shifted to flat fees and now handles 2.5x the volume at the same per-matter price — revenue up 40% while clients pay the same — has found the more sustainable answer.

Billing Model	Before AI	After AI	Revenue Impact
Hourly (research task)	8 hours billed	2 hours billed	-75% per task
Hourly (with volume increase)	8 tasks/month	12 tasks/month	-25% net (if demand exists)
Flat fee (per matter)	8 matters/month	20 matters/month	+40% (same client cost)
Value-based	Priced on outcome	Priced on outcome	Neutral to positive

The profession is moving from hourly to value-based billing. AI is accelerating that shift whether we are ready or not. 77% of legal professionals now use AI for document review, research, or drafting. If your billing practices have not adapted, you are not just behind the ethics curve — you are behind the market.

The European Dimension

Liga Paulina raises a point that American commentary overlooks entirely: "In many European jurisdictions, fixed-fee and capped-fee arrangements are far more common than in the US, particularly for corporate and commercial work. The billing dilemma that Oregon is wrestling with has already been partially resolved by market forces in Amsterdam, Stockholm, and Frankfurt. The question for European firms is not whether to abandon hourly billing — many already have. It is how to price the value of AI-augmented expertise fairly."

For European firms, the Article 4 dimension adds another layer. If your firm can demonstrate documented AI competence — showing that staff have been trained to use AI tools effectively and verify outputs rigorously — that competence itself becomes part of the value proposition. A firm that can tell a client, "We are Article 4 compliant, our staff are trained, and we use AI to deliver better results faster at a fair price" is in a stronger commercial position than one that simply bills fewer hours. [MODERATE CONFIDENCE]

---

The Competence Question

A partner at a 20-person firm in Hamburg described his billing dilemma to me last month. His insurance defence practice bills hourly to carriers. AI has reduced his research and drafting time by roughly 40%. His revenue has dropped proportionally because he bills actual time. He cannot do more work because his caseload is limited by court schedules, not research time.

His response was to stop using AI.

Sit with that for a moment. A lawyer who acknowledged that AI improved his work product chose not to use it because it reduced his revenue under the hourly model. He is billing more time. He is delivering slower results. And his clients — insurance carriers who are themselves investing heavily in AI — are paying for that choice.

Is that competent representation? Oregon's opinion ties competence to AI: lawyers must understand AI capabilities and limitations, and the implication is that failing to use appropriate tools when they would benefit the client raises its own competence questions. Not using AI when it would produce faster, better results — because the billing model penalises efficiency — is an ethical tension the profession has not resolved.

The lawyer who refuses AI to protect revenue and the lawyer who overbills AI-assisted time are making the same error from opposite directions. Both are allowing their financial interest to distort their professional judgment. The billing model should follow the work, not the other way around.

---

What To Do

1. Update engagement letters this month. Add AI disclosure and billing methodology. Sample language: "We may use AI-assisted tools for research, drafting, and review. You will be billed for actual attorney time. AI tool costs, if any, will be itemised separately." Do this before a client asks.

2. Track verification time separately. When using AI for drafting or research, log AI generation time and human review time as distinct entries. This creates a defensible record and helps you understand your actual efficiency gains. European firms should retain this documentation as part of their Article 4 compliance records.

3. Model alternative fee structures. For your five most common matter types, calculate what a fixed fee would look like based on AI-assisted time plus a reasonable margin. Test one or two with willing clients. The data will tell you whether the transition works for your practice. European firms already working under fixed fees should recalculate margins to reflect AI efficiency gains — and consider whether to pass savings to clients or reinvest in quality.

4. Read Oregon Opinion 2025-205 in full. It is available as a PDF and it is the best single document on AI billing obligations currently available. Even if Oregon is not your jurisdiction — and for European readers, it is not — your professional body is likely heading in the same direction.

5. Connect your billing transparency to your competence story. Do not treat AI billing as a cost problem. Treat it as a value demonstration. The firm that shows clients exactly how AI improved their work — faster turnaround, broader research coverage, systematic verification — while billing fairly for actual effort is the firm that wins trust. In a market where Article 4 demands demonstrated competence, transparent billing becomes evidence of professional capability.

---

Quick Reads

• Oregon Formal Opinion 2025-205 (PDF) — the full text, clear and well-reasoned, addressing billing, confidentiality, competence, and supervision in a single comprehensive opinion.

• 23 states have issued formal AI guidance — Paxton's tracker of state-by-state positions, updated regularly and essential for multi-jurisdictional practices. European readers: use this as a benchmark for what your own bar associations will likely address.

• MIT Technology Review: AI and legal employment — 47.8% of attorneys at large firms now use AI tools; the article provides useful context on what is actually changing in day-to-day practice.

• Attorney at Work: AI and billing practices — practical advice on aligning billing with ethics obligations, written for practitioners rather than academics.

---

One Question

If the hourly billing model punishes efficiency, and AI makes efficiency unavoidable, which one survives?

---

TwinLadder Weekly | Issue #12 | July 2025

Helping European professionals build AI competence through honest education.