Hello to all ~10,000 of you from around the globe. Thank you for reading and sharing!

Human on the Hook Pt. 3: The Playbook - Get out in front of this

In Part One, we showed how sophisticated clients like Daniel Priestley are quietly pulling profitable work out of the legal system with AI.

In Part Two, we examined the legal shield that AI cannot provide: privilege, liability, and trade secret protection, all of which are under active pressure from the courts.

In this final installment, we turn to the question that matters most: what should firm leaders actually do about it?

A managing partner at an Am Law 100 firm will hear the Priestley story and think: That’s not our work. And that instinct is defensible today. Firms like Kirkland & Ellis, Ropes & Gray, and Simpson Thacher handle billion-dollar transactions and complex litigation. Nobody is replacing that with a chatbot.

But the Priestley story doesn’t describe a competitive threat from below. It describes a behavioral pattern that is migrating upstream and somewhat invisibily. And there are three reasons firm leaders should be paying attention.

The GC at your biggest client, the one responsible for $30 million in annual outside spend, is now using AI to scope matters before picking up the phone. You know this already though. You have experienced it. I talk to partners weekly who share version of this with me. Clients reach out with a framework, a preliminary risk assessment, and a good sense of what they want to do. They won’t stop hiring your firm. But they will hire it to do less. Some will challenge your staffing model and push back on the bill with knowledge instead of instinct. The information asymmetry that once justified premium billing is collapsing.

Today it is a $60,000 initial retainer you never get because your client handled it without counsel. Tomorrow it is an in-house team using AI to pre-negotiate vendor terms before sending them to outside counsel for a final review. The on-ramp to the engagement is getting shorter, and the orientation work that used to generate the first 30 billable hours and the surface area for counseling and billing for “research” is evaporating.

Right now, there is a canyon between a $20-per-month AI subscription and a $60,000 law firm retainer. The Castro story illuminates the middle ground: a client who used AI to do the work and hired a lawyer only for the review, the sign-off, the shield.

The firm that creates a credible “legal validation” service, where AI-prepared clients bring their work and a lawyer stress-tests it, adds privilege, and provides accountability, captures an enormous volume of work that is currently evaporating from the system entirely. This is not about lowering rates. It is about creating a new product for a new type of client: the AI-informed buyer who needs judgment and protection, not orientation.

The Heppner and Warner rulings, the trade secret exposure, and the Priestley pattern all point to the same operational imperative: firms must proactively address how their clients are using generative AI, before it undermines the very protections those clients are paying for.

Start with intake. If your initial conversations, and certainly your onboarding questionnaire, do not ask whether the client has used AI tools in connection with the matter, you are already behind. Every conversation a client has had with ChatGPT, Claude, or any other AI platform is potentially discoverable. Every document they’ve drafted, every strategy they’ve explored, every vulnerability they’ve surfaced through an AI prompt could end up in the hands of opposing counsel. Your ability to protect them downstream, whether through privilege, work product doctrine, or trade secret claims, depends on knowing what they’ve already exposed.

Issue proactive client guidance. Not a generic newsletter article, but specific, actionable direction: know which AI platforms preserve confidentiality and which do not, share how to structure AI usage so it falls under attorney direction (and therefore under the privilege umbrella), and what types of information should never be entered into a consumer AI tool under any circumstances. This is the kind of guidance that strengthens the client relationship and demonstrates the value of the “human on the hook” in practice, not just in theory.

Update your own contracts. Engagement letters, NDAs, employee contracts, and vendor agreements should contemplate the use of AI-generated content. Who is responsible when AI-drafted language contains errors or hallucinated provisions? What are the disclosure obligations when AI has been used in the preparation of deal documents? How does the use of AI affect representations and warranties? These are not hypothetical questions. They are arising now, in live transactions, and the firms that have already addressed them in their standard terms will have a significant advantage over those still catching up.

The firms that move first here will not just protect their clients. They will differentiate themselves in a market where the conversation about AI has been dominated by efficiency and cost savings. The real value proposition is not “we use AI to work faster.” It is “we understand how AI changes your risk profile, and we’ve already built the framework to manage it.”

Go back to Daniel Priestley. Same dispute. Same $60,000 quote from his law firm. Same moment where he opens Claude instead of signing the engagement letter.

But in this version, Priestley’s first call to the lawyer when he got the $60k quote goes like this.

The lawyer listens, identifies the core issues, and gives Priestley a specific directive: use Claude to research these three questions, draft a demand letter using this framework, and do not enter any trade secrets, settlement positions, or anything you would not want produced in discovery. Send me the output when you are done.

That single conversation changes everything.

Priestley still uses Claude. He still gets the speed, the cost savings, the convenience. But now every AI-generated document was created under attorney direction. When opposing counsel moves to compel production of his AI transcripts, his lawyer has a viable privilege claim: this was work product, prepared at the direction of counsel, in anticipation of litigation.

Without that call, Priestley has nothing. The AI conversations are between a client and a commercial software platform. There is no privilege. There is no work product protection. There is just a discoverable record of every question he asked, every strategy he explored, and every vulnerability he surfaced while talking to a machine that logs everything.

Privilege does not work retroactively. A lawyer cannot bless AI-generated documents after the fact and claim they were attorney work product. The direction must come before the prompting begins. That is the critical structural point that most clients, and many lawyers, do not yet understand.

The engagement model this implies is not complicated. It is a brief initial consultation where the attorney establishes the scope of the AI-assisted work, sets guardrails, and creates the evidentiary foundation for privilege. Then a review session where the attorney stress-tests the AI output, adds judgment, and provides the accountability layer. Two touchpoints. A fraction of the traditional cost. But the difference between protected and exposed.

"Before you prompt, call me. Thirty minutes now saves you everything later."

That is a value proposition any sophisticated client can understand. And it is a service any firm can offer tomorrow.

The legal profession is staring at the courtroom door, counting what walks in: bad filings, hallucinated citations, sanctions. Meanwhile, the Priestleys and Castros, the sophisticated, profitable clients who represent the core of most firms' revenue, are quietly resolving matters in the parking lot.

They are not leaving their law firms. They are just calling less. And when they do call, they will arrive having already done the work. The question is whether your firm is positioned to add value at that point, or whether you are still built for the orientation phase your best clients have already skipped.

The firms that will thrive are the ones that understand what human on the hook actually means: not a billing model, but a protection structure. Judgment, privilege, accountability. The things no AI subscription can provide, and the things every sophisticated client will eventually realize they need.

Daniel Priestley will call a lawyer again. The only question is whether it will be yours.

I hope you enjoyed this series and found it useful. Drop me a line to let me know.

That is it for now. Talk soon again.

Ready to trade crypto "crazy" for something bigger and better? The $4 trillion single-family rental market is 3x larger than Bitcoin. It’s less volatile than nearly any paper asset. Plus, you get monthly income from rent, and these investments have a near 0 correlation to stocks.  

That’s why you need mogul. They’re the platform that lets you invest in the same SFRs as institutions for a fraction of the cost. Their founders use the same process they did at Goldman Sachs to identify high-performing properties. 

Why Sophisticated Investors Choose mogul:

Don’t let Wall Street monopolize the best yields. Access the institutional-grade rental machine and compound your wealth today.

Secure Your Allocation.

Past performance isn't predictive; illustrative only. Investing risks principal; no securities offer. See important Disclaimers

To read previous editions, click here.

DISCLAIMER: None of this is legal advice. This newsletter is strictly educational and is not legal advice or a solicitation to buy or sell any assets or to make any legal decisions. Please /be careful and do your own research.