Your Next Client Won't Call You. Their Agent Will.

Your Next Client Won't Call You. Their Agent Will.

A few months ago, we rolled out Enterprise ChatGPT licenses to all 130,000 employees at Thermo Fisher Scientific. Not just marketing. Not just executives. Everyone. The instruction was simple: use this wherever and whenever you can to make yourself better.

Overnight, every marketer had a copywriter. Every sourcing manager could get a pricing analysis in seconds. Every in-house attorney had a research assistant.

That's the obvious story. Here's the one that matters more.

When One Side Gets Agents, the Other Side Has To

In the division where I worked until recently, they sell roughly four million SKUs to biopharma companies and universities. Nearly two-thirds of that purchasing is already system-to-system. Machines buying from machines. Automated procurement engines moving billions of dollars of scientific supplies with limited human involvement on either side.

We didn't design this future. We adapted to it. And the pattern is predictable:

1. It starts with automation, basic ordering and invoicing.

2. Then procurement systems get smarter, enforcing contract compliance, benchmarking suppliers, surfacing spend visibility.

3. One side deploys agents first: in our case, running algorithms to optimize pricing across millions of SKUs, running with little human intervention.

4. Then the other side responds with their own agents to analyze spend, identify overpayment, and rank suppliers automatically.

Neither side chose this. Each side had to do it as a rational response to the market.

The important part isn't the technology. It's the pattern: as soon as one side becomes agent-powered, the other must follow, or a capability gap opens.

If procurement can evolve this fast, so can legal.

What Happens When the GC Moves Up That Curve

Imagine a general counsel receives a new matter. Cross-border investigation. AI governance question. A merger.

Here's what won't happen: they won't start by thinking, "which partner should I call?"

Their Legal Agent will ask: What does our matter history say about firm performance? Who has relevant sector expertise? Who communicates clearly and on time? Who caused billing problems last year?

That agent will shortlist firms, score them, rank them, and eliminate some before a human ever weighs in. Meanwhile, your firm's agents will be assembling proposals, optimizing pricing, generating staffing plans, and negotiating preliminary terms within guardrails.

What used to be a human-to-human relationship begins with machine-to-machine evaluation. The firms who don't understand this won't even know they've been eliminated.

The Era of Perfect Memory

At Mount Insights I use Claude. Here's what my Claude Enterprise does: I can query my emails, calendar, Teams chats, and every document I've touched. I can ask it to show me every conversation about a vendor over the past year, or summarize every missed deadline in a thread, or surface what we promised a customer on a call six months ago.

Now imagine a GC doing that with your firm.

Every email your partner sent. Every update an associate missed. Every billing dispute. Every MSA negotiation. Every redline, invoice, and pitch deck.

Their agent will know your firm better than your BD team does. And the GC doesn't have to remember how good you were. Their agent will remember everything. Responsiveness will be evaluated. Clarity will be evaluated. Predictability will be evaluated. Value will be evaluated.

Good news for firms who consistently deliver. A problem for firms who coast on relationships.

Content Is Now Training Data

We used to say content is king. In an AI world, content is training data.

Agents don't browse your website. They evaluate structure, specificity, schema, context, recency, and extractability. This is already playing out in retail and healthcare: zero-click search is rising, AI Overviews are diverting traffic away from sites, and generative engines are becoming the primary discovery layer.

If your content isn't structured and machine-readable, LLMs won't find it. If they don't find it, they won't recommend it. If they don't recommend it, you disappear.

Agents will look for precise matter descriptions, industry-specific expertise, clear "what to do next" guidance, structured bios, schema around sectors and courts, and signals of success, predictability, and value. If your website is a collection of outdated PDFs and vague practice group pages, you're invisible. If your content is specific, current, and consistently structured, agents surface you.

This is Generative Engine Optimization. This is Answer Engine Optimization. It's not SEO 2.0. It's competitive survival.

What You Can Do Monday Morning

This isn't a five-year transformation plan. It's a decision about where to start.

For BD: Read your last proposal like an AI would. Ask yourself whether a machine could understand why you're better. If the answer is no, rewrite it. Standardize experience data. Stop letting expertise live in three partners' heads and one stressed BD manager. Build proposal assembly into a system, not a scramble. Assume the client's agent reads your pitch first, while the GC is in another meeting. Shorter. Clearer. More specific. Less "trust us."

For Marketing Coordinators: Everything you upload is now data. Bios, case studies, deal lists. If it's vague, inconsistent, or outdated, the agent learns the wrong thing. Fix structure before style. Headings, bullets, clear summaries. Agents don't care about clever; they care about extractable. When creating content, ask better questions upstream: What was the outcome? What made this matter hard? Why did the client choose you? That's not busywork. That's signal.

For Marketing Leaders and CMOs: Decide what you want agents to believe about your firm. Not slogans. Signals. What problems do you solve? Which industries? At what level of complexity? Agents don't handle ambiguity well. Break the silos, because if BD, digital, and content aren't aligned, your firm shows up fragmented. Turn content into a system, not a publishing calendar. Own agent readiness as a growth mandate. You're not just promoting lawyers anymore. You're designing how the firm gets chosen, by humans and machines.

The Whole Truth

In the past, clients chose law firms based on relationships and reputation. In the present, they choose based on websites, rankings, content, and price.

In the (near) future, clients will choose based on what their agents tell them about you.

And their agents will tell them the truth: the whole truth, summarized cleanly, and benchmarked against every competitor. The firms that win will be the ones who made themselves legible, structured, and trustworthy, not just to people, but to machines.