White & Case announced a global rollout of Legora as its primary legal AI platform late last year.

43 offices.

29 countries.

But here's the detail that matters: they called it a "strategic partnership," not a software purchase.

Linklaters did something similar before that.

Goodwin invested in Harmattan AI for advisory workflows.

The pattern isn't "big firms buying AI tools."

The pattern is big firms selecting vendors who did deep diagnostic work before proposing anything.

That distinction matters more than the platform they chose.

Here's the sequence I see playing out at mid-market firms.

Someone on the leadership team sees a compelling demo.

The vendor's slide deck looks sharp.

The firm signs a contract.

Six months later, roughly a third of lawyers are actually using the platform.

The rest quietly went back to their old workflows.

What happened?

The vendor's "discovery phase" was a single call with a partner and maybe someone from ops.

They never talked to the associates doing the actual work.

They never saw the daily friction points that leadership doesn't know about.

So the recommendations didn't match reality.

And the ROI case that looked bulletproof in the presentation fell apart when actual usage data came in.

We've built a Vendor Evaluation Checklist with 20 questions that separate serious AI implementation partners from feature-demo vendors.
Print it.

Ask every potential partner.

If you're evaluating an AI implementation partner, here's what their process should include.

1. Interview the front lines.

Not just the managing partner.

Not just the head of legal ops.

A serious vendor talks to an estimated 6 to 10 people across your firm: partners, associates, paralegals, ops staff.

The partner sees the destination.

The junior associate sees every pothole in the road.

2. Map the workflow visually.

Whiteboards, Miro, whatever the tool.

The point is making invisible friction visible.

When you lay out how work actually moves from intake to completion, the bottlenecks become obvious.

3. Quantify the cost of inefficiency.

This is where it gets concrete.

A good vendor calculates: time wasted per employee, multiplied by the number of people affected, multiplied by their fully loaded hourly cost, multiplied by weeks per year.

That formula turns vague frustration into a specific number.

4. Propose solutions aligned to what they found.

Not to their product roadmap.

To your actual problems.

The recommendation should be driven by the diagnostic, not by what the vendor happens to sell.

5. Package it into a defensible roadmap.

A 15 to 20 slide deck that your partnership can scrutinize.

Quick wins first (builds adoption momentum), bigger implementations phased over time.

ROI assumptions that are conservative and transparent.

If a vendor can't do all five, they're pitching.

They're not diagnosing.

There's a meaningful difference between two common vendor models in the market right now.

The strategic audit typically runs roughly $8,000 to $12,000 over about 28 days.

It includes 6 to 8 interviews, validated discovery across staff levels, and a board-ready presentation with defensible ROI.

The rapid scan runs roughly $2,500 to $3,500 over about a week.

It's usually 1 to 2 calls with whoever's available, and the recommendations are based on what leadership thinks the problems are.

Both have a place.

But the trade-off is real.

The rapid scan is faster and cheaper, but it relies on self-reported pain points.

If leadership has blind spots (and they usually do), those blind spots carry straight through to the recommendations.

The strategic audit costs more upfront, but it catches the problems nobody told you about.

Before you commit, ask these directly.

"How many people on our team will you interview, and at what levels?"

If the answer is fewer than 5, or only leadership, that's a flag.

"Will you talk to junior staff doing the day-to-day work?"

This is the question that separates real discovery from executive hand-waving.

"How do you validate that the problems you identify are real, not assumed?"

Look for: "We triangulate across multiple roles."

Red flag: "We'll take your word for it."

"How do you model ROI?"

Good answer: Conservative assumptions, multiple scenarios, sensitivity analysis.

Red flag: A single optimistic number with no methodology behind it.

"What does the final deliverable look like?"

Good answer: A partner-grade presentation with objection handling and phased implementation.

Red flag: A generic slide deck.

Here's the part most firms don't think about during vendor selection.

AI automations aren't static.

Prompts need updating as models improve.

Security requirements shift.

New use cases emerge as the technology advances.

A vendor who did deep discovery already knows your firm's workflows, culture, and internal dynamics.

That institutional knowledge compounds over time.

They can proactively suggest the next automation opportunity before you even know to ask for it.

The firms getting compounding ROI from AI are the ones with ongoing partners, not one-off implementations.

This is how we structure ongoing partnerships at Cyberaktive: the audit isn't the end of the engagement, it's the foundation for a long-term relationship that keeps delivering value.

When evaluating an AI vendor, what's your biggest concern?

A) Will they actually understand our firm's specific workflows?
B) Can they prove ROI before we commit to a large investment?
C) Will the implementation timeline disrupt our current operations?
D) Will our team actually adopt and use what they recommend?
[Reply with your letter choice] I'll share the results in the next edition.

[Reply with your letter choice] - I'll share the results in the next edition.

The firms winning with AI partnerships right now aren't the ones who found the flashiest vendor.

They're the ones who demanded real discovery before signing anything.

Five steps, multiple interviews, quantified costs, a defensible roadmap.

That's the bar.

If your vendor can't clear it, find one who can.

Hit reply and tell me: what was your experience with AI vendor discovery?

I'm genuinely curious whether this matches what you've seen.

— Liam Barnes

We help law firms evaluate AI readiness and build implementation strategies that actually match how your firm works.

Grab some time to chat

(if you don’t see a suitable time, just shoot me an email [email protected])

How Did We Do?

Your feedback shapes what comes next.
Let us know if this edition hit the mark or missed.

Too vague? Too detailed? Too long? Too Short? Too pink?

👍 Loved it | 👎 Not for me | 💡 I Have a Suggestion

OR: [email protected]

Was this week’s newsletter forwarded to you?

Sign up, it’s free.

LAST WEEK’S POLL