Credibility as distribution

Anthropic and Microsoft are coming for legal. This is how you compete.

The existential question for every legal AI challenger right now is not about technology. Anthropic and Microsoft are moving into legal, and they will close most feature gaps over time. A challenger that wins on product alone is running a race it will eventually lose.

What makes this harder is that platform companies don’t just have resource advantages. They have distribution advantages, and they carry a significant amount of institutional trust already. Their brands are recognizable. Their contracts are already in place at most large firms. A lawyer who has spent years using Microsoft products doesn’t need to be convinced to try a Microsoft legal tool. That’s a meaningful head start that no challenger can simply outspend.

Robust distribution is a known competitive advantage in tech. VCs talk about it often. How do you scale sales (sorry, “GTM”), how do you enable your team, how do you build pipeline. And most of that conversation borrows from the broader software playbook: more layers & middle managers, more activity, automated outreach, tighter process.

That thinking has its limits in legal. Spoiler Alert: If you have a GTM team with healthy pipeline, but can’t seem to convert it to revenue, that’s usually a signal that sales activity isn’t the load bearing constraint. Something else is.

The crux of the problem is this: Sales activity and credibility are different things, and only one of them is the bottleneck. You can run a tight process, hit your call targets, have beautiful Salesforce dashboards, and still find that revenue fails to appear. Lawyers are not evaluating your GTM motion. They are evaluating whether you have earned the right to be taken seriously. No amount of pipeline activity closes that gap.

The real limitation is trust. Lawyers extend it to people and institutions they believe genuinely understand their work, their standards, and what happens when something goes wrong. That bar is specific, and it takes time to clear. Platform companies cannot manufacture it at the level of the profession.

For a challenger, it is the one unfair advantage worth building.

I. The evidence

The storied companies that built formidable positions in legal got there by earning the profession’s trust at scale. Most founders/investors miss something fundamental when they first encounter this market: law does not separate easily from the people who practice it. A lawyer’s judgment, their reputation, and the tools they rely on are bound together in a way that has no clean parallel in other areas of tech. When something goes wrong with a legal tool, it lands as an actionable lawsuit and potentially a professional liability problem.

That changes how lawyers evaluate vendors. It changes what it takes to win their confidence.

The standard is even higher when we’re talking about core legal work, or judgment-based tasks. This is where a lawyer’s training and professional standing are directly implicated in the output. It’s less obvious than it sounds. Legal research, for example, might seem like a straightforward support function. But finding the right case requires judgment: does this precedent actually apply, how does it interact with the facts, what weight does it carry.¹ That is legal reasoning. Keeping case files organized or routing a contract for approval is different. The distinction is not about complexity. It is about whether professional judgment lives inside the task. (more on this distinction later)

Legal AI, and specifically the recent generation of generative and agentic AI, sits squarely inside this territory. The trust bar is higher here than in other traditional legal tech categories such as CLM, e-discovery, or practice management. Legal AI is asking lawyers to outsource some level of professional judgment to a robot. The companies that have historically done an effective job earning that trust have done so in ways that are specific, observable, and worth examining closely.²

A. Westlaw

Westlaw (e.g. Thomson Reuters) is the oldest example and risks feeling dated, but it points to something deeper and undervalued. Westlaw understood the psychology of the profession well enough to embed themselves before lawyers even graduated. The student rep program meant Westlaw was the first research tool lawyers ever touched. They give free access to federal clerks, among the heaviest power users, to lock in that trust. I know this because I wasn’t a heavy Westlaw user until after my clerkship. By the time I made it to S&C and was given the option of choosing Westlaw vs. Lexis, it was a no brainer. It didn’t feel like a typical vendor relationship.

B. Casetext

Casetext went further in a different direction. You can find its story detailed in many places, but as someone who knew several people at the company, it was obvious to me that lawyer DNA was baked into every function: sales, marketing, product. One of my closest friends from law school was a successful sales rep there; while some elements of their GTM motion looked like the SaaS playbook, many of them did not. And when they began to release AI features that got to core legal work, ie. Prioritizing cases that were relevant for specific legal points, the GTM motion relied very heavily on the ability of their sales reps (who were lawyers) conveying the trust through credible explanations of the product.³

C. Harvey

This is the most recent example that is still playing out. Harvey came in with significant backing and elite firm endorsements, but there were plenty of well-funded legal AI companies that weren’t able to achieve Harvey has. What appears to have differentiated them was the early GTM motion: hiring ex-Biglaw lawyers, and letting client press releases do the work of social-proof based distribution rather than running a traditional conference plus outbound motion. The result (saw this on social media just yesterday) is a presence in over 50% of AmLaw firms in a remarkably short period of time.

Note on ALSPs as a relevant example

I’ll share a brief word on this since this is the space I play in, and there’s quite a bit of history here. Both Axiom and Ontra (called InCloudCounsel at the time), in their early days, leaned very heavily on trust-based GTM. Both hired teams of former lawyers, often with Biglaw experience or from top schools, as salespeople. It wasn’t clear that they had to take this direction—most other existing ALSPs didn’t, and were still able to scale. But because unlike those other ALSPs, Axiom/Ontra focused on core legal work—so they had a higher bar to clear to earn the client’s trust that they could handle those matters competently.⁴ With the benefit of hindsight, it was the right decision.

II. Legal-adjacent work

My argument weakens as you move away from products/services focused on work that relates to core legal judgment. These adjacent categories are worth naming quickly:

• CLM (e.g. Ironclad, Icertis, Docusign): Primarily about contracting process. Routing approvals, managing contract repositories, tracking obligations. The lawyer is present in the core workflow but the tool—with the exception of any AI add-ons--is not exercising judgment on their behalf.

• Ediscovery (e.g. Relativity, Everlaw, Disco): Primarily about basic-level filtering of documents. The buying decision is typically shared across legal, IT, and litigation support. The value proposition is operational efficiency at scale, and the sales motion reflects that.

• Practice management (e.g. Clio, MyCase, Smokeball): Supports the operational organization of law practice rather than substantive practice of law. Billing, scheduling, client intake. Closer to vertical SaaS than to anything that requires professional trust to sell.

The nuance worth holding onto is that even these companies have not been able to ignore the legal market entirely on its own terms. Each of them, at some point in their growth, had to develop cultural fluency with lawyers: through the profile of their hires, the structure of their GTM teams, the way they show up in conversations with firm leadership. The trust bar is lower in these categories, but it far from zero. Lawyers are the buyers or at least the influencers in almost every legal tech purchase, and they carry their professional instincts into every vendor evaluation regardless of what is being sold.

What about legal operations as a buyer?

Legal ops has become an increasingly influential buyer across the legal tech market, particularly for the adjacent categories described above. Legal operations professionals often come from outside the law; they are process-oriented, metrics-driven, and in many cases deliberately non-lawyer in their outlook. That is part of what makes them effective in their role. But it also means the trust signals that matter elsewhere in this market apply differently here. Selling to legal operations looks more like a sophisticated SaaS sale than a profession-first motion.

Many legal tech companies are selling to both lawyers and legal operations simultaneously. Legal operations often controls budget and procurement. Lawyers control adoption and renewal. You need both. But one buyer has to have primacy in how you build your GTM motion, your hiring profile, and your product positioning. Try to optimize equally for both and you end up speaking clearly to neither. The founders who navigate this well are honest early about which buyer they are fundamentally building their GTM motions for and treat the other as an important secondary relationship.

IV. The absence of a counterexample

It is worth noting first that there are not many companies selling something that genuinely touches core legal work. The market is thinner than it looks from the outside. That may partly explain why the examples are few. And this is what I believe is the most compelling aspect of my argument.

It is genuinely difficult to identify a company that took a traditional SaaS GTM motion and won in core legal work.⁵ The kind of significant penetration that Harvey has achieved, or that Westlaw built over decades. That company is hard to find. The absence is worth considering. Is that the absence of evidence? Or is evidence of the absence of evidence, evidence itself? (Don’t answer that question)

My view: The examples that come closest to being counterexamples weaken on inspection. Legal SaaS unicorns: Each of them has scaled meaningfully, but they also sit neatly within the adjacent category already carved out. The work they touch is operational, not judgment-based. Instead of disconfirming the thesis, they instead confirm the boundary.⁶

V. What trust-scaling looks like in practice

This is where challengers have a structural advantage the giants cannot easily close. The patterns below are observable across the companies that have been able to scale meaningfully in this market. Take these as observations about GTM structure, not rules or requirements. The weight you put on each one depends on your ICP and your product/service.⁷

A. Pedigree

The profile and background of GTM hires sends a signal before anyone has said a word about the product. When a former Biglaw associate or a HLS grad gets on a call with a senior decision maker, something shifts immediately. (And remember, by the way, that not all Biglaw firms are on the same level in the hierarchy.)

As a result, the relationship does not start as vendor and customer. It starts closer to peer and peer. That is a fundamentally different foundation to build from, and it is difficult to manufacture through training or messaging alone. Lawyers are evaluating credibility quickly and instinctively, and the right hire clears that bar before the conversation has properly begun. The specific pedigree that matters will shift with your buyer by the way. Some care about school, others care about former firm, and still others care about in-house experience. There are different badges of honor but the critical element seems to be having the right badge.

The underlying principle holds across all of these examples: the client should immediately grasp, from background alone, that this person has been where they are—and is an equal.

B. Hiring profile and temperament

The best client-facing people in this market have a particular blend. It is rare and hard to build. They carry genuine sales or customer success instincts; they are motivated to win business, they follow up, they move things forward, while retaining the measured, neutral polish that the legal profession expects from people it takes seriously. (This, by the way, cannot be solved for by just hiring ex-lawyers; many lawyers are not wired this way)

The right combination is partly about individual personality and partly about how someone fundamentally views the world. Lawyers, particularly at the upper end of the market, tend to operate—even thrive—in shades of gray. They are trained to hold complexity, to avoid overstatement, and to be skeptical of anyone who seems too certain. A client-facing person who sees the world in black and white, or who defaults to enthusiasm and urgency as sales tools, will feel off to a senior lawyer even if they cannot articulate exactly why.

Temperament compatibility with legal buyers can be developed at the margins, but the underlying disposition has to already be there. Hire for it. As a rough guide: the higher up the market you are selling, specifically senior partners, high-end GCs, and sophisticated legal teams, the more the neutral, measured end of the spectrum matters. Broader markets tend to tolerate a more traditional GTM temperament, though the baseline of professionalism the legal market expects never fully disappears.

C. Alignment in communication style

Attention to detail, responsiveness, and the quality of written communication are the accumulated texture of every interaction a client has with your company. They are observable, they compound over time, and they are among the clearest signals lawyers use, often without naming them, to decide whether a vendor genuinely understands their world. This dimension can be trained, though changing ingrained communication habits is genuinely hard. What you are looking for is a blend: people with a bias for action who also default to careful, precise communication rather than treating it as a tax on speed.

The calibration here follows the buyer. The more upper-end the client, senior partners, high-end GCs, and sophisticated legal teams, the more that responsiveness, detail orientation, and respect for their time become non-negotiable rather than differentiating. (Hint: 99% of enterprise CLOs are former Biglaw associates, so best to optimize for this cohort’s preferences)

At that level, a slow response or a carelessly written email does real damage. It signals that you do not understand what their time is worth or what their standards are. Lower in the market there is more tolerance for imperfection, but the underlying signal remains the same: every interaction is communicating, quietly, whether you are one of them or not.

VI. Conclusion

The giants will continue to make inroads into legal. Anthropic and Microsoft are already moving, and they will not be the last platform companies to see the opportunity. What their arrival ultimately means for the broader ecosystem, for pricing, for consolidation, for which categories remain viable for challengers, remains to be seen.

What is clearer is that there is a ceiling on how deeply they can embed themselves in the profession. The giants bring mass distribution, established brand, and significant product resources to the table. Those are real advantages and challengers cannot fight them on that ground. Competing on distribution or resources against a platform company is a losing game. The ground has to shift.

The asymmetry worth building is the one the giants structurally cannot close. Showing up as an insider rather than a vendor. That advantage does not announce itself with a large fundraising round or a tier one VC on the cap table. It shows up in the compounding of metrics over time (lower CAC, higher net retention, repeatability of sales) in the depth of relationships where a client is genuinely advocating for the product. That kind of trust has always been built slowly, on the profession’s own terms.

That part has not changed despite all the noise about AI. And I would not be surprised if it continues to be the case as time marches on.

1. This is exactly why we are seeing an onslaught of firms getting in trouble for AI hallucinations. Turns out using tech to automate some parts of the case selection process has led to some pretty terrible results. See If Sullivan & Cromwell Can File an AI-Contaminated Brief, Is Risk-Proof Use of the Technology Even Possible?

2. A parallel worth noting from outside legal innovation altogether. Legal headhunting, specifically placing senior lawyers into firms and in-house roles, is as close to core legal judgment as a services business can get without actually practicing law. To do it well, you need a deep understanding of how legal skillsets are actually composed, how to read pedigree signals accurately, how to stack rank firm reputations across practice areas and markets, and how to reason through the tradeoffs a candidate or client is making. Firms like MLA and others that have built dominant positions in this space are run by people with genuine standing in the profession, because the work requires it. You cannot SaaSify that judgment. The market has consistently penalized firms that tried to treat it like a generic search business. The same dynamic, in a category most people would never think of as legal tech at all.

3. I think vLex fits this model but am admittedly less familiar with their GTM approach pre-Clio acquisition

4. Other types of ALSPs, especially those oriented around price-sensitive “commoditized” work—doc review, due diligence processing, and high-frequency surface-level work—operate on a different logic entirely. UnitedLex, Integreon, Quislex, and others in that tier compete on operational efficiency and price. The buying decision is closer to procurement than professional trust. That is a legitimate and large market. But it is a structurally different one, and conflating the two leads founders to draw the wrong lessons from ALSP history.

5. If a good counterexample comes to mind please share with me in the comments, as I would love to reconsider my position (or add nuance) as needed.

6. Others in the legal AI space have grown and raised significant capital but have not yet demonstrated the same depth of penetration in core legal work at the top of the market. That may change. But the pattern so far holds.

7. Related: Some early-stage companies have tried to shortcut trust-building by convincing well-known lawyers to become clients , borrowing personal brand rather than earning institutional trust organically. This does not scale. You are growing on borrowed trust instead of the real thing—and will find scaling bottlenecks down the road.