
Walk the halls of any legal tech conference today and you’ll trip over dozens of AI startups promising to revolutionize law practice. Each booth features the same pitch: “We’re ChatGPT, but for lawyers!” The valuations are astronomical. The demos are slick. The value proposition? That’s where things get murky.
The legal tech world has become a gold rush, and most prospectors are selling fool’s gold. Every week brings announcements of new AI tools that are nothing more than thin wrappers around OpenAI, Claude, or Google’s models. These companies take a general-purpose language model, add a legal-sounding name, maybe some prompt engineering, and suddenly they’re worth millions. Or in Harvey AI’s case, $5 billion.
Let that sink in. Harvey AI, which essentially provides access to large language models with some legal flavoring, commands a valuation that exceeds the GDP of some small nations. For what? Features that a $20 monthly Claude subscription or Google Workspace already provides?
The proliferation is staggering. Legal AI tools are multiplying like rabbits. Contract review AI. Research AI. Brief writing AI. Deposition prep AI. Due diligence AI. Each claiming to be purpose-built for legal work, each demanding premium pricing, each essentially doing what you could accomplish with direct access to the underlying models.
This isn’t innovation. It’s arbitrage. These companies position themselves between law firms and the actual AI providers, adding minimal value while extracting maximum fees. They’re middlemen in expensive suits, and law firms are falling for it.
The problem runs deeper than overvaluation. Law firms signing multi-year contracts with these vendors are possibly making a critical strategic error. They’re betting on intermediaries in a rapidly evolving market where the underlying technology improves monthly. Today’s cutting-edge legal AI wrapper becomes tomorrow’s obsolete interface. Meanwhile, firms are locked into contracts, paying premium prices for increasingly outdated access to technology they could use directly.
Why This House of Cards Will Collapse
The legal AI bubble mirrors every tech bubble before it. Massive valuations built on thin value propositions. Investors pouring money into companies whose entire business model depends on other companies’ technology. Law firms, traditionally conservative with technology adoption, suddenly throwing caution to the wind because everyone else is doing it.
The fundamental question every firm should ask: What unique value does this legal AI vendor provide that I cannot get from direct access to Claude, GPT-4, or Gemini? Strip away the marketing speak and legal jargon. Look at the actual functionality. In most cases, you’re paying a massive premium for prompt engineering you could do yourself.
Consider document review. Multiple legal AI companies offer document review solutions powered by large language models. Their secret sauce? Prompts that tell the AI to focus on legal concepts. Any competent lawyer with an afternoon to spare could create similar prompts. Yet firms pay thousands per month for this “specialized” access.
Research platforms fare no better. They ingest legal databases and wrap them with AI interfaces. The AI doesn’t understand law any better than the base model. It’s just been prompted to format responses like legal memoranda. Again, something any associate could configure with basic prompt engineering skills.
The Better Path Forward
Smart firms should reject the vendor gold rush and build internal AI competency instead. This doesn’t mean avoiding AI. It means being strategic about implementation and skeptical about vendors selling repackaged access to technology you can use directly.
Start with direct subscriptions to major AI platforms. Google Workspace with Gemini costs a fraction of specialized legal AI tools. Claude Pro provides powerful language processing for less than most lawyers bill in an hour. Google’s Notebook LM is a favorite of mine. These platforms improve constantly, and you benefit immediately from upgrades without renegotiating vendor contracts.
More critically, invest in people, not platforms. Hire or develop internal futurists and explorers. These team members should understand both legal practice and AI capabilities. Their job isn’t to build AI from scratch but to identify opportunities, test solutions, and separate genuine innovation from expensive vaporware.
Create an AI evaluation framework. Before signing with any legal AI vendor, your internal team should prototype similar functionality using direct AI access. If they can replicate 80% of the vendor’s offering in a week, you’re looking at overpriced middleware, not essential technology.
Establish small pilot programs. Test AI applications on real work with controlled scope. Learn what works, what doesn’t, and what your firm actually needs. This hands-on experience becomes invaluable when vendors pitch their solutions. You’ll spot the fluff immediately.
Build prompt libraries and workflows internally. The “secret sauce” of most legal AI tools amounts to well-crafted prompts and integrated workflows. Your team can create these without paying vendor premiums. More importantly, you’ll own and control these assets, adapting them as needs change.
The Reckoning Approaches
The legal AI bubble will burst. Not because AI lacks value in legal practice, but because the current vendor ecosystem is unsustainable. When firms realize they’re paying Harvey AI prices for Google Gemini functionality, the correction will be swift and brutal.
Firms committed to long-term vendor contracts will find themselves trapped, paying premium prices for increasingly commoditized services. Those who invested in internal capabilities will adapt seamlessly, switching between AI providers as technology evolves.
The winners in legal AI won’t be the firms with the biggest vendor contracts. They’ll be the ones who understood early that AI is a tool, not a solution. Who recognized that sustainable advantage comes from how you use technology, not which middleman you pay to access it.
Stop signing contracts with AI vendors promising to transform your practice. Start building the internal capacity to transform it yourself. The bubble is real, the burst is coming, and your firm’s future depends on being on the right side when it happens.
The legal profession stands at an inflection point. We can chase shiny vendors and astronomical valuations, or we can do what lawyers do best: think critically, evaluate evidence, and make reasoned decisions. The choice seems obvious. The question is whether firms will make it before their competitors do.