George Hannah
banner
georgehannah1.bsky.social
George Hannah
@georgehannah1.bsky.social
Writing on Legal AI
Follow me, George Hannah, for more on legal AI.

Join thousands of legal professionals receiving the latest on AI.

Subscribe here for free: bestpracticeai.substack.com/
January 26, 2026 at 1:13 PM
AI won't simply kill the billable hour. Even firms using alternative fees still track time internally. The real challenge isn't changing what clients are charged - it's finding a new way to measure lawyer performance and capacity.

And that's a much harder problem to solve.
January 26, 2026 at 1:13 PM
My takeaway:

Firms use billable hours for bonuses, promotions, and performance reviews. It shows partners who's overworked, who's underutilized, and where to allocate resources. Remove it and you eliminate the primary metric driving firm economics and workforce management.
January 26, 2026 at 1:13 PM
2026
→ Gen AI completes tasks in minutes that took hours, fundamentally challenging time-based billing
→ Despite decades of criticism, 97% of firms still bill by the hour
→ Growing tension between traditional model and value-based pricing as profession faces transformation
January 26, 2026 at 1:13 PM
2007
→ Scott Turow publishes "The Billable Hour Must Die" in ABA Journal
→ Evan Chesler (Cravath presiding partner) writes "Kill the Billable Hour" in Forbes
→ Peak criticism from both elite practitioners and clients demanding alternatives
January 26, 2026 at 1:13 PM
1980s
→ The American Lawyer publishes firm rankings by hourly rates, turning rates into status symbols
→ Firms begin requiring minimum annual billable hours (1,800-2,000+) tied to advancement
→ "Big-firm star system" and competitive era begins, hours requirements rise steadily
January 26, 2026 at 1:13 PM
1940
→ Hale and Dorr (now WilmerHale) becomes first major firm to fully implement 6-minute increment billing
→ Smith describes it as "fair, logical, transparent" and "especially pleasing to businessmen"
January 26, 2026 at 1:13 PM
1914
→ Reginald Heber Smith at Boston Legal Aid Society creates the first timesheet system
→ Developed with help from Harvard Business School (not IBM) to manage 2,000 cases on limited budget
→ Initially just an internal productivity metric, not yet used for billing clients
January 26, 2026 at 1:13 PM
It's the idea. It's the execution. It's understanding the pain point well enough to do something about it.

I'll be doing a deep dive into the rise of these high agency "legal-quants" in in my newsletter this Sunday.

Don't miss it: bestpracticeai.substack.com/
January 23, 2026 at 12:59 PM
The ability to see a problem, pick up unfamiliar tools, and actually build something.

Even if it's imperfect.

If these functions can be whipped up in days, the moat in legal tech isn't the technology.
January 23, 2026 at 12:59 PM
Screenshare demos are popping up everywhere.

My takeaway:

The most valuable skill won't be technical knowledge or even legal expertise.

It'll be being a high agency individual.
January 23, 2026 at 12:59 PM
So far we've seen lawyers from top firms like Sullivan & Cromwell, Clifford Chance, and other top firms create:

→ UCC-1 financial statement batch drafters
→ Tabular review functions
→ High Court case trackers
→ CMA decision trackers
→ Case chronology generators
January 23, 2026 at 12:59 PM
No client data. No compliance headaches. Just pure experimentation with dummy data.

He even organized a hackathon in partnership with Manus AI, providing free credits for participants to build their own vibe coded products.
January 23, 2026 at 12:59 PM
He built an entire community around this movement – called "legal quants."

Lawyers, law students, and legal innovation specialists testing technology themselves.
January 23, 2026 at 12:59 PM
Jamie Tso vibe coded a tabular review function using Google's Gemini 3 and shared it on LinkedIn.

The post went viral within the legal industry.

But Jamie didn't stop there.
January 23, 2026 at 12:59 PM
→ His view that data quality, not model capability, will drive legal AI consolidation in 2026
→ Where enterprise legal AI actually sits in the broader market.

Link in the comments: open.substack.com/pub/bestpra...
January 22, 2026 at 1:07 PM
→ Why he thinks in-house teams are structurally different from law firms - and why that matters for product design
→ Whether European data privacy regulations create defensibility or just higher barriers to entry
January 22, 2026 at 1:07 PM
And the early results suggest something's working: 80% of their clients come inbound rather than through outbound sales.

We discussed:
January 22, 2026 at 1:07 PM
Instead of building another generalist AI tool, he co-founded LegalFly in 2023 - focused exclusively on in-house and procurement teams at regulated enterprises.

It's an interesting strategic choice, especially as we're seeing consolidation across legal tech.
January 22, 2026 at 1:07 PM
I spoke with Ruben Miessen, Ruben Miessen, co-founder at LegalFly who spent years leading the product teams at Tinder and Match Group, where sitting through weekly legal committee meetings helped him develop a deep understanding of the processes that caused bottlenecks.
January 22, 2026 at 1:07 PM
Join thousands of legal professionals getting the latest on Legal AI

Subscribe here: bestpracticeai.substack.com/
January 21, 2026 at 1:21 PM
My takeaway:

With lots of competition in the space, the winners will need to either:

• Build vertical-specific moats (M&A, employment, SaaS)
• Own the full contract lifecycle, not just review
• Integrate so deeply into workflows they become unreplaceable
January 21, 2026 at 1:21 PM