iiinova — where your digital self keeps evolving

OK, dinner's ready so I'm going to leave it there. But if this resonates with you at all, you should check out my Agile Attorney Podcast. I talk about this stuff A LOT. And I just did an 11-episode "Agile Attorney 101" series (starting with episode 101) that covers a lot of this ground. 17/

100

They also almost inevitably lead to process standardization. Which is a form of setting a quality standard. Which is essential for a functioning QA process. So a little bit of tool adoption can actually kick off a cascade of improvement. Also, visible work is less likely to fall through cracks. 16/

100

That's especially hard to do with knowledge work because it is hard to see. But this is why I'm such a huge fan of #kanban boards in law practices and have been beating that drum for over a decade. Other knowledge workers (like technologists and salespeople) have used them for years. They work. 15/

100

It starts with that honest reckoning with capacity. We need for the legal industry to allow it to be acceptable to say "my practice is full," or "my plate is overloaded," or even just "no." Overburdened systems, and overwhelmed people, make mistakes. Full stop. We have to know what "enough" is. 14/

100

So what can we do about it? Well, a better understanding of how gen-AI works, what kinds of risks using it entails, and developing processes to avoid or mitigate those risks is a start. But I think we need something deeper, and ultimately more cultural. 13/

100

Whats worse, many lawyers (and teams) are turning to AI to help with the QA function, but we've seen multiple cases now where that last bit of AI-assisted quality checks have actually introduced errors into previously good documents. Because we don't understand QA as a discipline, we miss those. 12/

100

Another problem is quality assurance. Compared to other industries, legal already has a relatively immature approach to QA (lots of individual, subjective review; few set quality standards and effectively zero automated testing). But AI use requires careful QA review, which takes up capacity. 11/

100

But also AI exacerbates our existing optimism bias, so we think we'll be able to do WAY more with AI than we could before, but really we can only do a BIT more. And even that bit more comes with a cost: More individual tasks lead to more switching costs between tasks and administrative overhead. 10/

Even when it works, AI is actually contributing to the overload problem (though it rarely works as well as the marketing promises). Partly this is because we have to adjust to a new "set point" on our AI-augmented capacity, it is harder than before to have that honest reckoning. 9/

Add to that the incessant drumbeat of legal marketing and legal technology businesses encouraging lawyers to take on even more work (with their help of course). And that was true before AI added a whole new layer of magical thinking to the mix. 8/

One of my catchphrases is that lawyers need to have "an honest reckoning with capacity," but this is easier said than done. Law firms don't traditionally use tools and systems that help us gauge capacity, load balance work, and even say no to new work when our current practice is full. 7/

Even those who aren't money motivated (trying to work the heavy hours load) often feel pressure to take on more matters simply because there are so many people who need help. Plenty of lawyers get overloaded because of their good intentions. 6/

100

Partly it is lawyer culture: we celebrate the workhorses, reward the midnight-oil-burners, and encourage taking on pro bono or other public interest work on top of already full workloads. It is extremely easy to wind up in a practice that overloads you with work. 5/

100

On the one hand, I get why it is happening. Lawyers are unquestionably overburdened for several reasons. One of them is structural: There simply aren't enough lawyers in the US to meet the demand for legal help. Law school classes shrunk post-great-recession combined with grey-wave retirements. 4/

100

But think of all of the contracts, business documents, estate plans, etc. that almost certainly have been drafted with AI — and therefore with high potential for hallucinations — over the past 2+ years. By lawyers and parties. We won't start seeing cases show up on those for years or decades. 3/

100

First off, the court cases we're seeing are just the tip of the iceberg. Feedback loops in legal are notoriously slow. We're only hearing about the hallucinated court filings because they have to be timely. Of course the time-pressure is partly why the lawyers are seeking shortcuts. 2/

110

John E. Grant ⚖️ @agileattorney.com · 29/03/2026

As someone who helps overwhelmed and overloaded lawyers get their workflows and processes dialed-in, This whole bad-AI-use epidemic just kills me because it is so totally avoidable. A few thoughts... 🧵

102

The countdown begins for his inevitable post: “I never wanted my humble attempt at promoting self-improvement to be political, but now that I’m being persecuted I have no choice but to go full MAGA.”

John E. Grant ⚖️ @agileattorney.com · 26/03/2026

I will say that Ted McGinley on Shrinking more than atones for his Revenge of the Nerds character

John E. Grant ⚖️ @agileattorney.com · 26/03/2026

omg, does that mean that Mar-a-largo face is really just an extended cut of post-bad-makeover Ally Sheedy in The Breakfast Club?

John E. Grant ⚖️ @agileattorney.com · 22/03/2026

An invitation to play would be almost enough to make me rethink my planned absence from TechShow this week.

John E. Grant ⚖️ @agileattorney.com · 18/03/2026

I don’t have to tell you this, but people can suck. It was a really great piece and I truly appreciate your positive framing and your personal perspective. Ignore the haters (as best you can)

011

John E. Grant ⚖️ @agileattorney.com · 17/03/2026

Respondent made a prescient, if unsuccessful, attempt at the “Marc Andreesen” defense: “It is evident from the record that Respondent has engaged in zero self-reflection as to his actions…” (This whole thing is wild; a stupid spat over double-sided printing led to a 2-year license suspension.)

000

John E. Grant ⚖️ @agileattorney.com · 16/03/2026

I will use that information to dream that they’ve seen the error of their shadow docket ways. But I won’t hold my breath.

110

John E. Grant ⚖️ @agileattorney.com · 13/03/2026

You know a story is well written when you know how it ends (because it is how that kind of story always ends), but still get sucked in. My only beef is why name Terry Rozier but not Chauncey Billups?

000

John E. Grant ⚖️ @agileattorney.com · 13/03/2026

I feel ya. Portland is dealing with our 3rd day in a row of this:

Weather radar showing the atmospheric river of rain over Oregon and Washington and stretching west over the Pacific Ocean

011

John E. Grant ⚖️ @agileattorney.com · 13/03/2026

Good point! Even if everything Miller says in this clip is true, the administration will wind up confused and angry that the American people still won’t give them accolades because the vast majority of us never wanted any kind of war — even a lopsided one — in the first place.

081

John E. Grant ⚖️ @agileattorney.com · 11/03/2026

Mine too! Like, you can pay $149 to still do sycophancy but not have a tripping hazard at the end of your legs all day. Or look stupid. Unless they are beyond doing sycophancy and have graduated to asceticism?

John E. Grant ⚖️ @agileattorney.com · 10/03/2026

That thread! To answer the question: AI has broken people’s brains. There’s this new societal-corporate expectation that everyone should be able to do more with less, because AI. Except there’s no consideration for the quality ramifications, or what the legitimate efficiency gains actually are.

Shambling Corpse of Randy Herman @randyhermanlaw.com · 10/03/2026

Reposted by John E. Grant ⚖️

Hello and welcome on the Terry Sanford Federal Building in beautiful downtown Raleigh, NC where I will be live-posting a show cause hearing starting at 4:00 ET

661121223

John E. Grant ⚖️ @agileattorney.com · 10/03/2026

Paging @juliedicaro.bsky.social & @chrisgeidner.bsky.social (and I continue to appreciate that @maxbernstein.bsky.social always includes links to actual court documents).

020

John E. Grant ⚖️ @agileattorney.com · 08/03/2026

Right up ‘til it’s December and kids are walking / riding to school in the dark. Not saying you’re wrong — I like the later sunshine (mostly) too. But there are legit arguments for permanent standard time instead. British Columbia is gonna let us all know what they think by this time next year.

300

Definitely a step in the wrong direction. But it is consistent with language I regularly see from people who are overly immersed in their own techno-utopian koolaid bubble. Don’t get me started on legaltech, where hyperbolic claims are making me lose respect for many folks I’ve trusted for years.

12 years after branding myself the “Agile Attorney” I get the attractiveness of the word. But, FWIW, the capital-A Agile approach is fundamentally about making things work at a human level. Maybe they’d argue they’re doing that? Feels like we are going to be in this uncanny valley for a long time.

100

Makes sense. Spending time with humanoids is obviously way less weird than dealing with robots. Especially ”agile” ones. 🙄 Seems like the kind of decision made by leadership that definitely enjoys eating a hamburger product.

110