Michael Bloom is the founder and CEO of Praktio, the training platform law firms use to enable new associates to develop practical skills through training that feels like doing the work. Before that, he ran the Transactional Lab & Clinic at University of Michigan Law School and taught at Chicago Law.

In this conversation, Michael and Mike discuss if AI is quietly making lawyers worse at their own jobs. They cover the MIT and endoscopist studies on AI-driven cognitive decline, why "practice-ready" was always a strawman, why skills and judgment can't be pulled apart, why junior associates say feedback has dried up, and the looming "tragedy of the commons" as AI eats the bottom of the associate pyramid.

Where to find today's guest:
https://www.linkedin.com/in/bloomich/
https://praktio.com/

📌 The Redline Newsletter:
Get the latest news, trends, and tactics in legal AI—straight to your inbox.
➞ subscribe free: https://www.spellbook.legal/newsletter

Where to find Mike Whelan:
https://www.linkedin.com/in/mikewhelanjr/

💎 Spellbook:
Draft and review contracts 10x faster
➞ try it free: https://www.spellbook.legal

Timestamps:
0:00 - Introduction
0:36 - Meet Michael Bloom and Praktio
2:36 - Why using AI on a task makes you worse at that task
7:04 - Driver’s seat vs. along for the ride
9:08 - The microscope test: amplification vs. atrophy
10:12 - The London cabbies and the cost of standardization
13:56 - Who actually owns lawyer training today
17:52 - Judgment isn’t taught, it’s built
19:20 - Inside Praktio’s feedback-loop sandbox
23:56 - The tragedy of the commons in the associate pyramid
27:44 - Who foots the training bill in five years
34:52 - Connect with Michael

Nika Kabiri is a decision scientist who has spent 20+ years studying how consumers pick who to trust, including for legal tech companies like Clio and Avvo.

In this conversation, Nika and Mike discuss what happens in clients' heads when a lawyer discloses using AI, why it often erodes trust instead of building it, why Gen Z calls AI images "boomer art," and what lawyers should actually signal if they want disclosure to help.

📌 𝗧𝗵𝗲 𝗥𝗲𝗱𝗹𝗶𝗻𝗲 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿:
Get the latest news, trends, and tactics in legal AI—straight to your inbox.
➞ https://www.spellbook.legal/newsletter


𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝘁𝗼𝗱𝗮𝘆's 𝗴𝘂𝗲𝘀𝘁: https://www.linkedin.com/in/nikakabiri/


𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝗠𝗶𝗸𝗲 𝗪𝗵𝗲𝗹𝗮𝗻: https://www.linkedin.com/in/mikewhelanjr/

💎 𝗦𝗽𝗲𝗹𝗹𝗯𝗼𝗼𝗸:
Draft and review contracts 10x faster
➞ try it free: https://www.spellbook.legal

𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀:
0:00 - Introduction
1:13 - The Hank Green AI backlash
2:55 - Why law is a credence good
3:48 - How consumers really pick a lawyer
8:29 - Why AI reads as a cheap signal
13:57 - Why disclosing AI erodes trust
17:16 - Where AI use is fine, and where it’s not
19:55 - How to disclose AI use
25:00 - Billing, ethics, and the AI price question
29:37 - Connect with Nika

Joshua Lenon is the Legaltech hype man and Lawyer in Residence at Clio, where he's spent a decade turning Clio's data across 400k accounts into the industry's most-cited look at what's really happening inside law practices.

In this conversation, Joshua and Mike walk through what 10 years of the Legal Trends Report reveal, including what clients are really annoyed by, why $400/month is the affordable-legal price point hiding in the data, what the "$27K at risk from AI" number actually means and why solo lawyers are surprisingly slow to adopt AI.

📌 𝗧𝗵𝗲 𝗥𝗲𝗱𝗹𝗶𝗻𝗲 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿:

Get the latest news, trends, and tactics in legal AI—straight to your inbox.
➞ join free: https://www.spellbook.legal/newsletter

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝘁𝗼𝗱𝗮𝘆's 𝗴𝘂𝗲𝘀𝘁:

https://www.linkedin.com/in/joshualenon/
https://bsky.app/profile/joshualenon.bsky.social
https://www.clio.com/resources/legal-trends/

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝗠𝗶𝗸𝗲 𝗪𝗵𝗲𝗹𝗮𝗻:

https://www.linkedin.com/in/mikewhelanjr/

💎 𝗦𝗽𝗲𝗹𝗹𝗯𝗼𝗼𝗸:

Draft and review contracts 10x faster
➞ try it free: https://www.spellbook.legal

𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀:
0:00 - Introduction
0:21 - Meet Joshua Lenon
1:59 - What the Legal Trends Report measures
5:30 - The 24-hour rule for landing new clients
6:00 - The async vs. synchronous communication mismatch
7:18 - Why clients feel relief (and lawyers miss it)
8:32 - Payment plans and the $400/month price point
10:50 - Utilization rate: 26% to 33% in 10 years
12:21 - Value creation vs. value capture in non-billable time
13:32 - Flat fees are getting more complex and more expensive
15:51 - What the "$27K at risk from AI" number really means
17:18 - The Jevons paradox and the latent legal market
21:00 - Are lawyers doing better work now?
23:33 - Why solos are slow to adopt AI
27:02 - Connect with Joshua

Brian Powers is the CEO and co-founder of Markups AI, creators of Agent Marco - an AI that negotiates routine contracts end-to-end, backed by a team of legal engineers who translate each client's expertise into rules the agent can dependably apply.

During the episode, he takes on the biggest question in legal-AI discourse: what is "judgment” really? He splits it into two parts (the rules you set up front and the ad hoc calls you didn't see coming), makes the case for humans as a premium asset, and works through what happens to young lawyers when the low-level work that used to build judgment stops existing.

📌 𝗧𝗵𝗲 𝗥𝗲𝗱𝗹𝗶𝗻𝗲 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿:
Get the latest news, trends, and tactics in legal AI—straight to your inbox.
➞ join free: https://www.spellbook.legal/newsletter

💎 𝗦𝗽𝗲𝗹𝗹𝗯𝗼𝗼𝗸:
Draft and review contracts 10x faster
➞ try it free: https://www.spellbook.legal


𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝘁𝗼𝗱𝗮𝘆's 𝗴𝘂𝗲𝘀𝘁:
https://www.linkedin.com/in/brian-powers-41465662/
https://www.markups.ai/

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝗠𝗶𝗸𝗲 𝗪𝗵𝗲𝗹𝗮𝗻:
https://www.linkedin.com/in/mikewhelanjr/


𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀:
0:00 - Introduction
0:22 - Meet Brian Powers
0:56 - Markups AI and Agent Marco
1:26 - What "agentic" actually means
2:36 - The two types of judgment
5:09 - Where humans fit in the workflow
7:52 - AI judges and the trust gap
11:35 - Waymo, and how trust is built
13:46 - Is legal judgment an information problem?
16:05 - Why humans become a premium asset
17:07 - How young lawyers develop judgment without low-level work
20:01 - Pointing lawyers straight at the hard problems
20:42 - Marco's theory: apply the playbook, flag the rest
24:03 - Connect with Brian

Natalie Knowlton has spent 20 years working on what everyone in legal calls "access to justice", a phrase she calls "the buzzword to end all buzzwords."

Now as the Associate Director of Legal Innovation at the Center on the Legal Profession at Stanford, she's focused on civil justice reform, family court, and the middle-market clients most lawyers forget are locked out too.

During this conversation, Mike and Natalie discuss what AI can and can't do for access, why self-help isn't the fix everyone wants it to be, why the adversarial system was already broken before ChatGPT showed up, and what practitioners themselves would have to change for anything to actually improve.

📌 𝗧𝗵𝗲 𝗥𝗲𝗱𝗹𝗶𝗻𝗲 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿:
Get the latest news, trends, and tactics in legal AI—straight to your inbox.
➞ join free: https://www.spellbook.legal/newsletter

💎 𝗦𝗽𝗲𝗹𝗹𝗯𝗼𝗼𝗸:
Draft and review contracts 10x faster
➞ try it free: https://www.spellbook.legal

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝘁𝗼𝗱𝗮𝘆's 𝗴𝘂𝗲𝘀𝘁:
https://www.linkedin.com/in/natalieanneknowlton/
Bluesky/X: natalalleycat

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝗠𝗶𝗸𝗲 𝗪𝗵𝗲𝗹𝗮𝗻:
https://www.linkedin.com/in/mikewhelanjr/

𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀:
0:00 - Introduction
0:37 - Meet Natalie Knowlton
1:07 - What "access to justice" actually means
6:07 - Self-help and its limits
7:23 - Why pro se filings are flooding the courts
9:37 - The three functions of lawyers and where AI fit
10:46 - AI in the adversarial system
13:13 - The emotional good of being seen and heard
14:16 - The Nippon Life / OpenAI case
15:37 - Law as a credence good problem
16:41 - Small legal AI vs. big general AI on accountability
18:30 - Regulation and the trust vs. capability question
20:03 - What practitioners need to do next
22:11 - Bottlenecks surfacing older problems
22:33 - Connect with Natalie

Alex Su talks to the people who hire lawyers. As Chief Revenue Officer at Latitude, an alternative legal services provider that builds on-demand teams for firms and in-house departments, he sees the resourcing decisions, the AI pressure, and the trade-offs play out in real time.

During our conversation, we spoke about why in-house teams keep growing, why "judgment" isn't something you type into a chatbot, and what lawyers need to underline when their bosses ask what humans still do.

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝘀𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 𝘁𝗼 𝗔𝗜 𝗶𝗻 𝗣𝗿𝗮𝗰𝘁𝗶𝗰𝗲:

Spotify: https://open.spotify.com/show/1RqIrtq...
Apple: https://podcasts.apple.com/us/podcast...

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝘁𝗼𝗱𝗮𝘆's 𝗴𝘂𝗲𝘀𝘁:

https://www.linkedin.com/in/alexander-su/
https://www.alexofftherecord.com/

📌 𝗧𝗵𝗲 𝗥𝗲𝗱𝗹𝗶𝗻𝗲 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿:
Get the latest news, trends, and tactics in legal AI—straight to your inbox.

➞ join free: https://www.spellbook.legal/newsletter

💎 𝗦𝗽𝗲𝗹𝗹𝗯𝗼𝗼𝗸:
Draft and review contracts 10x faster

➞ try it free: https://www.spellbook.legal

Hilary Bowman built Querious to solve the most frustrating moment in legal practice - when something unexpected comes up in a client meeting and the lawyer has to say "let me research and follow up."

After more than a decade as a healthcare attorney at law firms and in-house at enterprises like IBM, she founded Querious to sit inside attorney-client conversations on Zoom, Teams, or phone and surface issues, questions, and follow-ups in real time.

In this conversation, she flips the usual AI for lawyers script and asks what these tools actually do for clients - why most clients want a relationship partner instead of a black box, why timely follow-ups beat fancy portals, and how language tools finally let lawyers meet clients where they are.

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝘁𝗼𝗱𝗮𝘆's 𝗴𝘂𝗲𝘀𝘁:

https://www.linkedin.com/in/hilarybowman/
https://www.querious.ai/

📌 𝗧𝗵𝗲 𝗥𝗲𝗱𝗹𝗶𝗻𝗲 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿:

Get the latest news, trends, and tactics in legal AI—straight to your inbox. ➞ join free: https://www.spellbook.legal/newsletter

💎 𝗦𝗽𝗲𝗹𝗹𝗯𝗼𝗼𝗸:

Draft and review contracts 10x faster ➞ try it free: https://www.spellbook.legal

𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀:

0:00 - Introduction
0:50 - Meet Hilary
2:03 - What Querious does in a client meeting
2:45 - From healthcare law to founding a startup
4:10 - What clients actually want from a lawyer
5:21 - The lawyer as relationship partner and traffic cop
6:55 - The black box problem of hourly billing
7:30 - Bridging asymmetries between lawyer and client
8:03 - Why law can't have a Domino's Pizza tracker
10:36 - The most powerful LLM unlock for clients
12:38 - Where AI ends and creative lawyering begins
14:22 - Change costs and tool adoption
15:55 - Speed to value: narrowing the scope
17:45 - Building feedback loops with clients
20:18 - One thing to implement on Monday
22:15 - Connect with Hilary

AI tools won't solve your overwhelm if your processes are the problem. John Grant is a fourth-generation lawyer turned workflow consultant. He spent a decade in tech before practicing law, and another eleven years helping firms fix how work actually moves. In this conversation, he breaks down the "magical thinking" behind AI adoption, why the real bottleneck is almost never where lawyers think it is, and how speeding up the wrong step actually makes everything worse. He also explains the concepts of "definition of done" and "definition of ready," why legal has an immature relationship with quality assurance, and what courts are already seeing as AI-assisted filings flood systems that aren't built to handle them.

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝘁𝗼𝗱𝗮𝘆'𝘀 𝗴𝘂𝗲𝘀𝘁:

https://www.linkedin.com/in/johnegrant/
https://agileattorney.com https://greenlinelegal.com

📌 𝗧𝗵𝗲 𝗥𝗲𝗱𝗹𝗶𝗻𝗲 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿:

Get the latest news, trends, and tactics in legal AI—straight to your inbox.

➞ join free: https://www.spellbook.legal/newsletter

💎 𝗦𝗽𝗲𝗹𝗹𝗯𝗼𝗼𝗸:

Draft and review contracts 10x faster

➞ try it free: https://www.spellbook.legal

𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀:

0:00 - Introduction
0:42 - Meet John Grant
2:13 - What is Greenline?
3:18 - AI hasn't reduced work
4:30 - The optimism bias and the planning fallacy
6:00 - AI and the last mile problem
7:09 - Output vs. productivity: the developer study
8:06 - Effort is not value
9:21 - Reimagining what value actually means
10:00 - There's only one bottleneck
11:16 - Why improving the wrong step makes things worse
12:10 - AI-assisted pro se filings are flooding courts
13:45 - More but worse vs. less but better
14:00 - Where bottlenecks actually form in legal
15:27 - Legal's immature relationship with quality assurance
16:15 - Quality control is chosen friction
17:21 - Definition of done and definition of ready
18:06 - Failure demand and rework
19:29 - Lawyering is a caregiving profession
22:39 - The board game analogy: multiple ways to lose
23:32 - The Agile Attorney 101 podcast series
25:04 - Takeaway: AI won't solve overwhelm without design
25:38 - Connect with John

In this episode, Mike walks through the New York City Bar's recent opinion on using AI to record, transcribe, and summarize client conversations, and boils it down to four duties you probably already follow.

Then mediator Wendy Meadows shares her full workflow, including Fathom for capture, Clio for document production, Spellbook for quality control, and how it turned a two-hour MOU into a 20-minute job. Michael Eisenberg breaks down consent rules, one-party vs. two-party states, and why none of this is actually new.

📌 𝗧𝗵𝗲 𝗥𝗲𝗱𝗹𝗶𝗻𝗲 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿:

Get the latest news, trends, and tactics in legal AI—straight to your inbox.
➞ join free: https://www.spellbook.legal/newsletter

💎 𝗦𝗽𝗲𝗹𝗹𝗯𝗼𝗼𝗸:

Draft and review contracts 10x faster
➞ try it free: https://www.spellbook.legal

https://www.linkedin.com/company/spellbookai/posts/?feedView=all
https://www.linkedin.com/in/mikewhelanjr/
https://www.linkedin.com/in/scottas/

𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀:
0:00 - Introduction
1:46 - The New York City Bar opinion (Formal Opinion 20256)
2:45 - Risk 1: Consent and one-party vs. two-party states
4:48 - Risk 2: Confidentiality and ABA Rule 1.6(c)
6:06 - Risk 3: What if the client records the call?
8:46 - Why these rules aren't new
9:05 - Wendy's practice: mediator on Zoom all day
10:00 - Turn off the defaults
10:51 - Using it like an associate
12:27 - Drafting direct examination from a transcript
13:08 - Spellbook as the quality control layer
14:48 - Predicting the future: insights in context
16:01 - Why the good outweighs the bad
16:37 - Connect with Wendy and Michael

Contract review is tedious, but missing something can be expensive.

In this episode, Chris Brown of Pixel Law shares his screen as he reviews a 56-page Asset Purchase Agreement using Spellbook.

See how effective AI tools can help you:

- Surface unknown risks, - Draft clauses like indemnification and delivery terms, and - Update you on substantive legal issues within a contract.

Chris also shares how he uses Spellbook for a "second opinion," and why he believes AI will change the practice of law.

𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗳𝗶𝗻𝗱 𝘁𝗼𝗱𝗮𝘆'𝘀 𝗴𝘂𝗲𝘀𝘁:

https://www.linkedin.com/in/thepixellawyer/

📌 𝗧𝗵𝗲 𝗥𝗲𝗱𝗹𝗶𝗻𝗲 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿:

Get the latest news, trends, and tactics in legal AI—straight to your inbox. ➞ join free: https://www.spellbook.legal/newsletter

💎 𝗦𝗽𝗲𝗹𝗹𝗯𝗼𝗼𝗸:

Draft and review contracts 10x faster ➞ try it free: https://www.spellbook.legal

𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀:

0:00 - Introduction

0:37 - What contract review looked like before AI

1:52 - The document: a 56-page asset purchase agreement

2:11 - High-level risk overview

4:18 - Diving into indemnification

8:25 - Non-compete analysis and Florida law

10:46 - Breaking down the purchase price

15:08 - Reps and warranties review (35 issues in 9 pages)

17:52 - Running the same prompt twice for a second opinion

18:24 - Closing deliveries and what's missing

22:08 - Catching missing terms across the document

24:44 - Compare to Market feature

28:35 - What stood out most from the process

29:51 - When to trust AI vs. your own judgment

31:00 - How to get started with AI tools

32:09 - Wrap up