AI Tools for Practicing Lawyers

In this Field Note, Ron Drescher breaks down one of the most important—and misunderstood—risks in legal AI: hallucinations.

The episode begins with the recent Sullivan & Cromwell filing admitting AI-generated errors, with a close look at the now-famous Schedule A. While most commentary has focused on fake citations and misquotes, Ron highlights the more subtle—and more dangerous—types of hallucinations that appeared in that filing.

From structurally corrupted citations to mutated judicial language, this episode explores how AI doesn’t just make obvious mistakes—it makes mistakes that look like law.

Ron then expands the discussion to a broader framework, identifying both the most well-known hallucination risks and the lesser-known categories that are more likely to survive verification and make their way into filed briefs.

⚖️ What You’ll Learn

• Why the Sullivan & Cromwell Schedule A is more important than the confession letter
• Two underappreciated hallucinations:
  • Citation drift (hybrid citation corruption)
  • Mutated quotations
• The 3 most common AI hallucinations:
  • Fabricated cases
  • Real cases with incorrect holdings
  • Invented quotations
• Three lesser-known (and more dangerous) hallucinations:
  • Subtle semantic drift
  • Fake multi-case consensus
  • Logical hallucination (broken arguments that look complete)
• Why “just verify the citation” is no longer enough
• A practical verification framework for AI-assisted legal writing

🧠 Key Takeaway

AI hallucinations are no longer edge cases—they are part of the operating environment of modern legal writing.

The real risk isn’t obvious errors.

It’s the errors that:

• look correct
• pass a quick check
• and still make it into your brief

📥 Downloadable Resource

This episode includes a companion Field Note:

👉 “21 Ways AI Can Hallucinate in Your Legal Brief”

Use it as a working reference during your hallucination verification process—not as a one-time read.

🔧 The New Verification Standard

Before including any authority in a brief, confirm:

• Does the case support the proposition?
• Is the quote accurate and in context?
• Does the procedural posture match your argument?
• Has the legal standard shifted subtly?

🔜 Coming Next

Field Note:
12 Ways BigLaw Associates Are Quietly Optimizing AI in Legal Drafting

A practical look at how lawyers in high-stakes environments are adapting their workflows to use AI effectively—without getting buried in verification.

🎙️ About the Show

AI Tools for Practicing Lawyers delivers practical, no-nonsense guidance to help attorneys put AI to work in their practice right now.

Hosted by Ron Drescher, a retired bankruptcy attorney with over 40 years of experience, the show focuses on real workflows—not hype.

There comes a moment for every lawyer using AI when experimentation turns into real-world application. That’s where the real opportunities begin.

Ron and Heather talk to Colorado bankruptcy attorney Matt McCune, a 25-year practitioner who isn’t just talking about AI—he’s rebuilding his law practice around it.

Matt shares how AI is transforming the entire structure of a law firm and why the lawyers who embrace it thoughtfully will define the next generation of legal service.

Along the way, the conversation explores the tension between scale and responsibility, the importance of human oversight, and how AI can elevate—not replace—legal judgment.

Key Takeaways

1. AI as a Force Multiplier (Not Just a Time Saver)

AI isn’t just about speed—it’s about removing friction from everything surrounding the practice of law.

• Automating intake, communication, and SOPs
• Reducing repetitive client interactions
• Freeing lawyers to focus on judgment and strategy

2. Human-in-the-Loop Is Non-Negotiable

AI works—but only with oversight.

• Systems can glitch
• Workflows need validation and fallback checks
• Lawyers remain responsible for outcomes

3. AI Improves Client Experience (Where Lawyers Struggle Most)

The biggest complaint in legal services? Communication.

AI enables:

• Instant responses
• Automated updates (e.g., objections, case status)
• Pre-recorded or AI-generated client prep (e.g., 341 meetings)

4. Real-World Workflow Example: Client Prep Automation

Matt uses tools like HeyGen to:

• Deliver automated video explanations to clients
• Prepare them for key moments (like 341 hearings)
• Replace repetitive phone calls

Result: better-prepared clients and less time spent repeating the same explanations.

The FSJ Framework: How Lawyers Actually Adopt AI

This episode naturally walks through the full Flintstones → Simpsons → Jetsons progression:

🪨 Flintstones Lawyer

• Start simple: download an AI app
• Use voice mode while driving
• Treat it like a conversation with a colleague

🍩 Simpsons Lawyer

• Move from dabbling to building small assets
• Example: create a simple landing page using AI tools
• Begin experimenting with workflows

🚀 Jetsons Lawyer

• Build integrated systems
• Use closed AI environments for client data
• Automate document analysis and workflows

Practice Signal: Client Communication Breakdown

A real-world scenario highlights a common issue:

• Anxious clients
• Delayed responses
• Escalation to third parties

AI Solutions:

• Instant acknowledgment emails
• Automated status updates
• Tone-optimized responses
• Ethical analysis before responding

This is where AI shines—not in legal brilliance, but in consistent, empathetic communication at scale.

What’s Next

Matt McCune will be presenting at an upcoming National Association of Consumer Bankruptcy Attorneys webinar “Stop Dabbling, Start Deploying” a two-part series focused on real-world implementation

Resources Mentioned

• Matt McCune Substack BankruptcyAttorney.ai
• AI video tools -HeyGen
• Crossing the Chasm

Most legal AI conversations focus on drafting, research, discovery, and analysis. But what if some of the most valuable AI for law firms has nothing to do with writing?

In this Workflow Options minisode, Ron Drescher explores Foundation AI, a company focused on one of the oldest and most frustrating business problems in history: incoming document chaos.

Ron revisits his Four Buckets of a Law Firm framework:

1. Sales / Marketing / Revenue
2. Production / Fulfillment
3. Operations / Administration / HR
4. Personal & Professional Development

He explains why so much legal AI attention is aimed at Bucket #2 (drafting and legal work), while Bucket #3—operations—may offer some of the safest and fastest ROI opportunities.

Foundation AI appears to operate in Ron’s green-light AI zone by helping firms:

• ingest incoming documents
• identify what they are
• match them to the correct matter
• rename files properly
• place them in correct folders
• trigger tasks and alerts
• reduce delays and manual handling

Ron also discusses Casepeer, the PI-focused case management platform featured in the webinar, where Foundation AI was presented as the workflow engine feeding the operational hub.

A key caveat: Ron notes that Foundation’s current workflow appears tied to Outlook, with Gmail integration expected in the future—a major development if true, given how many firms operate inside Google ecosystems.

This episode also previews future discussions on Microsoft vs Google ecosystems for law firms, and why that choice may become increasingly strategic.

Sometimes the most valuable AI in law doesn’t write a brief.

It just gets the right PDF into the right file at the right time. 

What’s the cheapest way for a lawyer to start using AI without creating expensive ethical, security, or sanctions problems?

In Episode 010, Ron Drescher and Heather Gardner tackle one of the most common questions lawyers are asking right now: I’m ready to try AI, but I don’t want to pay an arm and a leg—and I don’t want to get into trouble.

The conversation starts with the premium end of the market, including Harvey and other enterprise legal AI products built for large-firm workflows, governance, and document intelligence. But most solos and small firms need practical, affordable solutions—not BigLaw pricing.

Ron and Heather explain the “AI tech stack” concept: choosing tools based on what work you actually need done rather than chasing hype. They revisit the Jeffers three-legged stool framework for legal AI governance: vendor security, proper configuration, and responsible lawyer oversight.

They then break down the current budget-friendly options, including Google, OpenAI, Anthropic, and Microsoft—with a strong case that Gemini for Workspace may be the best low-cost compliant starting point for many firms.

The episode also features a Practice Signal from Reddit’s BigLaw world: a burned-out associate billing 2400 hours and wondering whether there’s a better path. Can AI help lawyers build more autonomous practices outside traditional BigLaw structures?

Finally, the Flintstones–Simpsons–Jetsons segment recommends books for every stage of AI fluency:

Flintstones:
- A Lawyer's Guide to AI by Matthew T. Henshon (ABA, 2026)
  https://www.amazon.com/Lawyers-Guide-AI-Essential-Concepts/dp/163905684X
- AI for Lawyers by Noah Waisberg & Alexander Hudek
  https://www.amazon.com/Lawyers-Artificial-Intelligence-Transforming-Profession/dp/1119723841

Simpsons:
- Co-Intelligence by Ethan Mollick
  https://www.amazon.com/Co-Intelligence-Living-Working-Ethan-Mollick/dp/059371671X
- Tomorrow's Lawyers (3rd Ed.) by Richard Susskind
  https://www.amazon.com/Tomorrows-Lawyers-Introduction-your-Future/dp/0192864726

Jetsons:
- The Coming Wave by Mustafa Suleyman
  https://www.amazon.com/Coming-Wave-Technology-Twenty-first-Centurys/dp/0593593952
- The Future of the Professions (Updated) by Richard & Daniel Susskind
  https://www.amazon.com/Future-Professions-Technology-Transform-Experts/dp/0198841892

Key Topics Covered

• Why “cheap AI” can become very expensive if it creates risk
• What premium tools like Harvey offer—and why they cost more
• The Jeffers framework for safe AI adoption in law firms
• Why lawyers should choose AI based on workflow pain points
• Gemini for Workspace as a budget-conscious legal AI option
• ChatGPT Enterprise vs Claude vs Copilot comparisons
• BigLaw burnout and using AI to build independence
• Building an effective lawyer AI stack over time

Mentioned in This Episode

• Harvey
• Google
• OpenAI
• Anthropic
• Microsoft

About the Show

AI Tools for Practicing Lawyers delivers practical strategies for the modern lawyer—helping attorneys use AI safely, effectively, and profitably in real-world practice.

Even elite firms can get burned by AI hallucinations. In this field note, Ron Drescher breaks down the recent Sullivan & Cromwell filing controversy, where an emergency brief reportedly contained multiple citation and quotation errors that opposing counsel exposed first. The lesson is not that one firm slipped—it’s that polished AI output can create false confidence in any lawyer, especially when they're under severe deadline pressure.

Ron connects the story to his earlier Confession of an AI Hallucinator episode (where he confessed to sending out a memo containing hallucinated cases) and explains why time-stressed emergency filings are fertile ground for hallucination mistakes. He then pivots to a practical alternative: using AI as a research guide, not a research substitute.

Instead of relying on AI to generate authorities directly, Ron proposes using AI to create multiple Boolean search strategies, help navigate Westlaw/Lexis/Bloomberg features, and improve the research process while keeping lawyers anchored to real databases, real cases, and real citations.

The episode also introduces Ron’s “airport metaphor” for after-the-fact AI verification: if the promised shortcut requires hours of extra checking after the brief is drafted, maybe the traditional route would have been faster all along.

Key Topics Covered

• Why AI hallucinations are not just a solo/small-firm problem
• What happened in the Sullivan & Cromwell filing controversy
• Why emergency motions and deadline pressure increase hallucination risk
• The danger of polished but false AI output
• Why Ron is skeptical of AI-as-research-substitute workflows
• Using AI to generate smarter Boolean searches
• Using AI to help master legal research tools you already pay for
• Building briefs from verified authority rather than unverifiable AI citations
• The airport metaphor for inefficient AI verification workflows

Featured Insight

“This is what using the after the fact AI verification technique is like; if you have to go through this whole verification process after writing your brief, maybe you would have been better off using the old fashioned research tools instead.”

Resources & Deliverables

• Sullivan & Cromwell apology letter with Schedule A of disclosed citation errors: https://lawyeraitoolkit.com/deliverables

About the Show
AI Tools for Practicing Lawyers delivers practical, no-nonsense guidance to help attorneys use AI safely, effectively, and profitably in the modern practice of law.

In this episode, we’re joined by Carolyn Elefant, a longtime advocate for solo and small firm lawyers and the founder of MyShingle.com.

We start with Carolyn’s upcoming AI for Lawyer Moms workshop—why she created it, who it’s for, and how AI is uniquely positioned to help lawyers working in the “crevices” of their day.

From there, we shift into a timely discussion of Management Service Organizations (MSOs)—what they are, why they’re gaining traction, and the risks they may pose for solo and small firm independence.

We also explore how client expectations are changing in the age of AI, why lawyers can no longer ignore these tools, and how even small workflow upgrades can create meaningful time savings.

Finally, we wrap with a Practice Signal on managing client decision-making risk and an FSJ (Flintstones–Simpsons–Jetsons) segment featuring practical resources to help lawyers level up their AI fluency.

00:00 – Intro & Guest Welcome
Meet Carolyn Elefant and her work in the solo/small firm and AI space

01:00 – AI for Lawyer Moms: Why Now?
The AI adoption gap and why women lawyers may be at higher risk

03:00 – AI in the “Crevices” of Your Day
Using AI in small pockets of time for real productivity gains

04:30 – Workshop Focus: Claude, Perplexity & Workflow Integration
Moving beyond prompts to real legal workflows

06:00 – AI Ethics & Security: Practical Guidelines
SOC 2, data protection, and where to draw the line

09:00 – What Attendees Will Actually Do After the Workshop
Immediate, practical next steps

10:00 – What is an MSO?
Understanding Management Service Organizations

11:00 – “It Gets Up in Your Business”
Where MSOs move from helpful to intrusive

12:00 – AI Access, Cost & Co-Op Possibilities
Can solos share access to enterprise tools?

14:30 – AI for Small Firms: Progress & Challenges
Clio, vLex, and the difficulty of reaching the solo market

17:00 – Carolyn’s Background & MyShingle
22 years of advocating for solo lawyers and tech adoption

19:00 – Is AI Different from Past Tech Shifts?
Why this wave may be faster and more client-driven

20:30 – The Client Expectation Shift
“Why did this take a month when AI can do it in 10 minutes?”

22:00 – AI-Savvy Clients & Workflow Friction
From payment expectations to AI-generated documents

25:00 – Practice Signal: Managing Client Risk Decisions
Using AI for research, communication, and risk framing

29:00 – AI for Client Communication
“Show your work” and reduce friction

31:00 – FSJ Segment: Resources to Level Up
From Flintstones to Jetsons—where to start and how to grow

37:00 – Final Thoughts & Workshop Recording Info

•  Carolyn Elefant – MyShingle 
•  AI for Lawyer Moms Workshop 
•  Blaine Oelkers – 30 Day AI Challenge 
•  Zach Shapiro 
•  Bob Ambrogi - Lawnext.com
•  Bill Henderson – Legal Evolution 
•  Mark Cohen – Legal Mosaic 
•  Nate B. Jones - TikTok
•  Sabrina Romanoff (YouTube – Claude tutorials) 

In this kickoff “Workflow Options” episode, Ron Drescher takes a closer look at Ivory Mind, an AI document assistant designed to help professionals quickly understand and organize their materials.

At first glance, Ivory Mind didn’t seem to fit the frameworks Ron has been developing on the show—like the Three-Legged Stool and Folder Mania tests. But after a deeper look, a different question emerged:

This episode walks through that shift in thinking and explores where Ivory Mind may (and may not) fit in a modern legal workflow.

• Why some AI tools fail advanced frameworks—but still provide real value
• How document-driven AI tools can simplify everyday legal work
• A practical workflow for turning client conversations into structured work product
• Why many lawyers want AI benefits—without diving into the “AI rabbit hole”

• Ivory Mind is not a full-scale legal AI system — it’s a focused, document-driven tool
• It works best for:
  • Small to mid-sized matters
  • Document-heavy but manageable files
  • Lawyers who want simplicity over flexibility
• It is not designed for large-scale litigation workflows or deep system integration
• The real value is in:
  • Transcription
  • Summarization
  • Quick understanding of documents and conversations
  • Clean UI 

One of the most useful applications discussed:

📌 Client Meeting Capture Workflow

1. Record a client meeting (with appropriate consent)
2. Upload the audio file (WAV/MP3) into Ivory Mind
3. Generate:
   • Transcript
   • Structured summary
   • Key takeaways
4. Use the AI to draft:
   • Client follow-up email (adjusted to appropriate level)
   • Next steps / task list

👉 Result: A clean, searchable record of what actually happened—without relying on memory or handwritten notes.

This episode introduces an important concept:

Ivory Mind may be a strong fit for:

• Flintstones-level lawyers looking to ease into AI
• Lawyers who want structure without complexity
• Anyone who prefers a quieter, more focused AI experience

• No direct integration with cloud storage (Google Drive, Dropbox, OneDrive)
• No ZIP file ingestion
• No JPG/PNG (image) support (as of testing)
• Manual document upload required

As described by the Ivory Mind team:

• 🌐 Learn more: https://ivorymind.com
  
  

Sometimes the right question isn’t whether a tool fits your system…

…it’s whether it helps you get your work done more easily.

Know a lawyer curious about AI but avoiding the chaos? Share this episode—it might be the entry point they need.

AI can dramatically improve discovery work — summarizing productions, organizing documents, spotting patterns, building timelines, and making massive email chains usable. But most lawyers still lack a clear, defensible framework for using AI on discovery materials.

In this episode, Ron and Heather break down a recent Kansas federal court opinion and explain why it may become one of the most important practical standards for lawyers using AI in discovery. The key takeaway: you can use AI in discovery, but only if you do it in a way that is secure, controlled, and defensible.

The discussion introduces Ron’s practical framework for compliance: the Three-Legged Stool:

• The Vendor Leg — your AI tool must be capable of operating as a “closed” system
• The IT Leg — the tool must be configured properly
• The Lawyer Leg — your firm must use the tool with supervision, scope control, logging, and accountability 

Ron and Heather also discuss why consumer-grade AI tools are a red light for discovery materials, why simply buying an “enterprise” plan is not enough by itself, and how even solo and small firm lawyers can create a practical, affordable workflow that satisfies the emerging standard. 

The episode also includes:

• a Practice Signal on whether solos can gradually “merge” with larger firms by sharing systems and infrastructure, and
• a Flintstones / Simpsons / Jetsons breakdown of how lawyers at different AI comfort levels should think about discovery workflows. 

What You’ll Learn

• Why AI in discovery is useful but risky without a standard
• What the Kansas federal court opinion actually says
• The difference between open AI tools and closed AI tools
• Why “enterprise” is a useful buying shortcut — but not the full answer
• Why configuration, deletion, retention, and access control matter
• How solos and small firms can create a defensible workflow without overcomplicating it
• Why the “file cabinet test” still matters even after you buy a compliant tool

Free Deliverables Mentioned in This Episode

Ron created a set of practical, free downloadables to help lawyers operationalize the workflow discussed in this episode.

Download them here:

https://lawyeraitoolkit.com/deliverables

Included resources:

• Plain-language AI discovery protocols
• Proposed protective-order language
• Notice of intent to use AI on discovery materials
• Matter-level AI usage log template

These are designed to help lawyers move from vague concern to actual defensible implementation.

Key Takeaways

1. AI in discovery is not the problem — improvisation is

The issue is not whether AI can help in discovery. It can. The problem is that many lawyers are using it without a repeatable standard. 

2. “Enterprise” is a shortcut, not a safe harbor

If your tool doesn’t offer an enterprise-grade environment, that’s a major warning sign. But simply buying the higher tier does not mean your firm is compliant. The tool still has to be configured and supervised correctly. 

3. The right tool must be both safe and functional

A compliant tool that cannot actually handle your document workflows is still the wrong tool. If it can’t help you work effectively with a large discovery production, it may fail the practical test even if it passes the safety test. 

4. This framework can extend beyond discovery

Once a law firm builds a defensible AI workflow for discovery, that same thinking ca

In this companion Field Note to Episode 008, Ron walks through the practical steps lawyers can take to implement the emerging AI discovery standards discussed in Jeffries v. Harsco. He breaks the process into a simple three-legged stool: choosing the right vendor, properly configuring the tool, and handling the lawyer-side workflow and documentation needed to make AI use in discovery more defensible.

In Episode 008, I talked about the emerging legal standard for using AI in discovery and why lawyers need more certainty in this area. In this Field Note, I focus on the practical question:

How do you actually implement it?

The answer, in my view, is a simple framework:

To use AI in discovery in a way that is more defensible, you need all three of these in place:

1) The Vendor Leg

Start with the right environment. In practice, that usually means an enterprise-level AI tool — one that operates in a closed system, does not train on client data, provides a secure environment, allows deletion, and gives you a way to define who has access. 

2) The IT / Configuration Leg

Buying the right tool is not enough. You also need to configure it correctly. In this Field Note, I explain a practical workflow for doing that:

• upload the protocol,
• have the tool or your IT consultant walk you through the settings,
• configure the tool,
• take screenshots,
• and preserve proof of configuration. 

3) The Lawyer Leg

This is where the legal workflow becomes defensible. I walk through the downloadable documents that can help lawyers operationalize the standard:

• a proposed form of order
• a notice of intent to use AI in discovery
• the protocol itself
• and a log template for documenting AI use in the workflow. 

• Why the vendor leg is probably the easiest part of the stool to satisfy
• Why lawyers should not get paralyzed about choosing the “perfect” AI tool
• How to use the settings / configuration process to create a more defensible AI environment
• Why preserving screenshots, DPAs, and configuration emails matters
• How the lawyer-side deliverables can help you build a cleaner, more defensible workflow
• Why this issue is likely to become a normal part of law practice sooner than many lawyers think 

The downloadable materials discussed in this Field Note are available here:

https://lawyeraitoolkit.com/deliverables

That page currently includes:

• AI Discovery Protocol
• Notice of Intent to Use AI in Discovery
• Proposed Rule 26(f) / Order Language
• AI Discovery Log Template

You do not need perfection. But you do need all three legs of the stool.

If you have:

• the right tool
• the right configuration
• and the right lawyer-side documentation

…then you are in a much stronger position to explain and defend your use of AI in discovery. 

If you know a lawyer who is:

• experimenting with AI,
• using it informally,
• or avoiding it because they don’t trust it yet,

send them Episode 008, this Field Note, and the Deliverables page.

Because this is exactly the kind of issue where certainty matters.

Every AI tool claims it can read your folders. We actually tested that claim — across ChatGPT, Claude, Gemini, and Microsoft Copilot — using a real-world 48-document Chapter 13 case file. The results were all over the place.

Ron introduces the "File Cabinet Test": a threshold check that tells you whether an AI tool is actually seeing what you're giving it, before you trust it with anything that matters. Because if the AI is faking it — cherry-picking by relevance, wandering outside your folder, or missing documents entirely — that's not an AI intelligence problem. That's a visibility problem. And for lawyers, visibility is everything.

This episode is a deep dive into folder access, document security, workflow design, and what it really means to close the trust gap in your practice.

What You'll Learn

• What the File Cabinet Test is and why it's the first question you should ask of any AI tool
• How ChatGPT, Claude, Gemini, and Copilot each performed on real legal documents — and where each one surprised us
• Why "the AI comes to you" is a game-changer for law firm document security
• The hidden risk when AI output looks complete but isn't
• How a bankruptcy trustee's preference complaint workflow points to the next frontier of AI-assisted legal practice
• What Claude's new MCP integration with NetDocuments means for Jetsons-level lawyers — right now

Chapter Markers

• 0:00 — Introduction: Testing AI folder access across four platforms
• 1:05 — The test case: A complex Chapter 13 with 48 documents
• 2:00 — Introducing the File Cabinet Test (and why AI wants to fake it)
• 3:47 — Heather's ChatGPT zip file test: A home run — and a teaching moment
• 5:54 — Visibility vs. intelligence: The real question we're asking
• 7:30 — The million-document file: A tease for the roadmap
• 8:58 — ChatGPT connected to Google Drive: Same tool, different result
• 10:42 — Claude: Security, redaction, and a perfect file cabinet test
• 13:34 — Claude's limitations: Budget gaps and the timeout problem
• 15:00 — Analysis vs. ecosystem tools: How to frame the choice
• 15:59 — Gemini inside Google Drive: Close, but not quite
• 18:27 — The trust gap: Why small firm lawyers can't do what we just did
• 19:41 — Copilot: Why Ron wanted it to work, and what happened instead
• 22:19 — The Copilot saga escalates: Zip files, crashes, and a 20-file ceiling
• 25:57 — Where Copilot actually belongs in your workflow
• 27:30 — Workflow-first, tools-second: How to find your bottleneck
• 28:49 — AI malpractice on the horizon: The human oversight imperative
• 31:50 — Practice Signal: Preference complaints and the AI merge-print breakthrough
• 36:13 — FSJ Level-Up: Flintstones, Simpsons, and Jetsons recommendations
• 38:49 — Jetsons surprise: Claude + NetDocuments MCP integration, live today
• 40:34 — Ron's take: Why he resisted Claude — and why he changed his mind
• 41:35 — Closing thoughts: Making lawyers less afraid

Resources & Links

• lawyeraitoolkit.com
• ChatGPT Enterprise — openai.com
• Claude — claude.ai
• Google Gemini — gemini.google.com
• Microsoft Copilot — copilot.microsoft.com
• NetDocuments — netdocuments.com