Custom AI Training for a Fortune 500 Energy Legal Team

The situation

The legal team had broad interest in AI but uneven exposure. Some attorneys had used ChatGPT for one-off tasks. Others hadn’t touched generative AI. Legal operations was fielding AI questions from the business side without a framework for evaluating them. Leadership had budget for a single workshop — and it had to land for the whole department, not just the AI-curious early adopters.

The team works on Microsoft 365 across the board, evaluating specialized legal-AI tools (Legora) and using SharePoint as their de facto document home. Day-to-day work centers on land leases, title commitments, chain-of-title deeds, and contract negotiation for renewable energy projects across Texas.

They wanted training that fit their work — not generic “what is AI” content.

What made this engagement custom

Before the session, M41 ran three 30-minute pre-work interviews with designated team members across the lawyer / non-lawyer split. Those conversations shaped the agenda, the demo content, and the hands-on exercises.

The fictional Texas leases used in the workshop were named and structured to match the kind of leases the team actually negotiates. The deeds for the chain-of-title exercise spanned 1958–2019 across Reeves County, Texas — including a scanned image of a typewritten 1971 warranty deed, because that’s what the team deals with in the wild.

The session split into parallel tracks so both groups could go deep without sitting through each other’s content:

• Attorneys worked through a multi-clause contract playbook build, then ran a counterparty redline through it and produced a turn memo as the final deliverable — the same workflow they’d use on a real negotiation.
• Legal operations worked through title commitment review, non-monetary obligations extraction, lease data extraction, and chain-of-title analysis from the 1958–2019 deed set.

The opening “Responsible Use” session was delivered by the client’s own Director of Data Use & Privacy — not by M41. That single choice did more to legitimize AI as a department-wide practice than any third-party safety pitch could have.

What the team walked away with

• A working library of prompts tied to their actual matters. Not generic templates. The attorney track left with a playbook-build → redline-response → turn-memo pipeline they can run on real counterparty redlines. The non-lawyer track left with structured extraction templates for the lease and deed work they handle every week.
• A shared vocabulary for AI safety. Three calibrated-trust techniques — run it in parallel, get a second opinion, ask it to push back on itself — replaced the binary “is this safe?” anxiety that typically blocks adoption.
• A daily check-in habit designed to keep finding new uses for AI after the workshop ends. Without one, training fades in 2–3 weeks. With one, the team finds new uses every week.
• A map of where AI fits alongside SharePoint, the CLM the team is evaluating, the DMS they’re planning toward, and the specialized legal-AI tools under review.

What an engagement includes

Every M41 legal training engagement is built the same way:

• Pre-work interviews to shape the agenda around the team’s actual practice
• Custom-built workshop documents matched to the team’s matter types — leases, deeds, contracts, playbooks
• The workshop itself, typically half-day, split-track for mixed audiences
• A password-protected materials site with the recording, transcript, slides, demo files, every workshop document, and copy-paste prompts the team can return to
• A 60-minute open team check-in 2–3 weeks after the session
• Five 30-minute individual coaching calls within 45 days

Group training shows people what’s possible. The follow-on coaching is where each person can ask the question they didn’t want to ask in front of 55 colleagues.