Webinar takeaways: Turning legal data into AI readiness

AI is everywhere, but many legal teams still struggle with where to start. In this webinar recap, we share Pia Madissoo’s (EY Law Finland) insights on how to transform everyday legal data and processes into genuine AI readiness. 

The focus: what to prioritize, how to measure success, and how to build confidence without blowing the budget.

What you’ll learn:

• Where legal teams actually are with AI today, and why that’s okay
• The two foundations that matter most before any tooling: processes and data
• How to define “legal data” beyond spreadsheets and numbers
• A simple way to show early success and ROI
• Practical steps to prepare your team for AI in 2026

The reality check: You’re not late

Despite the noise on LinkedIn, most legal teams are not yet deeply invested in AI. In the 2025 EY Law General Counsel Study, respondents prioritized legal tech strategy and data strategy, while the active use of generative AI remains limited. That means you’re not behind if you’re still laying groundwork. You’re doing the right work.

Why this matters: You build ROI on strong foundations, not hype. Start with the basics that make AI safe, useful, and measurable.

Process first: Fix the pipe before you add pressure

If your intake, review, and handoffs are messy, technology will not magically fix them. It often amplifies what’s already broken. Start with simple process mapping.

Do this:

1. Sketch your key flows at a high level (e.g., sales contract review, vendor onboarding).
2. Write down the steps and owners. Keep it simple and visible.
3. Track the work in a system that captures activity and time spent. You need this data later to prove improvement.
4. Remove repeated friction: unclear intake, missing context, back-and-forth over versions, and approval bottlenecks.

Outcome: You gain visibility, faster cycle times, and the baseline metrics you’ll use to show AI impact later.

Data next: Teach AI with your context

AI is powerful, but without your data, it’s an “articulate ignorant” – fluent, yet missing your policies, risk appetite, and preferred positions. Enrich AI with your organization’s knowledge so outputs align with how you actually work.

What counts as legal data? More than numbers:

• Playbooks and positions: clause libraries, fallbacks, negotiation guidance
• Contracts and obligations: terms, metadata, renewal dates, DPAs
• Matter data: request types, business owners, turnaround times
• External counsel and spend: what you outsource, why, and how much
• Policies and templates: approvals, standards, security requirements

Practical move: If you use Copilot or a similar tool connected to your workspace, ask it to help inventory what’s already documented across your file storage systems. Use that list to prioritize what to structure and clean first.

Where AI can help right now

Legal teams report the most traction in:

• Drafting and review assistance
• Knowledge management and surfacing prior work
• Contract analysis for risk and obligations
• Learning and upskilling through guided modes

Start with a straightforward process and sufficient data signals. Select one workflow and optimize it from end to end.

ROI that leadership will trust

ROI is not abstract. It shows up in:

• Time saved on routine tasks
• Reduced external spend (especially overflow work)
• Fewer escalations and faster cycle times
• Better compliance with playbooks and policies

How to prove it:

1. Capture your baseline before changes (cycle time, external spend, queue length).
2. Run a small pilot with a clear scope.
3. Measure the same metrics after two to four weeks.
4. Share a brief win report with numbers and include a customer or business partner quote.

Early success builds momentum, budget, and buy-in.

A simple readiness pyramid

Think of AI readiness as a pyramid:

1. Processes: clear intake, documented steps, known owners
2. Data: structured playbooks, contract metadata, consistent tags
3. Tech: AI features and automation layered on top of clean processes and data

Most teams want to start at the top. Don’t. Invest the first two layers with discipline before even thinking about tech. The AI layer becomes safer and far more effective.

Two adoption paths that work

Path A – earlier-stage companies: Automate a single workflow end-to-end. Use agile delivery. Release, observe, refine. Move to the following workflow.

Path B – complex or mature organizations: Stabilize processes and data first. Map two or three high-volume flows, standardize intake, add the proper metadata, and only then introduce AI assistance. You’ll see faster adoption and cleaner ROI.

Change management: Start small, show progress, reduce fear

AI brings excitement and anxiety. Treat change management as part of the work:

• Set a small, clear goal for the first pilot.
• Involve a few champions from legal, procurement, and the business.
• Share a quick weekly update with one metric and one learning.
• Celebrate early wins. Address risks openly.
• Keep the human in the loop for sensitive steps.

Trust grows when people see stable outcomes, not slide shows.

A practical 30-day plan

 

Week 1	Week 2	Week 3	Week 4
Pick one workflow (e.g., standard NDAs or low-risk vendor agreements). Map the steps and owners. Note current cycle time and touchpoints.	Take inventory of the data you have: templates, clause variants, playbook rules, and past contracts. Add missing metadata that is relevant to this workflow (party, term, renewals, governing law).	Introduce AI assistance where it’s safe: clause suggestions, summary, or playbook checks. Keep human approvals for final decisions.	Measure impact vs. baseline. Document lessons learned and subsequent improvements. Share a one-pager of results with legal, IT, and a business sponsor.

FAQs from the session

Q: How can vendors help show early success?

A: Leverage prior implementation patterns. Guide customers on what to prep before implementation ever starts: clean inputs, essential metadata, and a tight, pilot-ready workflow.

Q: What if the team is change-resistant?

A: Acknowledge concerns. Start with a small, safe workflow. Involve people early, show progress weekly, and tie improvements to less rework, more precise guidance, and fewer late-night reviews.

🔑 Key takeaways

• You’re not late. Most teams are still building strategy and data foundations.
• Fix processes first. Clear intake and ownership make every tool more effective.
• Treat your knowledge as data. Playbooks, clause positions, and past deals teach AI your context.
• Prove ROI early. Track time saved and reduced overflow spend on one pilot workflow.
• Adopt in layers. Processes, data, then tech. That order creates predictable results and real confidence. 

Want to watch the webinar on-demand? You can do that here.