Saturday, 8 August 2026
Rather than asking lawyers to become expert prompt engineers, Texas law firm Jackson Walker LLP is using artificial intelligence to capture the judgement of experienced partners and turn it into structured workflows that can be reused across the firm.
Speaking at AI4 2026 in Las Vegas, Greg Lambert, Chief Innovation Officer at Jackson Walker LLP, outlined how the firm is moving beyond individual AI prompts towards playbooks that embed legal expertise into repeatable processes.
The session, titled From Prompts to Playbooks: Turning Your Best Lawyers’ Judgement into AI Workflows, focused on how law firms can scale institutional knowledge while maintaining professional oversight.
Jackson Walker is a Texas-based full-service law firm with approximately 550 lawyers and around 380 professional staff. Lambert said the firm’s role is not to replace legal judgement but to make existing expertise easier to apply across matters and practice groups.
AI becomes part of everyday legal practice
The firm currently uses several AI platforms alongside established legal research services.
These include Harvey for drafting, document review and summarisation, Microsoft Copilot integrated with Microsoft 365, legal research platforms Westlaw Precision and Lexis+ AI, as well as pilot projects using Anthropic’s Claude Enterprise to develop more advanced agentic workflows.
Lambert stressed that specialised legal research platforms remain essential despite improvements in general-purpose AI.
“General models are getting better, but they are not legal research tools,” he said, noting that authoritative legal databases remain necessary to reduce the risk of inaccurate citations and hallucinated case law.
Adoption driven by practice groups rather than mandates
Instead of requiring lawyers to adopt AI, Jackson Walker introduced the technology gradually.
The firm initially licensed around 100 Harvey users before expanding adoption incrementally as interest grew.
Today, more than 600 employees have access to Harvey, with approximately 84% using the platform each week, according to figures presented during the session.
Rather than measuring success by the number of users alone, Lambert said the firm identified influential lawyers within each practice group, worked with them to solve specific business problems using AI, and then relied on those early users to encourage wider adoption.
This approach generated demand across the organisation without requiring mandatory implementation.
Converting experience into reusable knowledge
Lambert argued that many experienced lawyers already possess structured decision-making processes built over decades of negotiating contracts, reviewing documents and advising clients.
The challenge is extracting that knowledge in a way that others can use.
Instead of asking lawyers to create AI prompts from scratch, Jackson Walker analyses historical document revisions and identifies recurring legal positions taken by experienced partners.
These patterns are then converted into structured rule sets before being reviewed and edited by the lawyers themselves.
“The playbooks already exist,” Lambert said. “They’re contained in years of document mark-ups and negotiated agreements.”
One example involved a commercial real estate finance partner whose previous loan agreement negotiations were analysed.
The firm reviewed six completed transactions and identified approximately 90 recurring negotiating positions across seven key sections of borrower-side loan agreements.
After review by the partner, those positions were refined into a 49-rule playbook capturing standard negotiating positions, acceptable fallback language and circumstances where the firm would recommend walking away from a transaction.
The completed playbook could then be used by junior lawyers reviewing similar agreements while still allowing experienced partners to exercise judgement where required.
Supporting associates as well as AI
Lambert said the resulting playbooks have also become valuable training resources.
Rather than relying entirely on traditional mentoring, associates can review the reasoning behind common negotiating positions before working on live matters.
The playbooks are written in plain language and integrated into Microsoft Word, allowing lawyers to compare new documents against established guidance and identify clauses that comply with, differ from or fall outside accepted practice.
The same methodology is now being applied across multiple practice areas, including commercial contracts and transactional work.
Document management becomes an AI resource
The firm’s document management systems, including iManage and NetDocuments, have become an important source of institutional knowledge.
AI can search previous work while respecting ethical barriers that prevent lawyers from accessing confidential matters outside their authorised client work.
Lambert described one example where lawyers needed information about a serial litigant.
Instead of relying on firm-wide emails asking whether colleagues had previously encountered the individual, AI searched historical records and identified four lawyers with relevant experience, including one who had already prepared an internal memorandum outlining legal strategies used in earlier cases.
The information was retrieved within minutes rather than requiring manual searches across the organisation.
Measuring outcomes rather than usage
Lambert cautioned that user adoption alone provides only a partial measure of AI success.
While weekly usage statistics indicate whether staff are engaging with the technology, they do not show whether AI is improving business performance.
Instead, Jackson Walker focuses on operational measures including turnaround times, matter progression and revenue generation.
The firm also monitors governance, document access controls and AI operating costs to ensure that automation delivers measurable value rather than simply increasing computing expenses.
Building systems instead of prompts
Looking ahead, Lambert expects legal AI to move away from individual prompting towards structured systems built around established workflows.
Rather than asking lawyers to become increasingly sophisticated AI users, future systems are likely to guide them through familiar legal processes, automatically requesting the documents and information needed for each matter while applying established playbooks behind the scenes.
He also argued that clients are becoming more receptive to AI-assisted legal work, particularly where efficiency enables firms to handle additional matters without compromising quality.
For Lambert, the long-term opportunity lies less in teaching every lawyer to write better prompts than in capturing decades of professional judgement and making it consistently available across the organisation.
The approach reflects a broader shift taking place across professional services, where AI is increasingly being used to standardise expertise and improve knowledge sharing rather than simply automate individual tasks.
© 2026 cij.world
