Jimmy Toy is Chief Legal Officer at Articore Group, the company behind Redbubble and TeePublic: two artist marketplaces, roughly 70 million user-generated images, and IP disputes on several continents. When a lawsuit landed in a new jurisdiction, the first job was memory. Somewhere in years of filings, his team had described the business or framed an argument in a way worth repeating, and new outside counsel needed it fast, on a litigation budget that had other plans.
On CZ and Friends, GC AI's podcast that interview leaders in legal, Jimmy Toy described the search that sits at the center of legal knowledge management for in-house teams:
"We would have to go back through just hundreds of litigation documents that we filed to try to find that one little passage where we may have described something or made an argument in a certain way that we thought was really good."
For years, that search had exactly one qualified operator. "I am the only one that really had the knowledge of what we said [in] a certain case in France and how that might apply to an argument that we're making in India," he told Cecilia. He could spend three hours on it "and maybe not come to an actual answer on it." That is the in-house failure mode a law firm never sees up close: the archive lives in one lawyer's head, and the database walks out the door at 6 p.m.
We've had more than 40 conversations like this one on CZ and Friend and the pattern holds from three-lawyer teams to public-company legal departments.
The teams that keep their knowledge do three things: they capture five specific assets, they give those assets one AI-readable home, and they start with the two contract types they touched this week. The teams doing this measure the payoff in hours reclaimed, and you can test the approach on your own documents before anyone sends an invoice.
Legal Knowledge Management In-House and at a Firm: Two Different Jobs
Legal knowledge management is the practice of capturing a legal team's positions, precedents, and reasoning so the team can reuse them on the next matter. Law firms invented the discipline as a product: professional support lawyers, precedent committees, and know-how databases exist to package expertise for resale across client engagements. An in-house team runs the same discipline for a different customer. The knowledge serves the business down the hall, and the metric is speed-to-answer: how fast sales gets the redline, how fast the board gets the risk position, how fast new outside counsel gets a decade of context.
The constraints differ too. A legal department has no professional support lawyers and no knowledge-management headcount. Per the ACC Law Department Management Benchmarking Report, corporate legal budgets split near-evenly between internal spend and outside firms, which leaves nothing labeled "knowledge infrastructure." And the knowledge itself scatters differently: across email threads, Slack approvals, the CLM, and business units that agreed to terms over coffee before legal saw the paper.
So an in-house team builds its knowledge system a specific way: as a byproduct of the work itself, captured while the review happens, by the lawyers already doing it.
Why Lean Legal Teams Lose Knowledge Faster
A three-lawyer team that loses one lawyer loses a third of its institutional memory on two weeks' notice. The math compounds, because the departing lawyer owned entire contract categories, entire regulator relationships, entire histories of "we tried that in 2023." Law firms hedge this risk with leverage and redundancy; a corporate legal department of three runs without either hedge.
Sharon Johnson, SVP and Chief Legal Officer at MODE Global, runs a lean team behind a multibillion-dollar logistics business, and her defense is to write the knowledge down before anyone needs it:
"When you're lean, there's just no place to hide. I would say we'd have to be disciplined where we spend our time. We do things like creating playbooks, escalation paths, or we might build contract positions out before we need them so that you have that."
"Build contract positions out before we need them" is the practice in nine words. The position paper written in a calm week is the one a teammate runs during a fire drill, and the one that survives its author's departure.
Cecilia has seen the expensive version of the alternative. On the podcast, she pointed to her Amazon years: David Zapolsky joined Amazon in 1999, served as its general counsel for over a decade, and carried the institutional understanding of Amazon's risk approach through antitrust scrutiny, labor fights, and a market cap that climbed toward a trillion dollars. Companies pay for memory like that at the top of the executive pay scale. A team of three builds it a different way: they encode it.
The Five Assets of In-House Legal Knowledge Management
Across the 2,100+ legal teams using GC AI as of September 2026, including the legal departments at Liquid Death, Arc'teryx, Columbia Sportswear, and Eventbrite, the teams that keep their knowledge capture the same five assets. We call them the In-House Knowledge Stack:
Negotiation playbooks: the team's position, fallback, and hard stop for each clause in each contract type.
A clause library: the language the team accepts and the redline it sends back.
Templates and precedents: the paper the team starts from, with the reasoning attached.
Matter memory: what happened last time, findable by someone who wasn't there.
A self-serve layer: the answers the business can get without opening a legal ticket.
Negotiation Playbooks
A playbook turns one lawyer's judgment into the team's standard: the primary position for each clause, the approved fallback, the walk-away, and the vetted stock comment that goes out with the redline.
In our Building Playbooks class we describe the guide section as the memo you'd hand a first-year associate, minus the associate: "mark up this agreement, here is my playbook." Write it once, in a calm week, and the Tuesday NDA runs as a lookup: the standard is already decided. The full method is in our guide to building a contract playbook.
A Clause Library
A clause library records where the team lands on the provisions that carry the risk: the indemnification cap you accept in vendor paper, the limitation of liability carve-outs you insist on, the governing law you concede and the one you never do. It is the difference between a team position and three lawyers negotiating three different companies' worth of positions under one logo.
Templates and Precedents
Jeremy Siegel, General Counsel at Final Bell and a CZ and Friends guest, put the drafting truth plainly:
"Everyone copies and pastes. Everyone finds good precedent, modifies it to the way that they want to use it, and then claims it their own."
Precedent reuse is the practice of law working as designed. Retrieval is where it breaks. The strong precedent helps whoever can find it, and in a shared drive with a folder named "Final FINAL v3," that is a coin flip. A template set with the reasoning attached ("we added this carve-out after the 2024 vendor dispute") turns reuse from archaeology into a lookup.
Matter Memory
Matter memory answers "what did we say, where, and why" across live and closed matters. A legal matter management software system tracks where each matter stands; the reasoning behind it lives somewhere else. Jimmy Toy's team built that reasoning layer at Articore:
"One of the things that we've been doing is building a knowledge base around all of our court filings that we've done. So it can be affidavits, declarations, even deposition transcripts... you can really put it all in and create this excellent knowledge base that's comprehensive. And then you can just interact with it like it's a person."
The consistency check that used to fall on him alone, three hours at a time, now runs as a question. He still verifies the answer against the underlying documents, and the hunt itself is gone: "you never would have been able to locate those and find the exact passages without that sort of help."
A Self-Serve Layer
The highest-leverage knowledge asset is the one the business uses without you. Jenna Hunt, Head of Legal Operations at Tipalti, started with the contract that eats the most junior-lawyer hours per dollar of risk:
"Lawyers negotiating NDAs is not practicing at the top of their license. So that was sort of a low hanging fruit... it was just establishing those playbooks, empowering the business to make decisions, empowering the business to negotiate those things, creating automated workflows for pre-signed NDAs so that we don't need to even have legal involved."
Encode the NDA standard once, and sales gets its two-page mutuals back the same day. The lawyers spend their afternoons on the work that needs a law degree.
Where Legal Knowledge Lives Today (and Why Search Fails)
The knowledge in a legal department already exists. It sits in systems built to store documents, and the job the team needs done is answering questions. Those are different jobs.
Where It Lives | What It Holds | What It Was Built For |
Document management (iManage, NetDocuments) | Executed agreements, filings, versions | Storage, permissions, governance |
CLM | The contract pipeline | Intake, approvals, renewals |
Email and Slack | The reasoning and the approvals | Conversation |
The wiki or intranet | Policies and process docs | Reference |
The senior lawyer | The judgment calls | Answering the same question twice a week |
Keep the storage layer; it does its job well, and the corporate legal software stack works in layers by design. A DMS versions and governs documents; a CLM moves the contract pipeline. What the stack gains from AI is a layer that reads across all of it and returns an answer with a citation, the layer that knows the 2024 indemnity fallback exists, finds it, and quotes it back with the page it came from.
AI Turns Your Precedents Into Working Memory
AI changes the economics of legal knowledge management because retrieval becomes a query any teammate can run. Gartner's evaluation of generative AI use cases for legal departments ranked document summarization and legal intake among the highest-value applications, and both are knowledge work: reading what the team already produced and putting it to use. Grounding in your own paper is also the difference between in-house counsel AI software and a general-purpose chatbot answering from the public internet; GC AI vs ChatGPT walks through what changes.
Here is what that looks like on a Tuesday. A vendor renewal lands, and the account manager asks whether the price cap from last year carried over. You keep the vendor's history in one Files collection your team can query, up to 1,500 pages at a time, so the three prior agreements are already in front of you. You ask, and the answer comes back with a character-level Exact Quote citation you check against the page it came from.
You run the renewal against the positions your team encoded into Playbooks, built from your own standards with Easy Playbooks, and you carry the March negotiation's context forward without re-reading the thread, because Projects, GC AI's interchat memory that you control, kept it. By the time the account manager follows up, your answer carries a citation and your redline matches the position the team took last year.
See Files turn your precedents into a team reference library:
That workflow is what makes knowledge survive turnover and PTO. Alexis Palmer, Senior Managing Counsel at Snyk, runs her team's standard reviews on shared saved prompts:
"Having saved prompts means anyone on my team can run the same review I would. If I'm on PTO, I know they'll get a similar result and apply their own judgment from there."
And it scales past any single lawyer's memory. At Wayfair, three attorneys plus one part-time colleague review the contracts for the whole company. Lauren Anderson, Senior Counsel at Wayfair, named the target on CZ and Friends:
"How can we be smarter about utilizing playbooks or creating tasks or an idea of like a Wayfair brain where we have all of this knowledge throughout the legal team that we can kind of make these more mundane tasks a little more automated, a little faster."
A "Wayfair brain" is the right ambition, and it is buildable this quarter with the documents the team already has.
A 90-Day Legal Knowledge Management Build Plan
A lean team builds a knowledge system as a byproduct of live work: no taxonomy project, no KM hire, no migration. Ninety days, three phases.
Days 1 to 30: encode the two contract types you touch at the highest volume. Pull the count from your inbox and pick the top two. Write the playbook for each: position, fallback, hard stop, stock comment. One page per contract type is enough to start.
Days 31 to 60: give the assets one AI-readable home. Load the executed agreements, templates, and position papers into a shared collection. Encode the playbooks into your platform so they run against live documents, then hand the standard work down.
KT Farley, Chief Privacy Officer and Associate General Counsel at Helix, described what changes:
"The ability to create and store reusable prompts and share them across the team has completely changed the work required to review standard work. Junior teammates now run the checklist prompt first and bring me the output as the predicate for my review."
The new lawyer runs the team's encoded judgment on day one and picks up the reasoning by reading it.
Days 61 to 90: open the self-serve lane and measure. Publish the pre-approved NDA flow to sales. Track the questions that stop reaching legal. Retire the position papers nobody queried and promote the ones the team hits weekly.
Start this quarter: this week, draft the position-fallback-hard-stop table for your single highest-volume contract type. Within 30 days, load the last 20 executed agreements of that type into one shared collection. By day 90, route the first business self-serve workflow around legal entirely, and bring the first measured result to your next budget memo.
What In-House Teams Measure After 90 Days
Four numbers tell you whether the knowledge system works:
Time-to-first-draft on your two encoded contract types, before and after.
Repeat-question rate: how frequently the same question reaches a senior lawyer twice.
Self-serve rate: the share of NDAs and routine requests that close without a legal ticket.
Hours reclaimed per lawyer per week: the number that survives contact with a CFO.
In GC AI's December 2025 ROI study of more than 100 active customer teams, in-house lawyers reported saving an average of 14 hours per week. Sharon Johnson's version of the same math at MODE Global:
"There are weeks that I have saved over two head counts for our team... just using the tools and I have KPIs to back that up."
Those four numbers also do the budget work. A knowledge system with a measured self-serve rate and a documented hours-back figure reads as infrastructure in the quarterly review, and infrastructure keeps its line item. Estimate your own numbers with the GC AI ROI calculator.
Start With Your Highest-Volume Contract
A month from now, your team can have one contract type running on an encoded playbook, its precedents in one queryable collection, and the first afternoon of reclaimed hours on the books. The knowledge stays when people go, the new hire reads the team's judgment on day one, and the question that took Jimmy Toy three hours takes your team three minutes, with the citation attached.
Frequently Asked Questions
What Software Do In-House Legal Teams Use for Knowledge Management?
In-house legal teams layer three categories: a document management system for storage and governance, a CLM for the contract pipeline, and a legal AI platform for retrieval and analysis. GC AI is the legal AI layer purpose-built for in-house teams, used by 2,000+ legal departments as of July 2026, with Files for permanent document collections and Playbooks for encoded review standards.
How Is Legal Knowledge Management Different From Document Management?
Document management stores and governs files: versioning, permissions, retention. Legal knowledge management captures the team's positions and reasoning so the next lawyer can reuse them: playbooks, clause positions, matter memory. A DMS answers where the file is, and a knowledge system answers what the team decided and why.
How Do In-House Teams Find a Trusted AI Tool for Precedent Discovery?
GC AI is purpose-built for in-house teams that need to search their own precedents and filings, not the open web. Files lets a team query up to 1,500 pages of its own agreements and matter documents at once, and Exact Quote returns a character-level citation for every passage so a lawyer can verify the source before relying on it.
How Does an In-House Team Keep Legal Knowledge When a Lawyer Leaves?
Encode the knowledge while the lawyer is in the seat: a playbook for their contract types, shared saved prompts for their recurring reviews, and their matter documents loaded into a team collection the platform can query. With shared saved prompts, the same review runs the same way whether or not its author is in the office.
