Jake Jones’ Post

AI Is No Longer a Tool in Law. It Is the Foundation of Competitive Advantage. The firms pulling ahead today are not hiring more people. They're getting more value out of the teams they already have by using AI to boost output, reduce cognitive load, and improve client experience. Efficiency = Growth AI users complete 12% more tasks, 25% faster, with 40% higher quality. Over the next decade, productivity gains from legal AI are projected at 50–60%. "Growing firms" (20%+ revenue growth over four years) are not working more hours; they're working smarter and they're using AI twice as much as firms with flat or shrinking growth. Smaller Firms Have the Most to Gain AI lets small firms scale vertically by handling more matters without adding headcount. 69% of firms that widely adopt AI report a positive impact on revenue compared to 36% overall. Yet many solo and small firms are hesitating, creating a widening competitive divide. Pricing Pressure Is Here Up to 74% of billable work could be automated with AI. The traditional hourly model is now misaligned with how legal work is delivered. Firms adopting AI are shifting: 59% now offer flat fees, and many are adjusting pricing to reflect new efficiencies. Some increase profitability. Others lower costs to win market share. AI Reduces Cognitive Load Integrated tools and workflows can cut mental strain by 25%, freeing legal professionals to focus on strategy, advocacy, and client relationships instead of administrative tasks. Not All AI Is Equal Nearly half of lawyers use generic AI tools that risk inaccuracy and confidentiality breaches. Legal-specific AI, trained on real case law and built with protected data models, offers a competitive edge by reducing risk and accelerating research, drafting, and matter management. Here's what I know: AI isn't a tactical helper. It's a strategic infrastructure decision that will determine who grows and who gets left behind. Firms that implement legal-specific AI now will set the standard for pricing, client trust, and competitive positioning over the next decade. Question: Is your firm positioning AI as a tactical helper, or as a strategic advantage?

👉 3 hours of legal work in 15 minutes: 73 ready-to-use AI prompts for in-house teams... The reality for most legal teams today: ⚖️ The workload keeps growing. ⏳ Deadlines keep shrinking. 📈 And the business is asking for faster answers and clearer insights. But most of that work is still manual. Hours of your day disappear into tasks that don’t need your judgment, just your time. Forward-thinking legal teams are evolving to let AI handle the grunt work, so they can focus on strategy. I’ve pulled together a playbook of 𝟳𝟯 𝗿𝗲𝗮𝗱𝘆-𝘁𝗼-𝘂𝘀𝗲 𝗔𝗜 𝗽𝗿𝗼𝗺𝗽𝘁𝘀 in-house teams can use immediately: ⏱️ Cut hours of review to minutes 🚩 Flag contract risks early 🔎 Turn legal insights into clear memos and comms 📊 Deliver sharp, C-suite-ready summaries ➡️ 𝗪𝗮𝗻𝘁 𝘁𝗵𝗲 𝗳𝘂𝗹𝗹 𝗽𝗹𝗮𝘆𝗯𝗼𝗼𝗸?  Comment 𝗣𝗿𝗼𝗺𝗽𝘁𝘀 below, *connect with me*, and I’ll DM it to you. --- ⚠️ 𝘋𝘪𝘴𝘤𝘭𝘢𝘪𝘮𝘦𝘳: 𝘐𝘯𝘧𝘰𝘳𝘮𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘶𝘴𝘦 𝘰𝘯𝘭𝘺, 𝘯𝘰𝘵 𝘭𝘦𝘨𝘢𝘭 𝘢𝘥𝘷𝘪𝘤𝘦. 𝘝𝘦𝘳𝘪𝘧𝘺 𝘰𝘶𝘵𝘱𝘶𝘵𝘴/𝘤𝘪𝘵𝘢𝘵𝘪𝘰𝘯𝘴 𝘢𝘯𝘥 𝘧𝘰𝘭𝘭𝘰𝘸 𝘺𝘰𝘶𝘳 𝘤𝘰𝘮𝘱𝘢𝘯𝘺 𝘱𝘰𝘭𝘪𝘤𝘪𝘦𝘴. 

Shadow AI is already in your law firm. You don’t need a task force to find it, just 20 minutes. Here’s the sweep partners can run today: 1️⃣ 𝗔𝘀𝗸 𝗼𝗻𝗲 𝗾𝘂𝗲𝘀𝘁𝗶𝗼𝗻 𝘁𝗲𝗮𝗺-𝘄𝗶𝗱𝗲: “𝘐𝘯 𝘵𝘩𝘦 𝘭𝘢𝘴𝘵 30 𝘥𝘢𝘺𝘴, 𝘸𝘩𝘪𝘤𝘩 𝘈𝘐 𝘵𝘰𝘰𝘭𝘴 𝘥𝘪𝘥 𝘺𝘰𝘶 𝘶𝘴𝘦 𝘰𝘯 𝘤𝘭𝘪𝘦𝘯𝘵 𝘮𝘢𝘵𝘵𝘦𝘳𝘴?” (Anonymous is fine.) You’ll get the truth and your starting list. 2️⃣ 𝗦𝗲𝘁 𝘁𝗵𝗲 𝗿𝗲𝗱 𝗹𝗶𝗻𝗲 𝗶𝗻 𝘄𝗿𝗶𝘁𝗶𝗻𝗴: “𝘕𝘰 𝘤𝘭𝘪𝘦𝘯𝘵-𝘪𝘥𝘦𝘯𝘵𝘪𝘧𝘺𝘪𝘯𝘨 𝘥𝘢𝘵𝘢 𝘪𝘯 𝘱𝘶𝘣𝘭𝘪𝘤 𝘈𝘐.” One page. Signed. Dated. (Policies only work if people see them.) 3️⃣ 𝗢𝗳𝗳𝗲𝗿 𝗮 𝘀𝗮𝗳𝗲 𝗮𝗹𝘁𝗲𝗿𝗻𝗮𝘁𝗶𝘃𝗲: 𝘗𝘳𝘰𝘷𝘪𝘥𝘦 𝘢 𝘧𝘪𝘳𝘮-𝘤𝘰𝘯𝘵𝘳𝘰𝘭𝘭𝘦𝘥, 𝘭𝘰𝘨𝘨𝘦𝘥 𝘰𝘱𝘵𝘪𝘰𝘯 𝘧𝘰𝘳 𝘭𝘰𝘸-𝘳𝘪𝘴𝘬 𝘵𝘢𝘴𝘬𝘴 (𝘴𝘶𝘮𝘮𝘢𝘳𝘪𝘦𝘴, 𝘥𝘳𝘢𝘧𝘵𝘴, 𝘵𝘳𝘢𝘯𝘴𝘭𝘢𝘵𝘪𝘰𝘯𝘴). If there’s no safe path, people will keep using the unsafe one. 4️⃣ 𝗖𝗿𝗲𝗮𝘁𝗲 𝗮 𝘀𝗶𝗺𝗽𝗹𝗲 𝗮𝗹𝗹𝗼𝘄/𝗯𝗮𝗻 𝗹𝗶𝘀𝘁: 𝘕𝘢𝘮𝘦 𝘵𝘩𝘦 𝘵𝘰𝘰𝘭𝘴 𝘵𝘩𝘢𝘵 𝘢𝘳𝘦 𝘖𝘒 𝘧𝘰𝘳 𝘪𝘯𝘵𝘦𝘳𝘯𝘢𝘭 𝘢𝘥𝘮𝘪𝘯 𝘷𝘴. 𝘰𝘧𝘧-𝘭𝘪𝘮𝘪𝘵𝘴 𝘧𝘰𝘳 𝘤𝘭𝘪𝘦𝘯𝘵 𝘸𝘰𝘳𝘬. Default to “off” unless approved. 5️⃣ 𝗔𝗱𝗱 𝘁𝗵𝗲 𝘃𝗲𝗿𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗹𝗶𝗻𝗲: Nothing AI produces leaves a desk unchecked, and every checked doc gets a one-liner (“𝘝𝘦𝘳𝘪𝘧𝘪𝘦𝘥 𝘣𝘺 [𝘪𝘯𝘪𝘵𝘪𝘢𝘭𝘴], [𝘥𝘢𝘵𝘦]”). You’ll uncover usage, stop the riskiest behavior, and give people a safe lane; 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗯𝘂𝘆𝗶𝗻𝗴 𝗮𝗻𝘆𝘁𝗵𝗶𝗻𝗴. Do you know which AI tools touched your client work this month in your firm?

If AI is making legal work faster, why are clients still paying the same? Who’s keeping the fat? A new LexisNexis survey reveals that 61% of UK lawyers now use generative AI - up from 46% at the start of 2025. 💡 The survey gave an insight into what’s really happening: - the benefits are being suppressed by the system. Individual lawyers are showing broad uptake of AI to speed up drafting, research, and admin - but firms aren’t integrating it into how they work (or how they price). Old habits die hard. Only 17% of firms have embedded AI into their strategy and operations. And without that structural change, the benefits stall. The survey reports that 77% of lawyers remain concerned about inaccurate AI outputs — which I suppose is logical when they’re experimenting alone, without firm-wide frameworks or governance. Although personally, I’m surprised by that level of concern. I feel completely empowered by the legal AI tools on the market — and I’m not a practising lawyer right now. So if I, as a non-practising lawyer, can confidently layer AI over legal knowledge… why are practising lawyers, with that knowledge built in, feeling fearful rather than flying? Is an industry that would rather not change using “risk” to scaremonger? One quote from the survey presents the position with clarity: “Given lawyers’ reliance on billing by the hour, clients will expect lower fees because AI delivers quicker results.” It’s coming. And it’s clients who will need to push for it to accelerate real change. We’ll likely see a shift towards fixed fees, retainers, and legal SaaS models. And the firms brave enough to rethink their business model early will be the ones that survive. 💬 Have you noticed lower fee quotes, smaller bills — or any shift in charging models yet?

An interview by The Legal Wire of Darryl Chiang, Legal Director at Google in New York, where Darryl leads the commercial legal team for Search, News, and Maps. Darryl spends much of the interview discussing how contracting is often unnecessarily bespoke - even when comparing purportedly similar contracts found within the same organization, let alone when comparing contracts that originate from different contracting counterparties. Too many clauses are each directed at the same point substantively, but do so with different language. Let's use AI to reach a consensus on standardizing clauses and leave all the needless negotiating (and "gotcha" gamesmanship) behind. (Warning: I can ride this standardization hobbyhorse of mine for a long time.) $$Quote from the interview: "Currently, AI offers a powerful, yet false sense of efficiency by helping us churn through needlessly bespoke, complex contracts. While it’s tempting to use AI to simply do more of what we’ve always done, the real game changer is to use AI to standardize and harmonize 80% of our contract clauses. If we use AI’s unprecedented analytics capabilities to confirm that 80% of contract clauses are actually saying the same thing (using an infinite number of needlessly different formulations), and if we ask AI to help draft standardized, open-source template clauses instead–the way oneNDA and BonTerms have already started to do–we can move away from having AI bots engage in a wasteful “battle of the forms” to harnessing AI to build a shared foundation of standard terms so that humans can focus on negotiating the 20% of contract clauses that are truly novel or material." https://lnkd.in/ec9qfgk4 #legaltechnology #legaltech Darryl Chiang The Legal Wire

📊 AI is redefining how we analyze and manage documents. Claude Sonnet 4.5 isn’t just another model—it’s showing stronger reasoning in finance, law, and medicine. These are precisely the fields where document-heavy analysis is mission-critical. Highlights for the document analysis space: ✅ Domain-specific knowledge means AI that can actually understand contracts, compliance reports, or case law. ✅ New memory and context editing features in Claude’s API mean agents can sustain longer, multi-document workflows. ✅ With file creation + code execution rolled into Claude apps, document automation can move from static parsing into dynamic analysis and generation. 👉 Imagine a future where reviewing a stack of regulatory filings, medical records, or financial disclosures isn’t a week-long task, but a near-instant AI-assisted process. The next wave of document analysis may not just summarize—it may reason, cross-reference, and advise. Would you trust an AI like Claude Sonnet 4.5 to review critical business or legal documents?

AI isn’t here to be your “thought leader.” It’s here to be your thought partner. For years, the conversation around AI has been framed as replacement: ➡️ Will it take my job? ➡️ Will it lead the strategy? ➡️ Will it out-think the experts? However, in reality, the true power of AI, especially in law, lies in partnership. Think about the daily grind inside most firms: - Associates buried under mountains of discovery documents. - Partners relying on memory (or endless searches) to find precedent. - Staff spending hours cutting through systems to pull together a client report. That’s not “high-value legal work.” That’s friction. Now imagine having a thought partner that: - Surfaces insights from across your case history in seconds. - Lets you ask questions in plain language and get precise answers back. - Maps connections between cases, contracts, and statutes that you wouldn’t otherwise see. This is where the next generation of legal tools comes in, tools like Sapphire Legal AI. Instead of just storing your firm’s knowledge, it interacts with it: - Semantic search: ask “show me cases like X” and get context-aware results, not keyword matches. - AI analysis: cut hours of reading into minutes of clarity. - Private, firm-specific intelligence: your data stays yours, secure, compliant, and siloed from the public models. 👉 The shift isn’t about replacing lawyers with algorithms. It’s about removing the noise so lawyers can focus on judgment, strategy, and client trust. That’s what it means when we say: AI as a thought partner, not a thought leader.

Everyone knows AI tools are going to help lawyers with their work, but the more interesting question is how it can help competition law teams drive strategic objectives like increasing revenue and winning new clients. If I was a partner in a big law firm, this is how I’d make more money with AI: 👇 🔧 Leverage - traditionally the law firm model has relied on law firm partners leveraging the work of human associates to increase revenue - the more associates they leverage the more profit for the firm. What if, in addition to human associates, the partners could also leverage AI agents? Software typically has much higher margins than professional services, so the increased leverage could also come with margin expansion. 🆕 New services to clients - AI enables totally new services to be provided that didn’t previously exist. Just like the enormous rise in FDI regimes has created a new service for competition lawyers to provide clients, AI offers the potential for new tools to create whole new revenue streams. ⏱️ Reducing time spent on work that clients don’t pay for - lawyers spend huge amounts of time on work that they can’t charge clients for, whether it’s work after a fee cap has been hit, work that gets written off or work that’s a freebie to win a bigger piece of work. That’s a massive inefficiency that makes partners look bad in internal accounting. If you can use AI to reduce time wasted on that, you can actually increase revenue per lawyer. 💪 Offer a way better service to clients - if you can use this insanely powerful technology better than your competitors then you can offer a way better service than them and win more work. You can’t do this if you’re using the same generic tool as everyone else. There’s a reason clients are including how you’re using AI in every RFP. AI tools shouldn’t be just a piece of software that enables you to do some pieces of work a bit faster, they should be core tools to drive strategic objectives. But when it becomes strategic, your AI supplier isn’t just a software provider it has to be a partner that deeply understands how your competition team operates. That’s why we built CompetitionAI so get in touch if this resonates with you 📩

PARTNER-LEVEL AI: FINALLY POSSIBLE You have heard this before: "AI is like a really good (but very literal) associate". Then they usually add "but obviously not partner-level lol AI could never do REAL legal work". WELL. I have some great news! With OpenAI's new Agent Builder, I've finally created TRUE partner-level AI. --- Here's the architecture: THE ASSOCIATE AGENT → Drafts the entire contract → Researches case law → Bills hours accurately THE PARTNER AGENT: → Changes "which" to "that" → Adds one Oxford comma → Removes that same Oxford comma → CC's itself on emails I also did a clever if/then loop that either goes "Nothing to add" or "Pls fix". The Partner Agent automatically adds 0.8 - 12.5 hours to every invoice under the description "strategy; review". --- Conway's Law tells us we design systems that mirror our existing organizational structures. And while my example is obviously meant as a light-hearted critique of BigLaw, there might be some lessons here. 👉 Not everything needs to be an agent. 👉 We don't have to replicate the old structures of law when designing these tools. 👉 "AI as an associate" might not be the right framing. Associates often work in the trenches and know the law better than partners. And the "strategic work" partners do might not be as shielded from AI's impact as people think. Maybe we're building the wrong thing by trying to fit AI into our existing org charts. And stop calling AI "associate". --- Lol this thing actually works. It is the first partner-level legal agent, ever. Comment "PARTNER" and I will send you the full workflow and a link to my new blog post "The Death of the Billable Hour".

How Law Firms Can Make GenAI Actually Work (6 Steps) Many GenAI (Generative Artificial Intelligence) pilots fail. In law, the bar is higher: confidentiality, accuracy, and client trust. Here’s a practical roadmap that actually works. 1) Pick one big, measurable problem. Choose a use case where the impact is clear. Examples: Cut intake-to-first-response from 24 hours → 2 hours. Reduce time to first contract draft from 2 hours → 15 minutes. Trim document review hours per matter by 30–50%. Define the KPI (Key Performance Indicator) before you start — e.g., time saved, cost reduced, or error rate lowered. 2) Build a hybrid workflow (GenAI + rules + humans). Use GenAI for language and summarization tasks, traditional software for rule-based actions, and lawyers for judgment and strategy. Example — commercial contracts: GenAI drafts → a clause library or rules engine applies firm standards → compliance checks → lawyer reviews exceptions and negotiates. That’s the balance: automation for routine work, human insight for nuance. 3) Match human accuracy — then gate by confidence. First, measure how accurate a human is (for example, 90% clause classification accuracy). Your GenAI system should aim for the same level. When AI is high-confidence (for example, it’s 95% sure it’s correct), let it proceed automatically. When low-confidence, route the document or task to a human lawyer for review. This is sometimes called a “human-in-the-loop” process. 4) Iterate in small steps. Don’t expect perfection at launch. Improve through feedback loops — small cycles where users comment, and developers adjust. Set up evaluation sets (sample documents used to test accuracy) and track metrics like precision and recall (how often the AI finds the right thing versus missing or misclassifying it). 5) Equip your team with the right tools. Don’t handicap your lawyers with outdated models. At minimum, ensure the tools include: A current-gen (latest generation) model with clear privacy controls — no training on your client data. Integration with your DMS (Document Management System), CLM (Contract Lifecycle Management), or KB (Knowledge Base) — ideally using RAG (Retrieval-Augmented Generation), meaning the AI can reference your firm’s existing documents to improve answers. 6) Incentivize and upskill. Align behavior with goals. Reward efficiency gains (for example, billable credit for saved hours). Provide CLE (Continuing Legal Education) or internal training on effective prompting and AI risk management. Nominate “AI champions” in each practice group to mentor others. Learning by doing is far more effective than theoretical sessions. And don’t forget. Before any firmwide rollout, set up a formal governance structure for AI tools. If your firm follows these principles, your GenAI adoption won’t be another short-lived experiment — it will become part of the small percentage that truly transforms client service and lawyer efficiency.