Articles by Daniel
-
Joining Luminance
Joining Luminance
My career has spanned major shifts in Enterprise technology and societal/cultural changes: Physical to Digital, On-Prem…
56 Comments
Activity
6K followers
-
Daniel Head shared thisAll technology for contract/legal work will be built this way. CLM and Legal-AI is coming together as a new category: Contract Intelligence.Daniel Head shared thisOne of the assumptions in AI has been that foundation models will always win. I'm not sure that's true. Legal work is a specialist domain. Contracts have their own language, conventions and risks. Understanding them requires specialised knowledge built from real-world legal data and experience. That is what led us to build Luna over the past year. This is our latest proprietary vertical AI model for contract intelligence and is the culmination of more than a decade of Legal-Grade AI development and trained on highly specific legal data. In testing, Luna achieved higher accuracy on contract understanding tasks and was consistently preferred by legal experts (in blind-testing) when identifying contractual concepts and extracting key information. I think we'll see more of this across enterprise AI. Foundation models will provide broad capability. Vertical models will provide deep expertise. An enterprise solution often requires a combination of the two. For legal AI, Luna is our contribution to that future. I shared some more thoughts here: https://lnkd.in/eewXH8UU
-
Daniel Head posted this"Legal-AI" has two problems: 1) it has a single department name in the title...but contracts are cross-functional 2) its premise is to make existing work faster...but AI is capable of transforming work. Improving individual productivity is table stakes. As far as legal work and contracts are concerned, the objective must be to unlock the value of contracts for the business as a whole; the finance lead checking liability caps, the procurement manager tracking supplier renewals etc...all from the same source of truth, the same continuously learning institutional memory. Contract Intelligence is the goal, with legal-AI as a critical component. Architecture requires an integrated system with a proprietary, domain-specific LLM alongside frontier models. No bolt-ons, no dependency on rented intelligence. Luminance has been building this for a decade. The recent wave of partnerships and pricing changes in the industry is validation that this is the direction of travel. Read more here: https://lnkd.in/e8FGXftr
-
Daniel Head shared thisAn incredible achievement from the Luminance engineers in Cambridge...Daniel Head shared thisToday, I��m proud to be able to share that a research project I’ve been leading at Luminance for the past 12 months has reached a major milestone and been officially released. When Julio Acosta-Cabronero planted the seed early last year, we initially set out to prove that we could leverage our proprietary dataset of legal documents to improve the sensitivity of an LLM to legal nuance, and use it to do things like summarising documents. We called the model “Luna". Since then, we’ve kept adding more capabilities to Luna. Now, after a dozen iterations, we’ve reached a point where it outperforms commercial LLMs on key contract understanding tasks, and can provide real value inside the Luminance platform. Frontier LLMs are getting more token-hungry, more expensive, and slower. The most rewarding part of this project has been developing a model which proves that you don’t always need a sledgehammer to crack a nut: Luna Crescent is small, fast, energy-efficient, and designed for specialised legal tasks which it executes well. Quantitative benchmarking and qualitative blind testing both confirm that Luminance users prefer Luna’s outputs, and so we’re pressing ahead to roll it out to our customers. A massive well done to Aron Molnar and Vlad Filip on my team for working tirelessly on this for months, and for obsessing at every stage over data quality and rigorous testing. On to the next version! 🌙 https://lnkd.in/eXCcKFxEThe Era of Vertical AI is Here: Introducing Luna Crescent – Luminance’s Proprietary Legal Intelligence ModelThe Era of Vertical AI is Here: Introducing Luna Crescent – Luminance’s Proprietary Legal Intelligence Model
-
Daniel Head shared thisMany tech companies have announced their intention to build proprietary AI models to improve speed, accuracy and provide optionality as pricing for generalist models evolves. Luminance is WAY ahead. Luna is trained, refined, and operational in a model-opinionated stack. The model generates text nearly 4 times faster than generalist tools, reaching 200-400 tokens per second. Work that takes 30 seconds on a general model takes under eight on Luna. Multiplied across an enterprise contract repository, those seconds become hours of recovered time and faster business decision-making. The benefits extend beyond quantitative metrics. Model ownership also delivers meaningful security advantages. Luna is deployed within Luminance’s own AWS environment, reducing reliance on external model providers and limiting data exposure to additional sub-processors. This gives Luminance greater control over performance, reliability and deployment, while helping protect customers from disruptions related to third-party availability, pricing or access. This is the difference between deep research and domain-specific development over time vs wrapper hype. Substance vs advertisements.Daniel Head shared thisToday we announced Luna Crescent - Luminance's proprietary vertical AI model, purpose-built for contract intelligence. The first wave of enterprise AI gave every company access to the same foundation models. But as AI becomes increasingly embedded in critical business workflows, a new challenge is emerging: some problems are simply too specialized to be solved optimally by models designed to solve everything. Legal work is one of them. Contracts are nuanced, context-rich and commercially significant. In Legal, “almost right” is wrong. Built on more than a decade of Legal-Grade AI innovation and trained on a high-quality subset of our proprietary legal dataset, drawn from over 220 million verified legal documents, Luna understands contracts the way they are written, negotiated and managed in practice. Compared to leading general-purpose models, Luna: - Delivers 5% higher accuracy on contract understanding tasks - Is consistently preferred by legal experts in blind evaluations - Generates answers nearly 4x faster Already powering AI agents across the Luminance platform, Luna helps transform contracts into a live source of business intelligence across legal, procurement, finance and operations teams. Because the future of enterprise AI won’t be defined solely by larger models. It will be defined by models built for the problems enterprises need to solve. Read more about why we built Luna and what it means for the future of contract intelligence: https://lnkd.in/ec_iVhCi
-
Daniel Head shared thisAlways good to share thoughts and perspectives with solid friends and professionals in the community. Thanks for having me on the new Pod Simon Gregory!Daniel Head shared thisDaniel Head joins us for Episode 1 at Primis Presents new series of "Modern CRO" 🎙️ This time diving into what really separates great teams: In his opinion it comes down to two things: 👉 The relentless ability to keep hiring and retaining top talent 👉 And the compound effect of hiring A players: great people attract more great people Because real scale isn’t about doing more yourself… it’s about building teams that can do it without you touching every part of it. Listen to the full POD here! 🔗 https://lnkd.in/eAe3euw3 🔗 https://lnkd.in/ew9gndAm
-
Daniel Head shared thisThe conversation I'm having most with enterprise legal teams right now isn't about speed. It's about trust and accuracy. Any AI can surface an answer. But what matters most to the organisations we work with is reliability and explainability. Grounded in authoritative sources. Backed by verifiable citations. Designed to inform real decisions. Shifting from AI as a productivity tool (systems of task) to AI as trusted decision-making infrastructure (systems of intelligence) is what will define our future. On May 21st, Luminance is joining LexisNexis for a live panel on how legal AI is moving beyond simple question-answering, what authoritative AI infrastructure looks like in practice, and what it means for in-house legal teams navigating complex decisions every day. If you lead or support an in-house legal function, I'd encourage you to register: https://lnkd.in/e-wJwEKi
-
Daniel Head shared thisThe deployment of Luminance agents across the contract estate just got bigger/broader...Daniel Head shared thisLexisNexis and Luminance have announced a strategic alliance that brings citation-backed LexisNexis legal content into Luminance’s contract workflows, with escalation into Lexis+ with Protégé for deeper legal analysis and drafting. The announcement is notable because it points to a broader pattern in legal AI: less focus on standalone tools, more focus on linking systems together inside existing workflows. For in-house teams, that may be where a lot of value sits, not necessarily in adding another interface, but in combining contract intelligence, authoritative content, and legal workflows in one environment. As Eleanor Lightbody put it: “AI is only as good as the data behind it.” Link to the full release in the comments.
-
Daniel Head shared thisWe're hiring - come and meet us on May 14th in London! The application and impact of AI across legal work and the contract estate is far greater than individual productivity gains and workflows. It enables things that weren't previously possible...like using agents to assess risk across millions of documents in seconds. This is why Luminance is growing...especially in the Enterprise. We're looking for talent across all GTM functions in all geos (office locations below). Come and meet us! And if you can't make it, reach out! Register here: https://lnkd.in/e9UmUBZK Luminance office locations: https://lnkd.in/ePhZUiMZ
-
Daniel Head shared thisTiming is everything. We can all work hard and, if we're lucky, do it within a great company with great people. But it's very hard to find a great company with great people with the right technology at the right time in an exploding industry with a massive TAM and accessible, modestly valued equity, that has a high chance of material growth. This is Luminance. And we're hiring. To all my friends, contacts and industry professionals both new and established...this is it! Positions are open across the business from Sales to Customer Success and Marketing to senior regional and global GTM leadership including CCO and CFO positions. Cutting through the hype in AI tech isn't easy. It's hard to separate high growth over-valued wrappers with a limited life span from high growth fairly valued Enterprise platforms with durable value. So we’ve just launched Voices of Enterprise AI at @Luminance so you can hear from clients yourself. It’s a series of conversations with General Counsel and other enterprise leaders on how AI is actually being implemented inside large organisations, and what it takes to make it work in practice. The first episodes cover how teams are approaching governance, redesigning workflows and balancing speed with risk. We’ll be releasing new conversations each month. Take a look here: https://lnkd.in/epKN9gCZ And come join us. Apply online (https://lnkd.in/ecXJQ8gQ) or message me.
Daniel Head commented on a post
3w
Welcome Dean! Delighted to be working with you! 😀
-
Daniel Head liked thisDaniel Head liked thisNo better way to kick off a new month than shining a spotlight on some of the people making Profound such a special place! 🚀 First up... Connor Joyce 🇺🇸🇬🇧 Connor moved from Boston to London as part of the founding EMEA team. Always smiling, with an infectious passion, energy, and enthusiasm. Exactly the kind of person you want alongside you when you're building something special. This morning, before most of London was awake, we kicked off September with an early morning ride around Regent's Park 🚴♂️☀️ Stay tuned for more EMEA team member spotlights... 👀
-
Daniel Head liked thisDaniel Head liked thisI technically managed Susan St. Ledger. But, she pretty much managed me. We hired Susan from Salesforce to help us take Buddy Media public. She'd brought in many of their largest customers. Most of her meetings with me went like this: Susan: "Why are you doing it like this?" Me: "We don't know any other way." Susan: "This makes no sense. We need to do it like this." Me: "OK. Then let's do it like that." I was the founder. But Susan had done this before. We hadn't. She challenged us because she knew better. And we would never have positioned the company to go public without her. Your ego wants agreement. Your company needs friction. Most founders hire people who follow their lead. The best founders hire people who challenge it. Who challenges you the most?
-
Daniel Head liked thisDaniel Head liked thisWhat a privilege to meet His Majesty The King and accept the King’s Award for International Trade on behalf of Luminance. Representing the company on such an occasion was a truly special experience and one I’ll remember for a long time. For those who know me, the chance to meet The King and talk all things AI was quite the combination. The award is a real testament to the pace of Luminance’s global expansion over the past year and to the work of the team behind that growth. I’m incredibly grateful to Luminance for the opportunity to represent the company and proud to have played a small part in marking the achievement.
-
Daniel Head liked thisDaniel Head liked thisWhen I evaluate a candidate, it’s a red flag if they haven't consistently stayed at companies for 4 to 5+ years. The average career spans 40 years, yet too many people look at their journey through a short-term lens. When a comp plan shifts, a new manager steps in, or things get tough, opting out for a quick salary increase or title bump is easy. Sometimes it's too easy to quit. Playing the long game takes grit. Here’s an example from my own career. When ExactTarget was acquired by Salesforce in 2013, most of the leadership team left in year one. I chose to stay and fight through the transition. It was one of the most challenging periods of my career, but it opened up opportunities I never imagined. I wouldn’t be at Databricks today if I hadn’t stayed. Our hiring philosophy at Databricks reflects this: -What we look for: Consistent 4–5+ year tenures. -What we forgive: A mistake or two. A short stint happens to everyone, and I’m no exception. -What concerns us: Multiple short-lived jumps in a row. This is something I’ve talked to my own children about, who are all early in their careers. If you’re 5 years in and already on your third company, it raises questions about whether you have the perseverance to navigate tough times. Stop thinking in 1-2-year sprints. Start planning your career in 5-to-10-year increments. The payoff will compound.
-
Daniel Head liked thisDaniel Head liked thisIt's classic: you never appreciate it enough until it's over. Two weeks ago, I said goodbye to Braze, five years after walking in the door. I could write something long and sentimental (I already did that internally), so here's the short version: I arrived in 2021 with a lot to learn. Four years as an IC taught me the hustle; the last year in leadership, working with the best team I've ever been part of, taught me everything else. I'm leaving having grown more than I thought possible, and none of it happened alone. It takes a village, and I got lucky with a brilliant one. Braze and its people will always be special to me ♥️, and I'll be your #1 fan on the sidelines.
-
Daniel Head liked thisDaniel Head liked thisThank you to Sean Farrington for having me back on BBC 5's Wake Up To Money yesterday morning, where we talked about the bond markets, Meta glasses, UK growth, and if society is going cashless (which made me realise the only cash in our house is in my daughter’s piggybank...??). Another great discussion on some of this week's biggest business news. 📻 Listen to the whole segment here: https://lnkd.in/eHckSfxCWake Up to Money - What's behind the bond selloff? - BBC SoundsWake Up to Money - What's behind the bond selloff? - BBC Sounds
-
Daniel Head liked thisDaniel Head liked this“Content costs plummet, experimentation speeds up, and composable intelligence changes how we make decisions.” - Andres (Andy) Trevino, Braze VP, Global Head of Solutions Consulting, just kicked off #BrazeCityxCity Sydney with a clear view on the future of marketing. Hundreds of marketers in the room today are debating the same thing. What does the future look like to you? #CustomerEngagement #MarTech #AI
Recommendations received
-
LinkedIn User
3 people have recommended Daniel
Join now to viewOther similar profiles
-
Neil Weston
Neil Weston
After more than three decades building and scaling technology businesses around the world, I've retired from executive leadership. I'm grateful for the people, customers and teams I've worked with, and I'm now focused on family, travel and selectively supporting founders and entrepreneurs.
Explore more posts
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More