European in-house legal teams feel real pressure to adopt AI, but ownership, budget and headcount haven't kept up. That's the headline of The State of In-House Legal Innovation in Europe, a new benchmark report from SpotDraft in partnership with Sarah Irwin of ITGC.
Over four months, the report surveyed more than 80 in-house leaders across 11 European countries. It's a deliberately senior view: 75% of respondents were GC or director level, and 54% were GCs or CLOs. The result is a data-led benchmark of how Europe's legal teams approach AI, legal tech, budgets and data, with 12–13 findings in total.
Download the full report → Link
Sarah hosted a webinar with three practitioners who put the data in context:
Below are five of the report's headline findings, with the panel's advice on acting on them. The rest of the data, including where teams are using AI today and how data-driven they are, is in the full report.
63% of respondents said the GC, CLO or head of legal drives innovation decisions. Only 11% said legal operations own it, and 17% said there is no clear owner at all.
Leadership clearly supports innovation, but few teams have carved out a dedicated function for it. Emilie Pounds confirmed this is very much part of the GC job now, and it's explicit: her exec team asks directly which tools legal is using and what has actually been implemented. In most scale-ups, she said, a dedicated legal ops hire is still a luxury.
Helen Loots warned that without an owner, innovation becomes a hot potato passed between the GC, junior lawyers and assistants. Her bigger point was about skills: being a great lawyer doesn't automatically make you good at legal ops. Whoever owns it needs real training in process, data and change management.
Two practical workarounds came up:
Sarah also found that innovation projects make great onboarding work for junior lawyers: low-risk, creative, and a natural way to build relationships across the business.
Efficiency demands topped the list of reasons teams feel pushed to adopt AI. Curiosity and experimentation, staying competitive, CEO and executive pressure, personal career security (selected by 47%), board expectations and FOMO followed. Vendor influence registered as no pressure at all.
The panel itself rated its pressure at 4–5 out of 5. Emilie described it as coming from every direction, top-down and bottom-up. But the career-security figure didn't resonate with her. She sees AI fluency the way she sees horizon scanning or using Slack: simply part of the job in a fast-moving business, not a threat.
Helen framed it through Richard Susskind's idea: lawyers won't be replaced by AI, but they may be replaced by lawyers who use it.
Sabrina welcomed the vendor's result. Pushing tools on lawyers who aren't ready doesn't work, she said; most teams who call SpotDraft are lean, stretched and can't add headcount. That makes a consultative role more valuable than a feature pitch.
Is AI actually giving time back? Helen's answer: yes and no. Big implementations need serious time investment. But small wins are easy to find. As a non-native English speaker, she uses Copilot to turn a frustrated draft into a polished, structured email in seconds. Her tip: spend one coffee break a week thinking about where AI could help, bank the small wins, then take on the next challenge.
Only 31% of respondents reported a bigger legal tech budget in 2026 than last year. 26% saw no change, 5% saw a cut, and a striking 22% have no legal tech budget at all. Budget was also the single biggest barrier to AI adoption (33%), ahead of team capacity, unclear ROI and change management.
The will is there; the resources often aren't. The panel shared how to close that gap.
Start with your own data. Sarah's simplest advice: have everyone on the team track what they work on, why, and how long it takes for 30 days. If the GC is spending hours a day on NDAs, no leadership team will want that to continue.
Fix the process before buying the tool. Helen has seen the same mistake at Unilever and at a family-owned engineering business: a flawed offline process, a new tool on top, and the hope it gets better. Ask your team what frustrates them, find the pain points, and measure the time they cost.
Compare internal and external spend. Helen also suggests looking at how much goes to outside counsel versus your own team. Spending four times as much externally is a strong argument for investing in-house.
Make the vendor build the ROI case. Sabrina's advice is to put the business case back on your vendor. They build them every day; you may be doing it for the first time. Two tips: learn when and by whom budgets are set internally, and give the vendor your real numbers (contract volumes, request counts, time per request). Your own figures carry far more weight with a CFO than generic stats.
Do the market homework. Emilie, whose CEO actively wants AI adoption, still had to justify a specific tool. A thorough review of the options, done by a team member, let her make a clear, segmented ask and explain why this tool rather than another. She also rated dedicated legal tech over a general model like Claude on security and on vendor responsiveness when you need something built or fixed.
Benchmark your budget against your peers. The full report breaks down legal tech budgets and the top barriers to AI adoption across Europe, so you can see where your team stands before your next budget cycle.
70% named accuracy as their top AI concern, followed by data security, internal capability and regulatory risk. Not one respondent said they were worried AI would replace them.
Why is the replacement story so loud in the media but absent among practitioners? Emilie pointed to history: the printing press, Amazon and the bookshops. Every new technology goes through a hyperbole phase. People who use AI daily see its gaps as clearly as its strengths, and lawyers in particular won't accept output that's only "70% right".
The new accuracy problem is coming from the business. Emilie described colleagues drafting contracts with AI and asking for a "quick sign-off", or running her in-progress markup through a chatbot and sending back a competing version. She has even reviewed an AI-written document commented on by someone else's AI: two models in conversation, and no one fully engaged. Far from saving time, it creates more work. Her fix is education and clear service levels. Recently she offered a colleague an AI-assisted quick look today or a full review later, and they agreed on the trade-off up front.
Helen's key line: you can never make the AI accountable. Responsibility stays with the person who sends the output or acts on it. She favours an 80/20, "done is better than perfect" approach where the risk allows it. A self-service NDA is fine; other contracts need guardrails so the business knows when to bring in legal.
Sabrina noted the tension in the data: wanting 100% accuracy implies wanting no human review, yet a perfectly accurate AI is exactly what would replace lawyers. Today's legal AI is built to save time on repetitive work, with a human still in the loop. And accuracy depends on context: when output looks off, giving the tool more detail behind the ask usually fixes it.
Respondents see Europe as competitive but behind the US. (Most came from the UK, Ireland and the Nordics.) Sarah challenged the "US innovates, Europe regulates" cliché: she's been impressed by how freely European teams are embracing AI, and wonders whether regulation actually forces teams to understand the technology properly.
The panel pushed back on "behind":
What about legal engineers? Helen sees them fitting best in larger organisations, and cautioned against hiring for a cool title. Data, process, technology and people all need to be in order first; otherwise extra people on top won't fix it. Sarah added that fractional support can prove the case before committing to a full-time hire.
The report closes with five recommendations for teams ready to move beyond experimenting:
Europe's in-house teams aren't afraid of AI, and they aren't waiting to be sold to. They're stretched, short on budget and often without a dedicated owner for innovation. The teams making progress start with their own data, fix processes before buying tools, keep a human accountable for every output, and make their vendors do the heavy lifting on the business case.
This webinar covered five findings. The full report has all of them, built from the answers of 80+ senior in-house leaders across 11 countries. Inside you'll find:
