Lucia Elizalde, a lawyer and behavioural scientist at Elizalde Consulting, has over 25 years of experience in law firms and in-house as a litigator, innovation leader and behavioural scientist. She moved into behavioural science after becoming frustrated that people do not change as quickly as organisations want them to. She now helps organisations secure genuine buy-in for change. Andy McDonnell, an attendee, joined with a market-focused challenge on junior lawyers and law firm economics.
Overview
Law firms are investing heavily in AI but often struggle to show a return. Elizalde argued that the answer lies in behaviour rather than the tool, endorsing the 10/20/70 rule cited in the session: roughly 10 per cent of value comes from the algorithm, 20 per cent from software, data and system connections, and 70 per cent from driving change within the organisation. Her concern is that firms treat behaviour as guesswork, when what people say and intend is not necessarily what they do.
Her approach starts with silent observation rather than surveys. In a two-year project, she sat with lawyers for an hour each week to watch how they used technology, then complemented this with usage reports and surveys. Usage data is easy to collect but says little about business impact. She contrasted a lawyer who uses a tool often for simple tasks with one who uses it occasionally for complex work, checking output carefully and improving the final advice; the second is the one connecting the tool to its purpose. Meaningful measurement needs a clear objective. If the aim is faster turnaround, measure whether a first draft reaches the client within 24 hours; if it is fewer errors, track how often lawyers must correct AI drafts. She also warned of an efficiency fallacy: if drafting takes 30 per cent less time, firms should ask what lawyers do with the time saved. Her framework links activity, behaviour and impact, with behaviour at the centre.
Elizalde said many firms buy technology out of fear of missing out or to meet client expectations, without defining why. McDonnell agreed that firms have little choice but to buy, but must understand how the technology will change their business, clients and people. A tension emerged on junior lawyers. Elizalde said AI should replace no one, while McDonnell argued that the market reality may mean fewer juniors and that the leverage model may no longer be viable. Elizalde saw no serious conversation yet on changing that model. McDonnell added that firms have never had to think this way and may lack the resources to do so.
Key Takeaways
- Define the objective before the purchase. Without a stated reason for an AI investment, you cannot decide which behaviours to measure.
- Observe people at work. Stated views and surveys often reflect what people think you want to hear, so watch real workflows and use usage data only as a supplement.
- Look beyond usage figures. Frequent use does not show business impact; ask what people do differently, in what circumstances, and with what consequences.
- Track what happens to time saved. Efficiency gains matter only if you know whether they go into client contact, business development, or elsewhere.
- Confront the leverage model. If AI reduces junior work, firms need honest discussion of profitability, talent pipelines, and the new skills juniors will need.