Where is your firm on the AI curve?
Praxity’s AI adoption Thought Leadership Summary Snapshot
Most firms now have access to AI and are integrating it into the business in different ways. What separates them is how far they have travelled from experiment to value.
Our AI adoption thought leadership summary below gives you a snapshot of the full report and allows you to plot your position and consider the questions you should be asking.
Access to AI is no longer what sets firms apart. Most have experimented, and many now use it every day for drafting and first-pass analysis. What differs is how much value they draw from it, and that has less to do with the tools than with the conditions underneath them. A firm whose data can be trusted and whose processes were designed with some intent will get more from the same technology than a firm still working around years of accumulated shortcuts.
That is why the more useful question is not whether a firm uses AI but how far along it has come. Firms tend to move through a recognisable progression, from getting the basics in order to using AI to create real commercial value. The report frames this as a curve with four stages, and knowing where your firm sits on it is the first step towards deciding what to do next..
Conversations across the Alliance point to a common shape in how
firms progress, even when the detail differs from one firm to the next.
It is rarely a straight line. Some firms move quickly through the early stages and stall later. Others spend a long time on the groundwork before advancing steadily. Progress also loops back on itself, as new tools or new data expose gaps that were easy to overlook before. Placing your firm on the curve is less about scoring yourself and more about seeing clearly which stage you are working in.
1. Foundations
The first stage is about confidence in information. Firms here often discover that AI tends to expose existing problems rather than create new ones. Data sits across disconnected systems. The same task is recorded differently from one team to the next. Outputs are only ever as reliable as the material feeding them, which is why this stage tends to be less glamorous and more decisive than the ones that follow. You are probably at this stage if the honest answer to "can we trust the information our tools are working from?" is "it depends." As Adam Brodie, Founder of Finerva, puts it in the report, "If information isn't in the central system, it's garbage in, garbage out." The work here rarely makes headlines, yet it sets the ceiling for everything above it.
2. Flow
Once information becomes more reliable, attention turns to how work moves through the firm. This is the stage where routine tasks start to be streamlined and the first real time savings appear. It is also where firms tend to notice how differently the same job is done across teams and offices, because automating a process forces you to define it first. You are at this stage if the tools are clearly saving time in places while the overall shape of the work has stayed much the same. The gain is real, though it is still efficiency rather than transformation
3. Insight
As friction falls away, the value shifts from producing information to interpreting it. People spend less time assembling numbers and more time explaining what they mean. Decisions that once waited on a report can be reached in the conversation itself. This is often where clients start to feel the difference, because the firm is quicker to the point that matters and freer to focus on judgement. You are at this stage if your teams are reaching the "so what" faster, and if AI is beginning to improve the quality of client conversations rather than only the speed of the output.
4. Value
The most mature discussions move past efficiency towards growth. Here the question changes from how to do the same work faster to what the firm can now offer that it could not before, whether that means new services or deeper client relationships. Some firms are already seeing AI shape how prospects find and choose them, well before any conversation takes place. The shift in mindset matters more than any single application. As Adam Brodie frames it in the report, "For every dollar of cost saving, I want firms to be thinking about four dollars of revenue-generating opportunities."
"For every dollar of cost saving, I want firms to be thinking about four dollars of revenue-generating opportunities.
Adam Brodie, Founder of Finerva and Partner at RouseFinerva
One theme runs through every stage, and it grows more important the higher a firm climbs. The further AI reaches into the work, the more human judgement has to be applied to what it produces. Confidence in source material and clear ownership of what the firm signs off are what keep the technology an asset rather than a liability. This is the point where a firm's reputation is most exposed and where the report is at its most direct. As Garth O'Connor-Price, Partner at William Buck, observes in the report, "Used well, AI is an early-warning system rather than a source of answers. It can flag something worth examining, but the moment a model looks most confident is often the moment to test it hardest. Responsibility for what we put our name to stays with us, and no algorithm changes that."
None of this happens in isolation, and that is where belonging to an alliance changes the picture. Firms at different points on the same curve can compare notes, so those just starting out can learn from the ones further along, while the ones in front can pressure-test their thinking against peers facing the same demands. A single firm learns only from its own experiments. An alliance of independent firms learns from everyone's. That collective view is difficult to build alone, and it is one of the clearest advantages of working within the Alliance.
Most firms will recognise themselves in more than one stage at once, which is entirely normal. The value of locating yourself is knowing what to prioritise now and what to leave until the groundwork can support it. The full report, From Experimentation to Value, sets out each stage in detail, with the diagnostic questions to help you place your firm and the full framework to guide where you focus next.
"Used well, AI is an early-warning system rather than a source of answers. It can flag something worth examining, but the moment a model looks most confident is often the moment to test it hardest. Responsibility for what we put our name to stays with us, and no algorithm changes that."
Garth O'Connor-Price, Partner at William Buck