What’s a robot’s daily rate?

By Fiona Czerniawska.

As we’ve mentioned before on this blog, artificial intelligence and the robotic process automation of the consulting industry is dominating the conversations we’re having with senior partners at the moment. It’s both a promise (no more war for talent!) and a threat (new entrants!).

But, for all the talk, it’s not your vision of the future that matters, but how you plan to get there, based on where you are today. We can imagine that, at some point, we’ll be sending out humans, Star Trek-like, to explore distant galaxies, but if we can’t envisage how we do that, then the idea will remain science fiction. Similarly, we can imagine a world in which partners bring with them, not armies of junior helpers, but tools that read clients’ existing systems, gather new data, and assemble all of this into a format that creates new insights and sparks profound debate. It’s possible, too, to imagine what these tools might look like. But do we know how we’re going to bill clients for this?

All professional services firms make money from people: in fixed price contracts, the fee is calculated on the basis of time and materials, just capped; even the vast majority of risk-reward contracts are calculated on the basis of inputs. Clients, so accustomed to this way of thinking, make judgements about the amount to pay based on the number of people involved. Deep expertise of specific individuals can sometimes change the detail of that calculation, but it doesn’t alter the underlying formula. So there’s a clear danger here that clients will see that automation means fewer people (remember: no more war for talent!) and bring the price down as a consequence. They’ll also look at technology firms (remember: new entrants!) offering similar services but almost certainly at a lower price. Of course, consulting firms can try to shore up their prices, such as by positioning their technology as something that supports and enhances the consulting process, rather than replaces it, but that’s going to be a difficult argument to maintain in a world that generally assumes that technology makes things cheaper. (Also, how do you demonstrate a better quality of insight?)

So let’s do a bit of imagining that’s closer to home. XYZ Associates is a typical, mid-sized consulting firm, employing around 50 people, with two partners. Those partners are typically charged out at an eye-watering level, but by the time you’ve taken into account the fact they cost the firm a lot and aren’t heavily utilised (they’re busy selling future work), the firm is lucky to break even where their costs are concerned. The vast majority of its profits are made from mid- and junior-ranking consultants, whose daily rate and high utilisation levels more than compensates even for above industry-average salaries. In our micro model, we have XYZ generating $2.3m profits on an $7.7m turnover. If you halve the number of non-partners, assuming that their jobs have been replaced by machines, revenues fall by half and profits by two thirds.

Aha, you say, that’s OK, because we’re going to charge clients for the use of the software that replaces these people. Well, maybe. For XYZ to make the same absolute amount of profits (which will surely be what it wants to do), it needs to charge the equivalent of half of its fees for its robot. True, you might decide to use your robot to compete for work on price, and accept lower profits as a consequence, but you also have to take into account the amount of money you’ve had to invest in software design and development. You also have to be aware of the consequence of pitching your new price too low, as you’ll find it nigh-on impossible to increase it later.

All of this will make for either a much less profitable consulting industry—or a very expensive robot. Either way, the choices firms make now are going to affect the money-earning capability of the consulting industry for decades.

Click here to read more about cognitive computing, robotics, and AI.