Automation will only eclipse part of consulting

Interviewed recently by the BBC on the impact automation is likely to have on the professional services sector, I was asked whether I thought that all the work done by professional service firms could be done by computers. The answer, to me at least, is clearly no.

Imagine the future of consulting—as we have in our report on the potential impact of robotic process automation and artificial intelligence—as a partial eclipse. The sun is the work that conventional human consultants do today. It, and the rays emerging from it, represent different aspects of the consulting process: discover; predict; advise; decide; design; implement; run and report. Technology is the moon. But, even as the moon moves across the face of the sun it won’t entirely obscure the latter’s light: Around the edge, a corona of human activity will continue to shine.

By analysing the amount of consulting work clients view as low-cost, standardised and repeatable, we estimate that just over 70% of the work done by today’s consultants could be automated. Artificial intelligence is more likely to have an impact on the high-value work consultants do, but we think the scale of its impact will be smaller, because it’s likely to take the form of specific tools that enhance the consulting process rather than replace it. However, the balance between low-cost and high-value, between what could be automated and what will probably not be automated, varies depending on the type of activity a consultant is doing. This is driven by the extent to which the inputs, process, and outputs can be standardised—and a good indicator of the latter is the extent to which any of those three things is overseen, managed, or undertaken by junior staff.

The table below is our estimate of the potential for standardisation: It’s very much a starter-for-ten, as this is an area we’ll be doing a lot of research about in 2018, but even in this very provisional form we think it highlights some interesting, perhaps important, points. It’ll come as no surprise that we suspect that standardisation is—or could be—highest in “discover” and “report” work. Clearly, firms have different attitudes to both these areas, with some preferring to take a more expert route, using senior people to craft an approach that’s unique to a specific client. But there are many areas of both “discover” and “report” where greater standardisation is possible. That’s not true for all areas: Consultants who are taking a board through a series of difficult decisions, based on the output of their earlier analysis, can’t follow a standard pattern but need to react to the subsequent debate.

The following chart takes that same data and puts it side by side, to create a provisional idea of how the impact of automation is likely to be felt at different stages in the consulting process. The activities that fall into the grey “shadow” at the bottom are the most likely to be automated, and those basking in the sunlight at the top are the least likely, with other activities somewhere in between.

Again, engagements differ—not every piece of consulting work will have all the stages we’ve mapped out here—so this is indicative rather than definitive. But it will, we hope, form the basis of a continuing debate in the coming year.