Are you going to be replaced by a robot or a Superuser?

OK. OK. Just hear me out. It’s not quite that simple.

There has been a lot in the tech press in recent months about Robotic Process Automation (RPA). That is, the use of software (with maybe a smattering of AI and machine learning) to handle high-volume, repeatable tasks that previously required humans to perform. Last year, the investment rock stars at Softbank (not content with having nailed it with WeWork and Katerra) put $300m into Automation Anywhere. Unlike task-focused bots, AA’s Digital Workers can be trained to do highly-skilled tasks, taking on the persona of a colleague in payroll, accounts or FM.

The processes and activities that lend themselves to RPA tend to have the following characteristics:

• rule-based;
• transactional (requiring limited judgement);
• electronic inputs and outputs;
• high volume and repetitive; and
• unlikely to change over time.

Quite a lot of what we do in engineering design falls within this list: for example, much of detailed design, loading calculations, heat calculations and quite a lot of CFD for starters. Clearly, engineering by its nature is rule-based and the laws of physics are unlikely to change over time.

So it could be that Elsie the Engineering Chatbot is set to replace the jobs of some engineers. But it is probably much more likely that engineering tasks will be undertaken by an algorithm or code that carries out the calculation in question. This Functions-as-a-Service (FAAS) model is being developed by Hypar.io. Hypar is a web-based cloud platform and API which executes users’ code, in Python and C#. Its data model supports IFC classification; it offers a common user interface and is BIM platform independent. It is effectively an ‘ecosystem’ in which algorithms can be shared or sold – an ‘app store’, if you like, for engineering processes. Importantly, once a process has been captured as an ‘app’ – that’s it. It’s there for everyone to use, and it does not need to be created a second time.

The Hypar founders have massive Autodesk pedigrees. Ian Keough started Dynamo at Buro Happold before joining Autodesk and overseeing its development there. Anthony Hauck led product R&D for the Autodesk Generative Design group. Investment has come from Building Ventures, who sold Sketchup to Google and Vico to Trimble. These guys know what they are doing. Expect to hear much about Hypar in the years to come.

So, if you factor in Hypar-type functions as a way in which engineers can be replaced by robots, the 1.4% figure is starting to seems very much on the low side.

“Yes John. We see your point. But this really doesn’t apply to us. Our jobs include lots of the factors that make computerisation very difficult: persuasion, negotiation, originality and social perceptiveness. So this really doesn’t apply to us.”

(Social perceptiveness? Really?)

OK, well in that case, I agree with you. You’re not going to be replaced by a robot. No. You’re going to be replaced by something else – a Superuser.

The concept of the Superuser and its impact on the AEC industry is set out in Randy Deutsch’s brilliant book [2], which was published last year. Superusers can blend together all the technology, data, interpersonal and design ingredients necessary to come up with a solution that meets the needs of the client. They are “T-shaped” people, where the horizontal bar across the top indicates not only their ability to deploy expertise from the widest range of disciplines, but also their skill in bringing the best out of other people. Deutsch stylises a Superuser as being a combination of architect, hacker, data scientist and algorithm builder.

The architects and engineers of the future are going to have to possess a range of skills and attributes way in advance of those that we see in the workplace today. Leading engineering organisations are already starting to respond to this need. Arcadis have sent more than 7,000 of their engineers and technical professionals on a digital base camp – a training programme run by Techstars. Buro Happold are embracing computational engineering, insisting that coding and scripting are for everyone. They are developing a ‘sandbox’ of tools to encourage engineers to solve problems using coding and data.

Speaking at the ACE Digital Leadership Conference last month, Caroline Gorski, Group Director and Head of R-squared Labs, set out Rolls-Royce’s requirement for 20% of engineers to be able to code in AI languages such as Python within the next 2-3 years.

So in the future, RPA (or Hypar-type functions) will be used for high volume, low risk work and for keeping clients up to date. Superusers will tackle complex problems requiring different forms of data, coding, software and economic analysis.

Whether you are replaced by a robot or a Superuser depends on the nature of your role. But your role will change. And you have a much better chance of becoming a Superuser than a robot. So why not give it a go? Python for Dummies is available from Amazon for only £13.63. (Well, it’s a start.)

And don’t be too miserable about it. At least you’re not a pastry chef!

More News