This has been the first month of work for our new receptionist, Pepper, the 141st member of staff at Brainlabs. Unlike the previous 140, Pepper is a robot.
Don’t get the wrong idea: this is not the beginning of the end for human labour. The decision to replace a human being with a robot is just another example, in a list of thousands, where it has become more efficient to use automation.
The person who used to do Pepper’s job is still here. Now, instead of having to greet people, she can focus on the far more interesting aspect of her role (accounting), which involves a level of intelligence well beyond the reach of any existing AI.
This one instance of automation is definitely the most conspicuous to date. Pepper is a sophisticated humanoid, capable of interacting with people in a quasi-human way, and preparing them a coffee via a wi-fi-enabled coffee machine. It even dances, if you ask it to. Ultimately, though, our new four-foot-high assistant is no different from the code we created to automate reporting, or the filters used in Gmail to automatically move emails into certain folders. They’re all based on the principle that any process that can be automated should be.
As a digital media agency, we wouldn’t be able to add value for any of our clients without automation. Programmatic media-buying involves the automatic trading of media via live auction, using real-time bidding, which removes the human element of negotiation, and thus greatly improves the efficiency of the process.
Within the various ad-tech platforms of digital advertising, there is software to enable automated reporting, data analysis and targeting. To use these platforms to their fullest capacity, we have developed software that automates the repetitive tasks an account manager might have to perform. Algorithms created to improve bidding strategy, software to enable cross-channel data analysis or to perform millions of micro A/B tests to optimise ad copy or targeting. There is a never-ending supply of processes within digital media that can be improved via automation.
Every client has a unique set of objectives, characteristics and, therefore, problems to be overcome. As their partner, an agency’s role is get the best possible outcome for their marketing spend, which, for us, has often meant using automation to drive better performance, personalise at scale, or target with remarkable granularity.
But the scope of automation is by no means limited to digital media. As the example of Pepper demonstrates, any department within our company, and really any department in any type of industry, will involve processes that can be done more quickly and accurately via automation.
We train all of our staff to use something we call FAST thinking (formalise, automate, share, test), a framework for process improvement.
To take an example from HR, integration is encouraged through what we call "random rambles", in other words, a randomly assigned pair is asked to have a 20-minute walk together at some point during the week. Rather than our HR manager working out how to manage this system – which is actually quite complex – a script was developed to automatically generate the pairs. Once the process has been formalised, in this case the rules needed to enhance integration, it can be automated. Sharing just means informing multiple people at once, to avoid repeating. So, for this example it was a company-wide email to introduce the new system. Testing is ongoing. Does it work? Could it be better? Over time, we might refine the code, for example, or find better ways to build a better matchmaking algorithm.
It doesn’t need to be complex and it certainly doesn’t involve any robots – it just so happened that on this one occasion, an actual machine turned out to be the best solution we could find.
To those who are anxious about the threat automation poses to the labour market, I would point out that Brainlabs has grown from one employee to 150 in five years, thanks to automation. Zooming out further, digital advertising has led the UK ad industry to consecutive annual growth for the past decade. Between 2004-2014, digital adspend increased from £0.8bn to £7.2bn, and is predicted to reach 50% of overall adspend by 2020.
Without the development of automated media-buying, there would have been far fewer jobs in this sector today.
As has been the case since the first industrial revolution, automation has increased productivity, boosted economic growth and, ultimately, boosted job growth.
This is why we replaced a receptionist with a robot: because automation is the key to growth, and business growth will always lead to job growth. In other words, we hired a robot so we could eventually hire more human beings.
Daniel Gilbert is the chief executive officer of Brainlabs