Gone but not lost: why digital doppelgangers are the future


Goodbye employee. Hello digital doppelganger

Losing an employee comes at a cost, but technology that taps into the leaver’s digital afterlife could help employers to bring new recruits up to speed. Oliver Pickup reports


It costs £30,614 on average to replace a departing employee, according to a report published by global forecasting analyst Oxford Economics(1). There are two key factors here: the logistical outlay of recruiting and absorbing the replacement worker, and the cost of lost output while a new employee learns the job.

The report suggests that at large organisations with more than 250 employees it takes, on average, more than half a year (28 weeks) for new workers to reach optimum productivity level. That number is only slightly better for small to medium-sized businesses (24 weeks). And for microbusinesses, with up to nine workers, that level can be hit within 12 weeks.

A failure to adequately transfer knowledge is the issue. While explicit knowledge is easy to catalogue (with handover notes and an onboarding process), it’s trickier to quantify, save and share implicit knowledge which relates to experience, context and interpersonal relationships.

The digital double

It’s no surprise, therefore, that businesses are looking to tech to help. Wouldn’t it be great if there was a way to create a digital doppelganger of the departing employee? Someone who retained their knowledge and experience and could get up to speed almost instantly?

In his dystopian near-future drama Black Mirror, author Charlie Brooker addressed the idea of a digital afterlife – where a person’s social media activity can be crafted after their death into a digital avatar that mimics their personality with eerie accuracy.

And it’s not just fiction – in the spring of 2016, three months after the death of her best friend Luka, entrepreneur Eugenia Kuyda developed Replika (‘an AI friend that’s always there for you’) by feeding thousands of text messages into a neural network to create a chatbot in Luka’s image. Anyone can now use this – inputting their own data, in the form of old text messages and social posts, to create a digital double.

The implications for the workplace are huge. “Corporations could quite easily key log an employee’s activities through their keyboard,” says Marcus John Henry Brown, a Munich-based technologist, speaker and author. “The type of work you do, what you write, and the patterns of your activity can all be recorded. Realistically, you could feed this data into an algorithm, and within a week it would have determined how you work.”

Transferable skills

While this technology is not yet being used (or with companies not openly admitting to it), progressive organisations are looking at other ways to aid knowledge transfer. TVA, Siemens, and Delta Airlines all routinely survey their best employees with the most tenure to find out the skills they need to preserve. Nasa’s Jet Propulsion Labs videotape their veteran employees lecturing on what it was like to be part of old missions, and use it to help socialise new employees.

Other companies are going one step further, developing tools to help them capture, share and reuse knowledge. Intel created Intelpedia, an internal, company-wide Wiki for employees to reference, while software company Panopto uses an enterprise content management (ECM) system that stores all content related to corporate processes, from documents to social content such as blogs, Wikis and discussion feeds.

Meanwhile, General Electric is using a cloud-based platform called Predix to record and store workflow for future use. New engineers have access to real-time analytical data and the responses of previous generations, which they can apply to new situations. In this way, future engineers can take advantage of colleagues’ past experiences, even after they’ve retired or left the company.

Why digital doppelgangers are the future

How can businesses quantify, save and share implicit knowledge when an employee leaves?


Algorithmic insight

Clearly, with all this data being collected, the stage is set for AI. One early adopter of the technology is Altify, which uses AI to coach salespeople, using insight gleaned from interactions with more than a million sales professionals.

The hope is that Altify will minimise the time spent onboarding new employees. “If a salesperson is expected to sell £100,000 every month, it’s crucial they get up to speed as quickly as possible,” says Altify’s executive chairman, Donal Daly. “Otherwise, every month equals £100,000 in lost revenue. If you have 50 new salespeople, it becomes a £5m problem.”

If your own company also has a lot of implicit knowledge data – ideally collected from your employees before they leave – there are AI tools that can help turn it into something useful. Deepgram transcribes insights from phone calls, video footage and online, while TalkIQ  offers critical insights about your customers’ conversations. Meanwhile, Sundown  automates repetitive tasks within your business. And if you’re looking to replicate tone of voice, Google Mail’s Boomerang Respondable is an AI assistant that helps you write better, more effective and more actionable emails in real time.

“When it comes to workplace knowledge and the prospect of a digital afterlife, there’s a constant process of evolution, change, and streamlining,” says Marcus John Henry Brown. “And it won’t be as bad as people fear. If technology and AI are able to take on more of the tasks that humans consider chores, our employment will become more enjoyable and rewarding. It won’t be about technology dehumanising the workplace – it’s allowing technology to take away all the work that dehumanises us.”


Oliver Pickup is London-based writer, specialising in business, lifestyle and technology

Additional reporting: Hannah Hudson


(1) http://www.oxfordeconomics.com/my-oxford/projects/264283