What Happens When No One Ever Leaves the Company?
Digital Twins powered by AI could bring up some interesting choices about who owns your work profile and the future of knowledge management, onboarding, etc, etc.
Way, way, back like a year and a half ago, before ChatGPT and AI in general started sucking all the air out of the room, there was a lot of talk around “digital twins.”
There was talk about digital twins in health care. Manufacturing. Even the metaverse. McKinsey defines a digital twin well enough as “a digital representation of a physical object, person, or process, contextualized in a digital version of its environment.” Lots of the talk around digital twins even when it was at the org level was creating a twin of the whole org. Those were all cool and had potential but now, I’m thinking about what it would look like at much more granular level, an atomic level. What would it be like for your entire team, org, and company, to have a digital twin of you and all your employees…forever.
Think about it like this: you join a company and immediately you’re given kind of a standard template of a twin. It lives on your AI-enabled laptop and can access all the knowledge of the enterprise systems. It’s built maybe on the job description and some generic indicators from your resume. Think of this initial instance as that blank profile picture when you join a new social network before you upload a pic of yourself. The trick is that as you interact with your new company’s systems - as you have meetings, take notes, take training, IM coworkers, write docs, provide your 360 feedback at assessment time; all of that gets collected and added to that default template. Over time then, as it collects more and more data, that generic picture resolves into something that looks more and more like you.
So you spend a few years at that company. You do excellent work but then it’s time to move on, to take on a different challenge at a new company. I used to call the problem of organizational knowledge retention, the collected wisdom of the group; the “bus problem.” The idea, as morbid and potentially Addams Family as it sounds, was that we needed to think about what would happen to that knowledge if any one of us walked out into the street and got hit by a bus.
Tough break for us and a very important lesson about traffic safety but what about the org and the lost knowledge? All of a sudden, maybe that’s not a problem. You may be stuck under a Sound Transit wheel but your digital twin is safe. It knows 90% of what you knew and the rest it can infer because it knows everyone else in the org. It knows your writing style, how you handled tough conversations, all the shortcuts you found through the byzantine maze of enterprise applications and thanks to the UI of the new LLMs, we can just chat with in like it was you on the other end of the IM. Let’s make the picture a little less violent and morbid, let’s just say you left the company to take that new job and you remembered to look both ways.
You pack up your desk - if you had to RTO - but do you pack up your digital twin? I mean it is you-ish. Your digital likeness in a lot of ways. Do you just have to walk away from it? Can you take it with you? If its staying behind, does it get paid? Do you get paid something like residuals if the old company wants to keep employing your digital twin or is it work product? I can imagine a time when your digital twin, just like you, carries the lessons learned and best practices that you accrue over several jobs. What if you brought that part of your twin with you when you came to your new company and that carried the preferences of how you liked to learn, work, and communicate, and that foundling twin could talk to the new job’s enterprise systems and all the other digital twins of the folks on your team, and it could find the happiest medium between all those styles and map out interactions based on those “vibes.” What belongs to the company and what belongs to you?
In one sense, this is a dreamscape for the KM folks. No more institutional memory going by the way. The ability to access information in a new way - like when you ask “hey how do you do X?” And someone says “I don’t know but Jane was really good at it.” Instead of now trying to track down Jane’s backfill and see if Jane left any relevant material, you can now just ping Jane’s digital twin - or would it be like the ghost of her digital twin since Jane isn’t here anymore? It’s like everyone is still here - just over there on Slack or Team.
Forget your school transcript or your LinkedIn profile - you provide access to your digital twin to the company wanting to hire you and you let them run your digital twin through some relevant exercises to see if you can actually or have actually done the work. Then if you get hired, you negotiate different access levels for longer times for your twin. Maybe you also negotiate, up front, what kind of information your twin will be able to take with you when you leave. Now we can start to scale up our models - from the individual to the team to the org - except now our model isn’t some generic blob of an org but an up-to-date replica of your org from the atoms up. That means the models of anything you run against it should be much more predictive of actual outcomes than others. Think about being able to to isolate for say the impact of training on performance because we can now isolate and control all the other variables? Think of all that value that could accrue to the org and then think of the value it could bring to you.
