Generative AI has been front and centre of the news for the last nine months and attention is often on existential risks, copyright claims or suspicions around deepfakes. However, there are a growing number of more positive ways it can be integrated into businesses.
One of those areas is customer service. The Samsung Neon people were a good example of what could be achieved with embodied AI. Samsung created an impressive suite of customer service agents whose profiles could match those of customers in need of help.
Now, any size of business can convert a one-dimensional bot into something more emotionally engaging by utilising a mix of new applications that have arrived on the mainstream market over the last twelve months. As Matthew Kershaw, V.P. of Commercial Strategy at D-id.com, a platform for creating digital people, explained, the market is exploding, “It is allowing people and organisations to create internal content for training, learning and development, and marketing at a fraction of the cost, and much, much faster”.
He explained that this allows clients to put video where video didn’t exist before. Where once we may have been presented with a boring document, now those can be turned into interesting videos or numerous imaginative scenarios. And that is where the corporate interest is coming from. He tells me that at his best guess the platform is signing a new deal every three or four days, with clients using generative AI to make training content better understood, more completed, and to make sales content more clicked on.
But how easy is it to create a digital person, also known as ‘synthetic media’ or, a ‘synthetic self’?
Matthew claimed it was very easy, “Our process is incredibly simple, you literally just upload an image, it could be a face you’ve generated, it could be a picture of you now, or it could be a picture of you when you were 20 years younger…Then you either type in text, and tell it the language you want it to speak, it could be one of 220 different languages. Or, you can upload spoken audio you’ve recorded elsewhere and that’s it. Then you press a button…a few seconds later a video is generated of that face, talking those words”. I also reached out to D-id.com’s major competitor for comparison but did not hear back.
It sounded so simple I endeavoured to try it out for myself. I knew I wanted one or more variations of an avatar that was modern, perhaps even futuristic and something that was a little ‘uncanny’, so that it had some resemblance to my real physical self but also looked a little artificial, and better than me, too. I could not create that myself so I engaged the team at Superheroic. The first thing they did was put me through their consultation, the Synth Boutique, to uncover what it was I really needed.
Having uncovered the cultural, commercial and functional needs for “The Synth”, I supplied a selection of headshots. The team then used Midjourney to blend a plethora of images, crafting a brand world and a series of characters for me. We agreed that the fourteenth character image was the best to begin with, and that became known as TF14. Thus TF14 became the first member of my ‘AI entourage’ as Ian Crocombe, of Superheroic, calls it.
The team then set up TF14 on the D-iD.com platform, using Eleven Labs, the voice cloning technology, to clone my voice for situations where I would not be able to record a bespoke voiceover every time. After a quick demo of how to use the platform, I was off on my own.
In fact I have been able to record my own voice and apply it to the image of TF14 every time I have needed to, and I have found a multitude of uses for this synthetic self in my business. I have sent TF14 off to follow up on emails, to reconnect in friendly ways with people across at least three messaging apps, to chase up invoices (something no-one ever really wants to do as their real self) and to front my podcasting promotional marketing. I have even provided customer reviews in avatar mode.
This is early days for digital people, and for my experimentation with digital personas in a business context but one can see that before soon there will be a shift from marketing, training materials and promotional content towards even greater utility, where agents such as TF14 will be used much like plug-ins. Crocombe thinks this is the next stage, “We’re especially interested in what happens when you hook up avatars to other systems which enable them to act as agents on your behalf, a simple example could be an out of office video, but there are lots of different scenarios these agents can be used for.”
Matthew Kershaw seems to agree, he told me, “I do look at the lifestyle of the Roys in Succession, these super rich people. They don’t hold on the line when they need to renew their insurance, they have a person for that. They have people who go out into the world and do stuff, their PR, their finances, and the buying of houses… well, it would be great to democratise that. I want to have digital people who can go out in the world and interact possibly with other digital agents, and negotiate on my behalf and get me the stuff I want without me having to hear “please hold, your call is important to us”.”
These days we do seem to face a daily struggle of domestic and business administration and bureaucracy, so the idea of a team of agents who can cut through that complexity should not be underestimated. These synthetic selves, or digital people, might not be substitutions for communicating with a human being but where that is impossible due to the sheer number of tasks to complete or the time available, a synthetic “me” or “you”, acting as an agent, working on our behalf, would do very nicely indeed.
You can listen to the full interview with Matthew Kershaw on The Future of You podcast.