Martin Thompson’s Post

Transcript

In this video series, we're covering AI in use in the enterprise, and I'm suggesting that we manage AI like an asset. And that's not just for the cost and the risk, but that's for managing the innovation, unlocking the innovation and allow people to innovate faster. So in the last video we spoke around the fact that AI is showing all the hallmarks of classic IT sprawl in the enterprise. Lots of experimentation, not much ROI thought about just yet, and not much governance around it. So in this video I'd like to cover what AI is actually in use. So according to ITAM Forum research, there were three main types that were identified. And I'll go through these in ascending order. So the lowest of the three in use was something called predictive analytics and machine learning, and that was around 15. Fourteen, 15% of organisations were using this. And this is whereby you might have like an assembly line or a supply chain or a system that's throwing out data and you can use an algorithm to analyze that data to predict what might happen in the future. So you might have an assembly line and you could predict when that assembly line is likely to have a fault. Based on previous faults, and then you can be proactive to to reach out to fix that before it fails and so forth. You'd be a lot more proactive. So that's predictive analytics and machine learning. And one of the examples I've seen of this in use is the pest control company Rentokil Initial. So they actually did facial recognition of rats so they could take video footage on their client sites, identify when a outbreak was happening and take proactive steps to actually address it. Now this is opposed to waiting for the next quarterly visit with that. client and discovering there's an outbreak and having to take remedial action so this allows no matter what the usage it allows organization to be a lot more proactive in supporting their customers predicting failure, predicting what might happen in the future so that's predictive analytics The next biggest AI in use in the enterprise was natural language processing at 17% so this is whereby you're analyzing text from users so the most common use of this is most. Likely to be things like chat bots or interactions with customer support or with customers around facilitating their support and their customer service process. And also what I've seen quite commonly is things like contract analysis using AI or you might assess HR documents or you might do a combination of both. You might do a chat bot analyzing, responding to questions on HR policy referring to. HR documents or something? So that's natural language processing. And the third example that we see in use in the enterprise and the most common around 29/30% is generative AI. So this is the ChatGPT, the copilot, you're interacting and it's generating content for you. And although lots of people experimenting with this, the most powerful thing I've seen recently on this is from the clothing chain Mango. And they actually. Generated so they took photos of the garments first of all, and then they generated a fictional digital model. And then they created a digital fictional background for the model and superimposed the new clothing line on this model. So it's like end to end campaign for this clothing chain. So just to think about what that eliminates, that's no photographer. There's no model. There's no two days in the Bahamas recording this photo shoot. And you're, you know, you're significantly altering the entire process. Now, the fear mongers would say there's a huge amount of jobs losses from that. But I would argue that actually there's as many jobs, if not more created during that digital process than have been displaced. And it means that actually that whole supply chain and the advertising of clothes can move a lot quicker. So those are the three most common uses of AI we see. In the next video, we'll explore some of the barriers to people actually implementing and experimenting with AI in the enterprise.