Insight of the Week: ChatGPT cools but real money is in Applied AI
It's looking increasingly clear that the popularity of the ChatGPT website is waning, having peaked at around 1.8B visits per month in May.
Some have jumped on this as an indicator that we're heading into the trough of disillusionment that I talked about in last week's email: The Hitchhiker's Guide to the AI Hype Cycle.?
But there could be many other reasons: the launch of their mobile app (number 5 in the productivity app chart on the Apple App Store), growth in usage of their APIs, and potentially a dip in usage by students who've been on holidays. Not to mention the growing traction of other AI service providers like Google, Meta, and of course Microsoft that uses Open AI tech anyway.
But this is irrelevant to us. These companies are in land-grab mode. They're not trying to make money with AI yet.
We are. We've got businesses to run. Targets to meet. Markets to defend.
So who's actually making money with AI?
A recent interview with Nat Friedman, an angel investor who until last year was CEO of GitHub, gave some interesting insight here.
He referred to a bunch of early-stage companies that are making tens of millions of dollars using AI to deliver personalized videos, complete bids for government contracts, and optimize search engine results and product titles and descriptions. These companies are applying AI to achieve valuable results, that their clients are willing to pay for.
Nat points out something interesting about these examples: they're not competing with existing software products but with parts of companies: the video production and editing department. The government bids team. The SEO team. As I've suggested before, Generative AI is like electricity for knowledge work - and these companies are putting that electricity to good use.
A lot of work, he notes, is text processing. Creating bids. Reviewing bids. Drafting contracts. Reviewing contracts. Writing clinical notes. Reviewing test results.
Nat continues: "where previously you would've hired an agency or built a team and now to a greater and greater extent you can maybe use [an AI] model."
So what's this got to do with Customer Service and Customer Experience? Well that's a whole lot of text right there! From the conversations your agents have with customers, to the documents that explain your processes and the manuals that explain how to use and troubleshoot your products. Of course, there's a lot of speech too, but that can be easily transcribed into text, and vice versa.
领英推荐
How might you apply AI to some of those challenges?
?
We're already working with customers on all of those use cases, with some very positive results. Applied AI is already busy at work in the contact center.
Later in the interview, Nat turns to the importance of data. Apparently Google is spending a billion dollars this year on generating data to train their AI models. If AI is electricity for knowledge work, data is the wind that powers the electricity generation.
And that's where you have a unique advantage. You have a huge source of data that's more relevant to you and your business than a billion dollars’ worth of Google's data: conversations with customers. Every single day your agents are having thousands of conversations with customers. For some of you, tens of thousands. For some companies we work with, they have over 1m conversations every day with their customers.
That’s just the contact center. What about the website? Right now your customers are just browsing, but what if they could chat to your AI? Then you’d really know what they wanted. And you’d be able to serve them there and then, before they go through to your contact center, or jump to a competitor’s website. We’re working with clients who have millions of hits per month on their website. What if every one of those hits was a conversation?
Just imagine the power of the AI models you could build with that data.?
So yes, traffic to the ChatGPT consumer interface may be falling, but valuable use-cases for AI are being exploited.
If you haven't started your Applied AI journey yet, check-out our Applied AI Playbook. In short: AI enable initial customer touchpoints, let the data that flows guide you towards even more valuable AI use cases, then manage your AI workforce.
The alternative? End up in a museum alongside steam-engines and fossil fuel powered cars. Electricity replaced those. Now electricity for knowledge work threatens to replace your business, your department, and maybe even you, too. Unless you learn to leverage it, and manage it.
Kerry Robinson?is an Oxford physicist with a Master's in Artificial Intelligence. Kerry is a technologist, scientist, and lover of data with over 20 years of experience in conversational AI. He combines business, customer experience, and technical expertise to deliver IVR, voice, and chatbot strategy and keep Waterfield Tech buzzing.