AI’s opposable thumbs
Dolphins and whales are smart, maybe even smarter than us. But without opposable thumbs, they can’t express their intelligence in technology or in ways that endure over generations.
Similarly, you’d have to be living under a rock not to have heard how smart the new generation of Generative AI tools are. Examples include ChatGPT, Bard, Claude and Ernie Bot as well as a range of emerging niche solutions from the open-source community. But for those smarts to be truly useful they also need opposable thumbs, or at least a digital version of them. It’s one thing to talk to a smart machine, but until they can interact on our behalf with applications, and the real world, their utility is limited. But this is changing, particularly in the enterprise.
At work, interacting means having access to business systems from customer relationship and products through to internal communications and financials. To turn this access into digital opposable thumbs, businesses need solutions embedded in the public AI services, AI incorporated into business processes and AI in the hands of the IT team managing the solutions.
When internet search engines matured and became truly useful around the turn of the century (many will remember AltaVista), the clean, minimal, search interface of these tools became the starting point for many work tasks on the Internet. In a similar vein, AI-based tasks have found their starting point in the minimalistic chat interfaces popularised by emerging platforms like ChatGPT. It seems that use of the Internet and AI will converge through plugins, which are the emerging equivalent of mobile apps popularised by Apple’s iOS and adopted by Android.
ChatGPT has done for AI what the iPhone did for smartphones, so it is no surprise that ChatGPT’s plugin store is having the same effect. Launched in March 2023, at the time of writing there are more than 500 plugins available, and more are being added daily. We can expect the enterprise software vendors to soon start adding plugins directly for enterprise users to access their solutions as well as toolkits to make it easy for business to build services for their customers to access their services from the convenience of the AI chat screen.
But the real revolution for business will come when the power of Generative AI is embedded in the heart of their processes. Much has been made of the ability of AI to mimic human customer service operators, but there is even more opportunity to reduce the need for humans to intervene at all. We’ve long seen great benefits from automating business processes from provisioning to invoicing. But with every automation has come new complexity and inevitably human intervention has been required to deal with the huge number of exceptions that are thrown-up. Whether it is provisioning a new utility service or opening a bank account, the model customer goes straight through but for too many real-world situations there is something unanticipated or out of the ordinary. This is where Generative AI excels and will enable major productivity enhancements.
Finally, the configuration of our enterprise solutions is tremendously complicated, keeping technology professionals busy with collecting requirements, configuring systems, and maintaining the solutions. Generative AI, when embedded in the tools they use for these activities, has the potential to revolutionise the way they work. While developer tools are emerging, the real change needs the enterprise software vendors to support Generative AI-driven configuration and ongoing maintenance. In the near future, even the stability of the operating environment will be maintained by the new breed of Generative AI models.
The major software vendors have almost all announced generative AI plans which touch on all three of these approaches. As these get released it will force organisations to think again about their enterprise data and solution architectures.
Perhaps this is why so many, including myself, challenge the idea that generative AI will cause jobs to be lost. There are huge productivity and service opportunities, but there is also a lot of work to be done. Each past wave of technology has streamlined the workplace at that time and created demand for technology to deliver so much more.
Chief Information Officers need to put themselves at the centre of their organisation’s AI response, assessing both the short-term priorities and business plans that leverage the long-term opportunities. More than anything, businesses will see tasks, and even jobs, that the technology is capable of taking on but will need help to find ways to integrate generative AI with the data that it needs and the real world where the benefits are realised. These will be the emerging digital opposable thumbs of AI at work.