Nir Evron, Senior Generative AI Lead for Microsoft’s GBB (Global Black Belt) EMEA, spends an increasing amount of time working with EMEA’s largest organisations on generative AI projects. 

Microsoft is at the centre of the generative AI storm that was whipped up earlier this year as more people gravitated towards Open AI’s ChatGPT, which was released to the public at the end of 2022. Microsoft is one of Open AI’s biggest backers, having invested $13bn (£10.3bn) in the company, and the hardware that underpins ChatGPT runs on Microsoft’s cloud. 

The sudden uptake and interest in generative AI was a complete surprise to his team, Evron says. They expected a much slower-burn growth in interest and usage. “We expected we were going to have to tell people about this technology and teach them what this technology can do. ChatGPT grew to 100 million users in two months. All of a sudden everyone knows what generative AI is.”

As the interest in generative AI increased beyond expectations, it experienced an inevitable backlash. Politicians and business leaders talked about the threat of AI – not just to jobs, but also, in some cases, human existence. To Evron, the criticism is largely unfounded.

“The concerns that people have, some of them are genuine, some of them are not. But like anything that is transformative, people need to adapt. The type of discourse that we see at the moment is statements like all accounting jobs or media jobs or what-have-you are going to go. Nothing is as black and white as that.”

Evron will be talking about the possible impacts of generative AI at the ICAEW Annual Conference in November. People shouldn’t compare AI with, say, the invention of the motor vehicle, which did away with an entire industry built around horse-based transport. It’s more subtle than that, he says.

“AI is not sentient. It’s not good enough to make decisions. People won’t trust it to make decisions, and the regulator won’t allow it to make decisions that affect people’s wellbeing. All of those things contribute to the fact that while it is a useful technology, it’s not designed to replace humans. It’s designed to supercharge them with abilities that they didn’t have before.”

As more businesses adopt the technology, it is becoming clearer how it will transform the way we work. Evron has had conversations with CEOs who have used it to increase productivity and profitability in their organisations. Healthcare professionals are considering how it could be used to sift through and consolidate patient data, saving doctors and surgeons time that they could be using on patient care. 

“We never had the tools before to do something like this at scale,” Evron explains. “AI was useful, but it required a lot of work to make it work. Now we have something that does all of these things out of the box because it’s intelligent enough. That is the transformation behind this technology.”

Previous AI models have been able to handle structured data. Generative AI can sift through unstructured data and summarise the relevant parts. This would typically require a lot of human time and effort, but tools such as ChatGPT could drastically reduce that effort, says Evron.

“Earlier technologies weren’t intelligent enough to understand text and context. Accountants need to deal with a lot of unstructured data, from tax policies and regulation to company reports and non-financial disclosures.”

We now have the ability to feed generative AI hundreds and thousands of documents, ask questions about those and get an immediate answer with a reference to the document that it came from, Evron explains. This could reduce hours worth of work down to seconds. “That was previously impossible, at scale and with the level of accuracy and competence that GPT4 is capable of. These large language models are super effective at doing that.”

This can help to reduce the amount of time people spend working on time-consuming tasks that take them away from the primary function of their jobs, says Evron. “An accountant’s job is not keying information into an Excel sheet, but they have to do it as part of their job. Can we automate that? Can we help them turn unstructured data from documents into useful data that they need in order to make decisions quicker? Can we automate manual tasks to free up valuable time? That’s where we can use generative AI.”