Deloitte has extended its AI consulting capabilities with the launch of Generative AI practice built around the core of what the company is calling its Generative AI Market Incubator.

The company says the move is in response to the ability of the rapidly emerging technology to unlock host of new marketplace applications while also enabling organisations to radically re-imagine their products and business models.

In a report published in the second half of 2021 called The AI Dossier, Deloitte identified 6 common ways the firm says AI creates business value;

• Cost reduction
• Speed to execution
• Reduced complexity
• Transformed engagement
• Fueled innovation 
• and fortified trust.

But reflecting how quickly AI has exploded into public consciousness with the launch of services such as  ChatGPT and Midjourney, the report makes no reference to ChatGPT, although Deloitte has been discussing the technology in other reports since at least 2021.

According to Kellie Nuttall, Strategy and Business Design Leader at Deloitte, “The key point is we are moving really quickly around Generative AI specifically, but it's not from the standing start. We've got a practice in Australia of over 650 people who work in the AI and data space. We established our AI Institute two years ago and we've had dedicated teams and consulting across the firm focus on AI and data.”

She said the company had been building capability organically through hiring, but she also flagged the acquisition of Intellify, a data science and machine learning specialist a year ago.

Nuttall said the current AI capabilities translated well into the emerging world of generative AI.

“At the core of any of these tools is data. So you need like a practice and capability that understands how to bring good data in these AI models. What we probably see a little bit differently with generative AI, which we haven't seen as much in the past, is there's the large language models that many of the big hyperscalers will put out and invest in, like OpenAI did with ChatGPT, for example. You don't need to build those models from scratch because they've been made by all the hyperscalers for you. But what you will need to do is be able to get your data into those models and be able to tune those models to be more relevant for your particular data that you've got.”

Use cases

Nuttall told Mi-3 that some of the early use cases will be consistent across any industry. "These horizontal use cases will include things like customer feedback, sentiment classification, or personalized conversations, chat bots, et cetera, that could be in any industry."

"And then there's going to be your industry vertical specific applications. But then if you look at something like healthcare, it might be 3D images of anatomy to help patients understand."

She described generative AI as the first democratised form of the technology. "You don't have to be a data scientist to actually play with it and understand tangible benefits. Anyone can log in and have a go and actually see how good it is.:

One of the issues that is starting to emerge is company IP leaking into large language models, most famously with Samsung, where engineers in company's semiconductor division used ChatGPT to check source code.

This reinforced the need for businesses to get sound advise on AI practice, she said. "The key thing is IP and your data sets that you're using to train models, being really mindful around uploading any data from your organisation, organisations you're working with. I think anyone that has been working in the AI and data game for a long time will probably already have pretty good capability in place of security and privacy."

Incubator

Built around a team of engineers, Deloitte’s Generative AI Market will initially focus on the rapid development of Generative AI pilot programs; along with demonstrating proofs of concept, and most valuable practices

The company says an R&D team will work with the firm’s alliance partners to train and tune Foundation Models. 

A spokesperson says the teams gained experience through early adoption and experimentation with Generative AI technologies through 2022, and further scaled through Deloitte’s local acquisition of Intellify and hiring of deep AI specialists. 

“Additionally, the practice works in concert with the AI Institute to bridge the AI talent gap and train thousands of practitioners on a wide range of AI skills, including model development and prompt engineering,” according to a Deloitte spokesperson.

The rapid rise of generative AI caught many in the industry off guard. McKinsey and Company for instance completely missed it in its December 6, 2022 report on A.

The advice sector is now scrambling to catchup  in a field that is likely to spawn a whole new clutch of careers and capabilities according to industry analysts.

Constellation Research founder Ray Wang for instance identified more than half a dozen new roles in a recent tweet including, prompt engineers, algorithm inspector. human machine trainers, .AI ethics Managers, deep  fake reviewers and authenticity checkers.