Your next business strategy may be shaped by generative AI, with KPMG and other management consultancies embracing a technology that, new research shows, executives are racing to implement regardless of ongoing concerns about its accuracy and security.
Having launched a generative AI service called KPMG KymChat for internal use by its 265,000 employees in March, the firm this month expanded the service’s scope – allowing clients to access an ‘Accelerator’ solution that complements its consulting services with access to KymChat and the use cases, lessons learnt, and features on which it was trained.
The service – built using Microsoft 365 Copilot and Azure OpenAI Service tools – comes as KPMG and Microsoft announced a multi-billion dollar partnership expected to generate more than $17 billion ($US12 billion) in incremental business for KPMG.
KPMG “is embracing the future,” said KPMG International global chairman and CEO Bill Thomas, who said the deal “will help harness the power of our multidisciplinary model by ensuring that our people always have the right expertise, skills, and tools.”
“We believe that AI Is key to unlocking sustainable growth in a way that will build a better future for our people, our clients, and society…. It will also help make KPMG a more agile and resilient business that continues to be an interesting and exciting place to work.”
By training KymChat on proprietary data rather than using the massive and often inaccurate mass-market ChatGPT – a recent NordVPN survey found 40 per cent of users ‘often’ noticed mistakes in the chatbot’s output and 37 per cent ‘sometimes’ spot errors – KPMG has joined the growing ranks of companies working to make generative AI enterprise-ready.
New initiatives enable any business to build its own generative AI platform, with Databricks recently debuting Dolly – a white-labelled solution, based on the open-source EleutherAI’s Pythia family of AI models, that is said to allow any business to create a generative AI tool for $150.
“Everyone assumed that you need a giant model to get this ability, and people hadn’t tried to get it with a much less expensive model,” explained chief technologist and cofounder Matei Zaharia.
“We showed that if you add 50,000 examples of conversations and train it on those, suddenly it becomes pretty good at instruction following.”
“That shows that the secret of getting this ability isn’t necessarily a huge model, or lots of training data; it’s having the right data to get the model to do this – and this means a lot of organisations are going to want to own and build their own models.”
Excited about AI, but far from ready
While major consulting firms trumpet the possibilities of generative AI – Deloitte, for one, has built a compendium of legitimate business use cases for generative AI while Gartner identified five others – studies suggest their customers are eager but still unsure how to implement it.
Fully 74 per cent of respondents to a recent Capgemini survey, of 1000 executives in 13 countries, said the potential benefits of AI outweigh concerns about the technology – with 40 per cent having already established teams and allocated budget.
That corroborates a recent Gartner straw poll of 2544 executives that found widespread enthusiasm over ChatGPT has driven 45 per cent to increase their AI investments – with 68 per cent saying that the benefits of generative AI outweigh the risks.
Business solutions provider Avanade has also found widespread optimism about AI, with 85 per cent of organisations expecting it to increase their revenue growth by 2025 – yet just how that will be achieved is still very much up in the air.
Two-thirds of respondents admitted that their organisation lacks “sufficient checks and balances in place to mitigate the potential risks and harms of AI” – issues that recently led the Australian government to limit its use to ‘low-risk’ situations.
“The importance of adopting an AI-first mindset cannot be overstated,” said Avanade global head of advisory Jillian Moore, warning that “there is no end point to AI readiness.”
“To seize the growth opportunities of AI and mitigate risks for unintended consequences through continual change, leaders need to consider more than the technology implications of AI.”
With 48 per cent of respondents yet to implement responsible AI guidelines, new Avanade generative AI consulting services have helped companies like recruitment provider PageGroup – whose managing director for group data, insights and activation, Alex Bates, said the technology would “supercharge the vast knowledge and experience of our people.”
“It’s helpful to have a framework to help us to learn fast,” he said, “so we can prioritise actions that enable our people with AI and harness the business value of the technology in a secure and transparent way.”