Train up ChatGPT like a staff member: Morgan Stanley



Firms introducing artificial intelligence (AI) to their businesses should treat it like a new member of staff to train up.
Speaking at the Stockbrokers and Investment Advisers Association (SIAA) conference in Sydney, Jeff McMillan, head of analytics, data and innovation at Morgan Stanley Wealth Management, explained how the firm had implemented AI to the business.
Morgan Stanley had implemented a Next Best Action tool that acted as a platform for personalised communication and client engagement.
He specifically touched on how AI platforms learnt and became more specific through user interaction.
“There’s a perception that technology is a virus which will have unintended consequences. It is not a virus,” McMillan said.
“It is no different to hiring a junior member of staff who you have to train up, that can take years. The technology will only operate within the confines of what you let it do and operate with the values and ethics of your business.
“If someone said ‘AI made me do it’, then they didn’t have the proper safeguards and risk controls in place and shame on them.”
He said 90 per cent of Morgan Stanley’s advisers were using the firm’s AI tool on a monthly basis and it was helping them “in every metric” of their business including client retention, engagement, and revenue.
Morgan Stanley had met with OpenAI, creator of ChatGPT, some 18 months ago and decided to use the system with its advisers to help them provide advice more effectively and efficiently.
“I was very jaded, I don’t usually walk away from these meetings excited but the responses it was giving was better than our own boss’ responses. So we looked at how we could engage with it as we realised the world would never be the same,” McMillan said.
“It isn’t about machine versus human, it’s about how do we help people do their jobs more effectively?”
The next stage for take-up of AI would be which companies developed new tools fastest such as Salesforce, Bloomberg, and Refinitiv and how they would use data to do so. JPMorgan Chase had already announced it was developing its own IndexGPT product to deliver investment advice.
In a message for Australian wealth managers, he said: “I would be paying close attention to existing providers and asking if their strategy would benefit your clients. Many of them won’t be successful.
“Will the incumbents be able to move quickly and exploit their data or will more nimble start-ups be building stuff from scratch?”
Recommended for you
Money Management examines the share price of financial advice licensees over one year to 31 March, with M&A actions in the final quarter having a positive effect for two licensees.
A $3.5 million settlement for victims of Melissa Caddick has been approved by the Federal Court following an initial agreement last December.
The Reserve Bank of Australia has delivered its first rate decision since the introduction of a new board structure last month.
Digital advice provider Otivo has launched an interactive tool, powered by artificial intelligence and Otivo’s own advice engine, to help answer client questions.