AI took a turn recently with the launch of generative AI, which goes beyond simply interpreting existing data to create new work. OpenAI’s ChatGPT, a large language model trained using vast swathes of text on the internet, is the best-known example.DADO RUVIC/Reuters
Sign up for the Globe Advisor weekly newsletter for professional financial advisors on our newsletter sign-up page. Get exclusive investment industry news and insights, the week’s top headlines, and what you and your clients need to know.
Artificial intelligence (AI) has leaped from niche technology into the mainstream in the past few months. Research suggests that financial advisors are interested in how it can help their practices.
Accenture PLC surveyed 500 advisors across North America last year and found almost nine in ten were eager to adopt AI tools in their daily work to automate time-consuming and manual tasks.
Many AI solutions help with generic business tasks across many sectors. Examples range from transcribing audio with Otter.ai to knowledge management with Mem Labs Inc. AI that will enhance your text with Wordtune or even write it for you with CopyAI Inc.
Jacqueline O’Flanagan, financial services industry lead at Microsoft Canada Inc. in Toronto, says she’s already seeing advisors use AI to claw back precious time by handling mundane jobs.
“These include data entry and extracting of insights from existing data sets, pitch book generation, investor report summarization, and compliance checks,” she says.
AI can also be your invisible meeting assistant. Fathom records video and audio calls and uses AI to transcribe the audio and generate a bullet-point summary of what was said.
Advisors don’t have time to review an entire transcription or audio file, says Fathom chief executive officer Richard White in San Francisco. Instead, advisors will read the summary and click through any highlighted points to review the associated text or audio.
“They will have better recall and be so much more prepared for that next call,” he says. “They can pick up where they left off [with a client].”
Fathom also drops this information into the advisor’s customer relationship management system of choice.
Mind your language
AI took a turn recently with the launch of generative AI, which goes beyond simply interpreting existing data to create new work. OpenAI’s ChatGPT, a large language model (LLM) trained using vast swathes of text on the internet, is the best-known example.
Microsoft Corp., which invested US$10-billion in OpenAI in January, has folded the latest version of the technology into its tools and services. In March, it announced Microsoft 365 Copilot, which is an assistant for its productivity suite that completes complex tasks based on natural language prompts. An advisor could ask it to create a presentation using the notes from a meeting or draft a proposal letter for a prospective client.
Some LLM output can seem almost magical.
“LLMs have huge potential in productivity gains, knowledge mining, and intelligent automation areas,” says Chris Mar, partner and national transformation leader at PricewaterhouseCoopers LLP Canada. However, he also highlights some dangers.
LLMs use probability to guess at what output they should produce, but don’t understand what they’re producing, Mr. Mar says. That can lead to inaccuracies.
For example, a New York court recently found research for a legal brief created using ChatGPT riddled with false case references, invoking disciplinary measures for the lawyer who created it.
“One should generally avoid over-relying on the output as is,” Mr. Mar adds. There still needs to be a human in the loop.
Specialist AI for advisors
LLMs can become more accurate when trained on smaller amounts of specialized text such as internal client documents or tax research, says Frank Attaie, general manager of technology at IBM Canada Ltd. The company uses LLMs and other kinds of AI in its Watson brand of AI products.
IBM worked with independent mutual fund dealer Investia Financial Services Inc. last year on a proof of concept system designed to help advisors cope with growing administrative workloads.
The system held natural online conversations with clients to automate changing address information. The system would ask why the client was changing their address, using AI-driven language processing to understand the responses and offer other services automatically.
For example, the AI might ask if a divorcee moving out of their house wanted to review the beneficiary on their life insurance policy. It might offer financial management tips for the recently divorced and offer to set up a meeting with their advisor, says Louis DeConinck, Investia’s president.
He sees AI as a way to lighten advisors’ load rather than replace them.
“If we do AI right, it can help you do meaningful work where it’s important,” Mr. DeConinck says.
Replacing mundane form-filling exercises with conversational interfaces leaves time for advisors to expand their books of business, perhaps targeting the next generation of younger clients who will become profitable as their wealth grows.
Krish Banerjee, managing director and applied intelligence lead at Accenture PLC’s Canadian operations in Toronto, says advisors will adopt solutions specific to the sector gradually.
“There will be an increased need for these vertical capabilities,” he predicts. “They will impact how you’re being more relevant to your clients rather than just being more productive and efficient.”
These advisor-specific AI tools are still in the early stages, but experts believe they will offer more guidance in areas ranging from portfolio rebalancing to due diligence and investment research.
However, they all agree AI-powered tools will always be assistants with a human hand remaining firmly in control.
For more from Globe Advisor, visit our homepage.