Insignia Financial-owned licensee Shadforth Financial Group is incorporating new AI-enabled tools developed by the MLC Expand platform – also owned by Insignia – into its advice process in a move that is estimated to save advisers about 15 minutes on every task it is applied to.
In June this year, Shadforth chief executive Terry Dillon told the Professional Planner Licensee Summit in the NSW Blue Mountains that the licensee had embarked on ambitious three-year plan to double the number of clients it serves from its current 10,000.
Dillon acknowledged that working with salaried advisers made it simpler in some respects to get all of them working on a common tech stack, which in turn smooths the way for implementation of AI-backed processes.
Dillon tells Professional Planner that the licensee’s “big need” is for additional capacity, administrative efficiency and process simplification to allow its advisers to spend more time facing clients.
“The first two areas that Expand and [Expand chief executive Liz McCarthy’s] team have built for us is the ability for the AI… to read client service agreements and the data from SOAs to, one, populate this year’s fees for our annual fee renewals; and secondly, to read the data out of the SOA to pre-populate the application forms for clients,” Dillon says.
The average time saving produced by the AI automation has been estimated at 15 to 20 minutes per client.
“We have 17,000 reviews a year. This is a big, big change, and we do 10,000 client service agreements a year,” Dillon says.
“This saves us incremental amounts of time and adds to our efficiency in a world where we’re sort of crying out for a Big Bang change in efficiency, but it’s actually more practical to just chip away and make yourself more efficient with tools like this.”
Insignia estimates that if all advisers using the Expand platform enjoyed the same benefits as by Shadforth has to date, the aggregate time saving would be around 500 hours a month.
Support staff to reap the benefits
Dillon says that of the 270 or so employees of Shadforth, around 100 are advisers and the rest are support staff, who are expected to be the key beneficiaries of the first round of AI integration.
Those staff “serve 10,000 families, doing all of the really important administrative and client service stuff; this will save them time,” he says.
“There will be other AI transformations in the future that will save advisers more time.”
MLC Expand’s McCarthy says improved efficiency not only frees up an adviser’s time but also produces financial benefits for the advice practice.
“What I’ve noticed is, at any one time an advice practice, or a practice like Shadforth or even Rhombus, the network… can have millions and millions of dollars of revenue outstanding,” she says.
“This how they get paid. So, it’s both efficiency, but it’s also making sure that they’re getting their revenue in a timely fashion. It’s actually creating a lot of upside in a number of places. And what I’ve noticed is that it’s not necessarily reduction of cost in the back-office, what I’ve noticed more so is practices doing more with what they’ve currently got – servicing more people, doing higher-order tasks, doing tasks that add better value than what you would have been doing by re-keying information into forms.”
The AI innovations rolled out by Expand are built on a large language model, which McCarthy says tends to work best when the information fed into it is all in a consistent format. The variety and multiplicity of forms across the advice profession have posed a challenge.
“That doesn’t mean to say we can’t build capability around different types of forms, but the proliferation of forms is a challenge for AI and large language models, that is true,” she says.
“We haven’t seen a lot of those issues to date, because most of most of the people we work with are co-creating with us and narrowing down the number of forms they’ve got anyway, because it makes sense. And when you work with large organisations like the Rhombus network or Wealth Today or Count, everyone’s starting to realise the benefit of doing things in a similar fashion.
“That’s not say that you can’t customise what goes out to the end client. It’s more about how we’re sourcing it and what we’re putting it into in the system.”
Sound privacy protection
Dillon says he is not overly concerned about risks posed by allowing AI systems to handle private client data and being involved in the creation of critical documentation. He says there are already risks in the human-based manual processes that AI will eventually replace.
“When people are manually keying 10,000 client service agreements a year, there are errors, there’s human errors that just are happening,” he says.
“[AI] is actually safer in terms of making mistakes that cost people money. Secondly, the way we have dealt with privacy issues and other things is… rather than the AI reading the data straight out of the SOA, it’s actually taking the SOA data through Xplan, which protects us from the privacy issues that we were concerned about. So the question was, am I happy that AI is reading SOA data out of Xplan? I’m comfortable with that.”
Dillon says Shadforth is also rolling out an AI-powered file note tool “that will not just be a transcript of the meeting but will [also] turn the transcript into a compliant file note, but also then become a paraplanning request”.
“And in stage two that paraplanning request will be templated up with obvious advice, outcomes given. We’re a fair way away from that, but that is coming down the track. We won’t ever let – I don’t think – AI pick investments and get out of its lane.
“At the moment, we have a high-net-worth adviser with a complex client, we send two people in: one of them is effectively scribing and the other one’s being the adviser. The tool we’re launching will be a great transcript of the meeting. It’ll make sure we don’t miss anything. It’ll coach us, eventually, on how to be a better adviser.
“That’s not what we have now, but that’s where we’re headed.”
McCarthy says the next project for Expand will be to streamline and simplify the production of ROAs.
“That’s where a lot of the work is, that repeatable year-in, year-out, review, using the same large language model to pre-populate a change in investment structure or the weighting of a plan,” she says.
AI brings exciting times for financial planning businesses. However it appears their will be a tiered approach yet again. Where the smaller boutique AFSL will be able to adabt and use AI where appropriate and the bigger AFSL groups like Rhombus will be left behind while they decide if they will allow their advisers to use AI or what restrictions they will put around it.