AI agents in finance may look more like steady improvements than total overhaul, explains Brad Joudrie
Earlier in May Anthropic announced the launch of ten AI agents built specifically for the financial services industry. The agents, which plug directly into Claude Cowork and Claude Code, are meant to be autonomous and capable of covering different forms of research and client work. That includes pitchbook creation, meeting prep, earnings revies, financial modelling, and market research. The announcement was framed as a source of ‘disruption,’ with Anthropic CEO Dario Amodei telling Reuters that if software companies fail to adapt to AI, they could “lose market value, go bankrupt.”
Brad Joudrie says his platform is already built to work with AI agents. Joudrie is the CEO of Conquest Planning, a Canadian financial planning platform used by a significant segment of Canadian advisors. He explained that these new agents are not likely to replace expert planning systems like Conquest. Instead, they will function as coworkers, capable of using Conquest to generate and update financial plans. Advisors should now be ready to use these agents’ capacity for conversational prompts and autonomous work to run systems they already rely on.
“What [Anthropic] is trying to do is essentially provide a single pane of glass around running a wealth management business. Supporting consumers, supporting financial professionals and advisors, supporting the firms ultimately with agentic workflows that allow for more conversational ways of approaching really anything that would be supported within a customer journey,” Joudrie says. “They’ve talked about these 10 new capabilities, but they’ve also talked about partnering with expert systems to support the use of these more intentional tools on top of this wealth experience so that you can still get the single pane of glass, still get the conversational approach, still get the workflow management efficiencies, but not lack any of the expertise that these expert systems are providing.”
To prepare for the rise of these agents, Joudrie says, the Conquest team has already created a Model Context Protocol (MCP), an open-source standard created by Anthropic that allows AI models to connect to external data sources, tools, and software. An advisor using an Anthropic agent can now plug that agent into conquest to run planning functions. Rather than the advisor updating each new piece of data manually to reset the plan, Joudrie sees a shifting world where the Advisor goes to Claude and Claude then uses Conquest as a “headless” platform.
An AI agent capable of running financial planning software might seem more disruptive to advisors than to a platform like Conquest at first glance. If consumers can just ask Claude to update their plans, they may not need an advisor. Joudrie argues, though, that the last person in the picture to be removed by AI will be the advisor. He echoes Peter Nolan, Head of Asset & Wealth Management at Anthropic, who said the same thing. Joudrie notes that Conquest has already been available as a direct to consumer platform, and that hasn’t devalued the role advisors play. The introduction of AI agents shouldn’t change that either.
Joudrie insists that Conquest was built to run “headless” accessed through an Application Programming Interface (API) and now through an MCP. That makes it a useful tool to work with an AI agent. Looking at the risk of AI disruption to the industry as a whole, though, Joudrie cautions against simply adding AI for its own sake.
“What financial professionals need to do is take a step back and look at the business that they’re trying to run and the strategy around execution within that business and then find the technology that’s going to support those workflow experiences in a more efficient way for them,” Joudrie says. “Instead of just being excited by all the new bells and whistles and toys, it’s being intentional with what you’re trying to solve for.”
Joudrie notes that this intentionality doesn’t mean advisors should avoid adding AI agents. Conquest has added its own agent to its Strategic Advice Manager (SAM) tool called SAM Guide. That agent doesn’t touch anything outside of conquest and Joudrie says it doesn’t hallucinate. It allows users to make requests in common language that will then deliver answers based on the data in conquest. It’s an agent that Joudrie says is designed to meet the industry’s key expectation: that output is consistent, accurate, and repeatable.
Taking this kind of intentional and industry-rooted approach to AI can be beneficial for advisors, tech firms, and other industry players in Joudrie’s view. He thinks that any part of an advisor’s workflow that they feel is constraining them could be fertile ground for an AI agent.
“The change is here. So there needs to be some sort of embrace of change. If I’m a financial professional and I have a firm, I want to think about how I’m delivering my experience, how I’m running my business and what sort of technology I can leverage to improve and support those experience changes,” Joudrie says. “I go back to however, being intentional with those decisions. There are so many choices that exist out there and with the advent of people building and leveraging more agentic experiences and even building experiences off of agents. Focus on the partnerships with the firms that you think are being intentional in solving a specific need that is a problem for your business.”
