Connecticut Wealth Management, an independent registered investment advisor with $5.2 billion in assets under management, 45 advisors, and serving more than 1,300 client families, has hired Subatomic in a five-year partnership to build an AI Operating System that will unify the firm’s data and deploy AI Co-Workers supervised by its human staff.
The first phase is scheduled to go live before the end of 2026.
The system will allow CTWM to maintain its 35-to-1 client-to-advisor ratio while adding capacity for the firm’s team, according to Kevin C. Leahy, founder, president and CEO of Connecticut Wealth Management.
The AI Operating System sits across CTWM’s custodian, portfolio management system, CRM, planning and tax software, meeting notes, phone system, email and documents, creating a unified experience, according to the firms.
An AI Chief of Staff assigns tasks to AI Co-Workers, with an initial focus on work around client meetings, including preparation, follow-up emails and calls that generate tasks, the low-hanging fruit that many of the AI notetakers serving the advisor market are already tackling.
“An important distinction is that we are building this in the client’s [CTWM’s] infrastructure, helping them future-proof it and be able to continue using their own data, which they continue to own,” said Sam Sova, co-founder and CEO of Subatomic, who added, “They had the foresight to hire AI when everyone else in wealth management seems to be buying AI.”
“This is an alternate labor force that will rapidly become a natural extension of the human teammates you have,” he said.
At CTWM, the advisor model is built around what the firm calls a diamond team structure that brings together a managing advisor, financial advisor, analyst and client service professional.
Sova said that the Subatomic operating system runs within CTWM’s security environment with role-based permissions, and Subatomic’s Deep Lens product records every action AI Co-Workers take, allowing the firm’s staff or auditors to review it. It also tracks a firm’s AI usage from a token and cost perspective.
“The majority of RIA firms we meet with have a Frankenstack, 10 or 15 different applications or platforms in their stack, but too often they are only using say 30% of its potential and the human is still the middleman in the mix trying to make that whole system work,” said Sova, noting that seldom is there a single login across everything a firm is using, meaning that many advisors today still have to log in to multiple systems for various tasks.
“It’s the big unlock now,” said fellow Subatomic co-founder and CTO Karl Simon.
“Maybe they [an RIA firm] have gone to the unified data model and have a unified data layer, but they are coming to realize with native AI you can finally break out of the traditional construct of the world of tools that were out there,” he said, adding that while newly minted breakaway firms might have a clean slate to design their own technology stack from scratch, most established firms will already have a diverse stack and one in which much of what they have is thought of as best of breed.
“Such firms see what we have as a way to simplify their architecture,” said Simon.
Sova said another important aspect of Subatomic’s approach is that it is model-agnostic, meaning its platform is not based on any particular model.
“It allows them flexibility not only for now but in the future, to pick from whatever model will be most helpful, efficient or cost-effective,” he said, using as a hypothetical example that if some earlier and less expensive version of Anthropic’s Claude remains fine for building certain tasks or purposes, the firm’s AI can use that.
Sova said that it remains too complex to provide transparent pricing because the several firms Subatomic had worked with thus far (CTWM is not the first, just the first to publicly acknowledge the scope of work being done) were each different in what they needed.
“There are other firms that want to embed Subatomic’s workforce,” said Simon, “they are taking it in maybe smaller bites at a time, say with one internal team at a time.”
Numerous firms have announced what are being called agentic platforms for wealth management firms, sometimes referred to as agentic wealth operating systems.
These platforms include those from Advisor360°, Altruist, Anthropic, Apex Fintech Solutions, Datalign Advisory, Humanity Labs, Jump, Onevest, Zeplyn and Zocks, among others.
And while the WMIQ 2026 WealthStack Study indicated that only 16% of firms were actively running at least one production AI agent, it is clear that many firms are considering next steps toward adopting AI.
“With a mission to turn life’s aspirations into reality for our clients and team, we must continually evolve,” Leahy said.
