Pontera floats itself as AI's trust layer for 401(k) access

Pontera has built its business on getting software into 401(k) accounts on behalf of financial advisors, a position recordkeepers including Fidelity have resisted, and it now wants a second job—standing between the machines that do the reaching. In a Sept. 24 blog post titled 'Agentic AI reframes the debate over financial access,' Pontera senior director of public policy Ben White wrote that a trust layer becomes mission-critical once AI agents begin transacting with other AI agents, and that his New York City firm can serve two roles, one as a trust layer for consumers and one as a protection layer for institutions.

White frames the post as takeaways from the 2026 FDATA summit in Toronto earlier in September, and the gap he points to is one the industry has not closed: much of the financial ecosystem is not yet built for agent-to-agent communication, many firms lack the means to distinguish a person from an authorized agent, and not every financial institution has the infrastructure to govern automated activity on its own. His answer is the trusted intermediary—a party that validates users, scopes permissions, minimizes unnecessary data sharing, and maintains a record of what the machines did—which he ties to the industry's pursuit of hyper-automation and the services that would ride on it.

No financial institution appears in the post, and the coverage does not say whether any recordkeeper has agreed to such an arrangement. RIABiz, which reported on the blog, placed it against Pontera's campaign to win 401(k) access from Fidelity, a standoff the outlet described in part as a cybersecurity question because recordkeepers dislike clients handing their passwords to outside firms that capitalize on them, and RIABiz reported that the threat of AI-agent overload may give Pontera a leg up in that campaign, its headline describing the pitch as possibly solving the impasse. Our own reporting in early October found the credential-sharing dispute unresolved.

Into that standoff came Schwab, which RIABiz reported has announced it is prepared to release Charley AI agents to serve its millions of clients. The coverage does not describe Charley's capabilities, its timing, or whether it faces clients directly, and that last detail is the one that decides whether advisors supervise the tool or simply inherit it. The direction is at least set: as this publication reported in September, Schwab's Claude integration costs an advisor $240 a year per seat, which prices AI as a custody line item rather than a back-office experiment.

For a practice, the practical content of Pontera's pitch is narrower than the framing, because advisors who manage held-away plan assets already depend on credentials a client hands over, and White's list—validate the user, scope the permission, limit the data, keep the record—describes what a careful advisor does with that login now. The difference is tempo: agents act continuously rather than when a human logs in, which makes a standing validator a more plausible product than a one-time consent screen, and that puts the vetting burden on whoever authorizes the agent.

That burden is the same one running through the retirement business, where, as this publication has argued, the economics have been moving toward getting paid inside the plan rather than at the rollover desk—the ground Pontera's software occupies, since the assets it manages stay in the 401(k) instead of moving to an IRA. If a validating intermediary becomes the gate through which advisors reach those accounts, terms for a service an advisor has already sold to a client would be set by a party the client never meets.

Compliance lands there too, and the split this publication has tracked, between practices gaining leverage from AI and practices losing it, applies here, with the burden falling on whoever clicks approve. Supervisory procedures written for humans with logins do not fit a machine that acts between two institutions at machine speed, which is roughly the gap Pontera is selling into: validate the user, scope the agent, keep the log.

The counterparty problem is not hypothetical: Fidelity is the recordkeeper in the credential dispute and, as this publication reported in early October, the custodian that told RIAs with less than $100 million in assets to move by June 2027, with no fee option to stay. Michael Kitces called that hard no bizarre, noting that past minimum increases let small firms pay to remain. Whether custody terms and plan-data access get negotiated in the same room is not something the coverage establishes, and advisors deal with the same counterparty on both.

What would move the discussion from positioning to practice is a recordkeeper putting its name to a trust-layer arrangement, but the post does not name one, RIABiz's hedge on Fidelity is not a deal, and the two firms making the loudest claims about agents operating inside client accounts have not published the mechanics. In the meantime, an advisor who gets asked to authorize an agent can start with questions the post does not answer: which institution holds the credential, what the agent is permitted to touch, and where the record of its activity lives.