Charlie: The Agent That Works for the Consumer — and the Bank
As an executive who has spent decades living at the intersection of consumer data and financial services — first as a strategist, and most recently as Chief Data Officer of Visa - when Bob Hedges asks “Who does the agent work for?”, you can take the question seriously.
His article in Open Banker is a sobering read. It talks about the existential moment that banks find themselves in. It’s worth reading in its entirety, but here’s the payload:
“The companies developing agentic AI tools to address consumer financial questions are not doing so out of altruism. They are assembling the most comprehensive consumer data infrastructure ever, and they intend to monetize it. The agent … is not working for you or your customer. The agent is working for the platform that owns the interface, accumulates the data, and builds customer profiles for future marketing purposes.”
Of course, the ‘companies’ Bob refers to are the major foundational AI vendors. And the threat posed to financial institutions is the same threat across many industries; complete disintermediation and the dominance of a few platforms that know the global population more intimately than we know ourselves.
Inrupt sits in a rare position, at the nexus of personal data trends, platform economics and the web’s evolution. We’ve seen this threat coming for a long time. In fact, I wrote about it a year ago in my piece, “OpenAI’s Memory Trap.”
So, if you’re staring at the imminent threat of total disruption in your business, what can you do about it?
Well for banks, Bob’s prescription is clear: “The bank that positions itself as the trusted steward of the consumer's data and financial identity — not the monetizer of it — will occupy a strategic position that other competitors cannot challenge.”
We agree. Better yet, we’ve been building the technology that makes it possible for organizations to fulfill that role of trusted steward in the digital lives of their customers, whilst building a new and vibrant business. We’ve named this new tech “Charlie.” Charlie is the AI agent that works for you and your customers, not the AI platform.
What Charlie Actually Does
Here’s the roadmap that Bob lays out for banking executives:
1. Shield consumers’ identities before their data reaches AI platforms.
2. Give consumers genuine control and consent over what gets shared.
3. Delete and de-identify PII before it travels anywhere.
4. Give consumers portable memory across platforms.
5. Assemble intelligent context that works for the consumer — not the platform.
That is a precise description of Charlie.
Charlie is Inrupt’s personal AI intermediary, and it can be tuned specifically for banking customers. Built on Sir Tim Berners-Lee’s Solid open standard and delivered through Inrupt’s Enterprise Solid Server, Charlie puts a consumer’s data in a vault that the bank can help them manage.
When a consumer asks an LLM a financial question, Charlie pulls only what’s relevant from that vault, strips PII, replaces other personal details with anonymized equivalents, and presents the user a simple consent before anything leaves. The AI platform gets enough context to give a useful answer, but nothing it can use to profile or track the person.
Charlie learns as much about the consumer as any platform AI. But it’s all kept in the person’s memory bank, in their vault. A consumer-controlled record of their personal data and AI interactions that travels with them across LLMs.
When managed by the bank, Charlie can draw from the bank’s own data — spending patterns, income history, account balances — so he can surface genuinely intelligent context that no general-purpose AI can touch. The bank holds this data. Charlie lets the consumer wield it.
Return to Bob’s question: Who does this agent, Charlie, work for? Both the bank and the consumer.
Consumers get personal AI guidance without being taken advantage of. Banks get a new high-engagement consumer channel, plus the right to request access to rich, current customer data when it helps them provide better service. And the LLMs? They get no personal data. Just enough to learn and get smarter, without being creepily intimate.
The Window Is Real. But It’s Not Staying Open.
Bob is not just an ally of consumer protection and a smart bank strategist, he’s also become a close advisor of Inrupt.
Over the past months, Bob has worked with us to pressure-test Charlie against the critical challenges banks face with LLMs . His experience make him exactly the right advisor to help banks understand what’s actually at stake — and what to do about it.
And the timing could not be more urgent.
In May, OpenAI, in partnership with the data aggregator Plaid, announced the launch of a new personal finance experience in ChatGPT. With this offering, banks no longer face an implicit risk of disintermediation on the horizon. It’s here. Day by day undermining your consumer relationships and leaking your data to an immensely capable competitor. When a consumer connects their financial accounts inside ChatGPT, the interaction data belongs to OpenAI. The bank becomes a ledger, not a relationship partner.
Charlie gives banks a different path. The trusted layer between their customers and the AI ecosystem. A genuine steward — bank managed, technically capable, consumer-controlled, and built on a standard designed from the start to put individuals in charge of their own data.
Sir Tim Berners-Lee proposed this vision nearly a decade ago. The banking industry now has every reason to make it part of their future.
If you’re a banking executive thinking seriously about your agentic AI strategy, we should talk.
