In 1993, a New Yorker cartoon gave us the line: “On the Internet, nobody knows you're a dog.” Today, the more expensive question is whether the person asking to move your money is a person at all. Soon it may be an agent you have authorised to act for you.

I think fintech is underestimating the business of accepting digital identity. Issuing someone a digital ID is only the beginning. The harder job is persuading another bank to trust it, pay for it and agree what happens when it fails. That is where I expect a valuable new financial network to emerge.

The technology has an awkward name: verifiable credentials. Think of a digital ID whose origin and authenticity a service can check, without starting the entire identity check again. My interest is in what happens when that ID becomes useful beyond the organisation that issued it.

The history of card networks offers a useful lesson. Your card works because banks and merchants have agreed to accept it under shared rules. The card itself is only one part of that arrangement.

I would judge a digital ID by the same practical test: where can I use it? A beautifully designed wallet that works with one provider leaves most of the original problem intact. The valuable work is getting the next institution to join, then the next.

The W3C model below shows the basic arrangement. An organisation issues a credential, you hold it, and another service checks it. Common standards allow that exchange. They cannot make the receiving bank comfortable with the decision.

A digital ID passes from the organisation that issues it, through you, to the service checking it. Source: W3C, Figure 1 (W3C Document License).

September's U.S. guidance matters because it gives banks a clearer basis for considering these credentials. FinCEN and other banking agencies clarified how mobile driving licences and other digital credentials can be used within existing customer identity checks. Banks still have to meet their obligations and decide what they can accept.

That qualification is central to the opportunity. A regulatory clarification creates room for a market; it does not create the network. Proof's September launch of a portable credential for banks is one commercial attempt to fill that space. I will be more interested in acceptance across institutions than in the launch itself.

Europe is putting more weight behind adoption. EU digital identity wallets are due by the end of 2026, with acceptance obligations for covered banks following in December 2027. Those deadlines should get identity out of innovation presentations and into operational decisions.

For a bank, I think the sensible response is to start with one useful customer journey. Can someone bring an accepted digital ID to open an account, share fewer documents and still satisfy the bank's checks? Proving that would tell me more than another ambitious wallet announcement.

Microsoft Security's overview shows how a reusable credential works in practice.

Banks and fintechs pay to check identities. Customers pay in time and repeated disclosure of personal information. If a trusted check can be reused, there is a potential saving to share.

One possible model is to pay the organisation that issued the credential when another business relies on it. I think of that as interchange for trust. It is a commercial hypothesis, and the fee would have to earn its place. If accepting a credential costs as much as repeating the check, the argument weakens quickly.

My instinct is that distribution will matter more than many identity startups expect. Banks already have customer relationships. Card networks have experience bringing institutions together under common rules. Specialist identity providers bring expertise in checking people. The winner will need enough of those capabilities to make acceptance a routine business decision.

The quality of the original check still matters. The U.S. National Institute of Standards and Technology (NIST) diagram makes the basic work clear: collect the evidence, check it is valid and confirm it belongs to the applicant. Reusing a weak check would just spread the weakness.

Three steps in checking an identity. Source: NIST, Figure 1.

There is also a limit to the card analogy. A bank cannot assume that accepting somebody else's credential removes its own responsibilities. It will need confidence in the checks, a way to handle fraud and an agreement about who bears the loss. I expect that negotiation to be harder than connecting the software.

If I were building in this market, I would start with a group of businesses that already need to trust each other. Give them a reason to accept the same credential, agree how mistakes are handled and make the real saving visible. That seems a more credible route to growth than asking consumers to download another wallet and hoping businesses follow.

For banks, the temptation will be to treat this as another compliance project. I think that misses the opportunity to improve the customer relationship. A customer who can carry their identity elsewhere is easier for a competitor to onboard, too. That should make incumbents pay attention. Better identity could lower the cost of acquiring customers while making it harder to keep them through administrative friction. I would welcome that shift: banks should compete on the service they provide after the identity check.

We started AID:Tech in 2015, bringing digital identity and payments together. More recently, our KARE platform has put that work into practice in U.S. disaster relief, using the Algorand blockchain to connect survivors' identities with assistance.

In a 2024 pilot in Mississippi and Florida, hundreds of recipients received relief after confirming their identities. KARE uses digital identity standards from the World Wide Web Consortium (W3C) to help survivors access support across participating organisations.

KARE connects the relief organisation, the survivor and the payment. Product screens from the Algorand Foundation case study.

That work is why I keep coming back to acceptance. A digital identity has limited value if the next organisation asks you to start again. For someone recovering from a disaster, the delay can stand between them and help. The useful network is the one that lets trust travel with the person.

I see the same problem in finance. Opening a second account can mean handing over the same documents and proving the same facts again. Each institution has its own obligations, but I struggle to believe that repeating so much of the work is the best model we can build.

I suspect agents will push this market forward faster than consumer wallets alone. People put up with repeated identity checks. An agent expected to act across several services runs into a different problem: each service needs to know who authorised it and what it is allowed to do.

Imagine Aoife in Dublin asks an agent to find a cheaper energy supplier and arrange payment, with a limit of €200 a month. The supplier needs to know that Aoife is a real customer and that she gave this agent permission. Her bank needs to know that the payment falls within that permission. Neither should have to trust the agent simply because it can hold a convincing conversation.

This is the experience I would want: Aoife approves the task, can see the limits and can cancel the permission. The agent cannot take out credit or add unrelated payments. If her phone is lost, there is a recovery process that protects her without leaving her locked out indefinitely.

That is an illustration of what participating providers could build. The difficult part is making those permissions mean the same thing to the customer, the supplier and the bank.

The European Commission illustrates how a wallet could help someone choose and share documents for a bank loan. Agent permissions would add a further step. Source: European Commission.

The work by Visa, Mastercard and Ant International on identifying agents across networks is worth watching for that reason. Payment businesses have a direct interest in knowing whether software is entitled to spend. Giving an agent a payment method only solves part of the problem.

We should also be careful about what “verified” means. Proving that a unique human exists behind an account, as World ID seeks to do, is different from establishing that person's legal identity for a bank. Neither, on its own, proves that they authorised this particular payment. Those distinctions will matter far more when software starts spending money for us.

Alex Blania discusses proof of humanity and AI with Empire. It is a useful part of the debate, alongside the separate questions of legal identity and permission.

I would put more weight on one credential being accepted by several independent institutions than on another wallet launch. I would want to see lower costs, fewer repeated checks and a recovery process that works when a customer loses their phone. Those are signs of a useful network.

Privacy is part of that test. A system that makes identification easier could also make people easier to track. Sharing only the information a service needs helps, but so do limits on what businesses can request and retain. People also need a workable route to essential services if they cannot use a smartphone.

I would be wary of a model in which one provider becomes the gatekeeper for every financial interaction. The convenience of a reusable identity should give people more freedom to move between services. If it creates another dependency they cannot leave, we have traded one problem for another.

For founders, the commercial question is where they can win acceptance. For banks, it is which checks they can safely reuse and which responsibilities remain theirs. Those are the conversations I would prioritise now.

Our work on KARE has made me optimistic about what digital identity can do when it is tied to a real need. It has also made me impatient with treating the wallet as the finished product. The next organisation has to accept it. Until then, the network is a promise.

Nobody knew you were a dog. Now the challenge is knowing who stands behind an agent and whether it has the authority to act. My bet is that the businesses which make those answers dependable will build the next valuable financial network.

Subscribe to Synaptic Finance for weekly analysis on the intersection of AI, blockchain, and the future of finance. Connect with me on LinkedIn for more on what I’m building.

Written with AI assistance. Views are my own.