Kodak didn’t die from the camera

In 1975, a young Eastman Kodak engineer named Steven Sasson assembled the first digital camera. It was the size of a toaster, recorded to cassette tape, and took twenty-three seconds to capture a single black and white image.

Kodak owned the future of photography before anyone else had glimpsed it. Then it spent the next three decades very deliberately not building that future, because the future did not sell film, and film was where the money lived.

This is the part of the Kodak story the idiom always misses.

Kodak did not die because it lacked the technology. Kodak invented the technology. It died because it could not bring itself to cannibalise the business model that paid its bills, and it kept telling itself a more comfortable story right up until the story ran out. The camera was never the threat. The margin was.

I mention this because the advice profession spent a recent afternoon at the Professional Planner Licensee Summit debating whether it faces a Kodak moment.

In the course of reassuring itself that it does not, three of its more thoughtful leaders demonstrated precisely the reflex that killed Kodak. Not complacency. Something subtler and more durable.

The instinct, when confronted with a structural problem, to relabel it until it reads as a manageable one. Watch the move three times and it stops looking like coincidence.

The first relabel: evolution, not disruption

Geoff Lloyd, the Conexus Financial chair and a former executive at both BT and MLC, was not convinced the profession should worry about a Kodak moment at all. “Not sure this is quite a Kodak moment,” he offered.

“This is an evolution of the way in which we can build and deliver services and products.” He liked, he said, the AI plus human element. He was less worried about full disruption by an AI agent because advice is more than generating a recommendation; someone still has to execute it.

On execution, he is right, and it is the strongest point made all afternoon. A large language model can produce a defensible strategy for a household in seconds. It cannot open the account, move the money, manage the behaviour through a drawdown, or sit across the table when the plan meets a redundancy. That gap is real and it is the profession’s moat.

But notice what the relabel does. By recasting disruption as evolution, the framing quietly assumes the moat sits where the incumbent would like it to sit.

Lloyd reached for the United States robo advice experience as proof, arguing those models came and went because they were banks pushing one product, and that this was always the deficiency.

The history is almost the inverse. The flagship American robo platforms, Betterment and Wealthfront, were independent venture-backed startups, not banks, and they did not come and go; they are still here.

The largest pools of robo-managed money did not end up with the disruptors at all. They accreted to the incumbents who copied the model, Charles Schwab and Vanguard, who already owned the client and simply bolted automation onto the relationship.

That is not a story about a fad that failed. It is a story about who captured the value once the technology commoditised one layer of the stack.

Which is the actual Kodak parallel, and it is not reassuring. The risk to advice was never that an AI agent replaces the adviser wholesale. It is that AI collapses the cost of the strategy layer to near zero, and the margin migrates to whoever owns the client relationship and the execution rails.

Andrew Gregory, UniSuper’s chief advice officer, saw this clearly when he noted members now arrive carrying readouts from ChatGPT or Claude, using the machine not instead of an adviser but to renegotiate the relationship with one. The consumer is already repricing the strategy layer in her head. Calling that evolution does not change the direction of the water.

The second relabel: a service, not a product

The cleanest relabel of the afternoon was also the most consequential, because it concerned conflict, and conflict is the thing the entire post-Hayne architecture was built to manage.

Lloyd’s observation was crisp. ASIC, he noted, is now scrutinising separately managed accounts because they are a product, as opposed to managed discretionary accounts, where it is a service.

Then the rhetorical flourish:

“Where did the vernacular come from that we’ve got a problem with vertical integration of a service? I never even heard of vertical integration of a service; I have for a product.”

It is a tidy line, and it dissolves nothing. The conflict in a vertically integrated model has never lived in the legal wrapper. It lives in two things: discretion and an in-house revenue line.

When an adviser exercises discretion over a client’s money and routes some of that money to a service their own group earns a margin on, the incentive is identical whether the regulator files the arrangement under SMA or MDA, product or service.

The label changes ASIC’s surveillance perimeter. It does not change the direction in which the dollars flow or the hand on the lever that sends them there.

Neil Younger, the Entireti Group chief executive, demonstrated the same logic with admirable candour, and in doing so revealed its internal seam.

Entireti, he said, made a deliberate decision never to enter asset management, because that would be too high a bar around conflict. It also declined to become a platform or super administrator; on the grounds it lacked the scale to compete. Both are genuinely principled lines, drawn precisely where the conflict becomes hard to defend.

And yet the firm runs managed portfolios, earns a margin on them, and does not regard that as a conflict, “because it’s centred around advice.” But portfolio construction is asset management. It is the selection and weighting of assets inside a client’s account, performed under discretion, monetised by the licensee.

The activity Entireti has ruled out as too conflicted to touch and the activity it performs for profit are the same activity wearing different clothes. The carve-out is asserted with conviction. It is never actually reasoned. “Centred around advice” is doing the entire load-bearing job, and it is a phrase, not an argument.

To be fair to everyone in that room, Younger drew a distinction that genuinely matters and deserves defending. He refused to lump the Shield and First Guardian failures, where ASIC has alleged in court a deliberate intent to benefit distributors at the expense of clients, in with ordinary structural conflict.

He is correct. Alleged misconduct and managed conflict are different categories, and a profession that cannot tell them apart will regulate itself into paralysis. But the existence of a worse problem does not certify the lesser one as solved. It just makes the relabel easier to wave through.

The third relabel: structured appropriately

The third move was the most familiar, because it is the one every integrated institution reaches for eventually. UniSuper, Gregory acknowledged, operates a vertically integrated model. But the fund, he said, is “structured appropriately.”

It has a group executive accountable for advice, an advice business unit with independence and a direct line to board committees, and it owns and operates its own approved product list.

This is governance language, and governance language has a specific function in these conversations. It answers the question “is the conflict managed?” while leaving untouched the prior question of whether the structure should generate the conflict in the first place.

Owning and operating your own approved product list is offered here as evidence of control. It is equally a description of the conflict itself. The fund decides which products its advisers may recommend, and the fund manufactures products.

Independence reporting lines mitigate that. They do not dissolve it, and a single entity carrying both a Best Interests Duty for its advisers and a Best Financial Interests Duty for its members holds two obligations that can pull in opposite directions on the same afternoon.

The deeper tell is in the justification. The model exists, Gregory explained, to solve access to education, guidance and advice at the scale of the membership. It is a serious and sympathetic aim. But the numbers he supplied undercut it.

UniSuper, with one hundred and sixty six billion dollars under management, employs seventy-nine advisers against six hundred and eighty thousand members. That is one adviser for roughly every eight thousand six hundred members.

Whatever that structure is solving, it is not solving scaled access to comprehensive advice, because the arithmetic forbids it. The scale argument and the delivery capacity point in opposite directions, and the gap between them is exactly the space into which the relabel expands.

What the reflex is actually protecting

Here is the inversion worth sitting with. Stop asking the question the profession kept asking that afternoon, which was some version of “is our label defensible?” Ask instead the question a hostile regulator would ask.

If you drew the org chart purely as a map of money flows, stripped of every euphemism, where does the client’s dollar travel, who touches it on the way, and who earns on each touch? Draw that diagram and the labels stop mattering. Service or product, evolution or disruption, structured appropriately or not, the lines on the map are identical.

The reflex is not dishonest. That is what makes it dangerous. Every speaker who reached for it is, by every available signal, trying to deliver good outcomes.

The relabel survives precisely because it lets you keep the margin while genuinely believing you hold the moral high ground. It resolves the cognitive dissonance between what the business model requires and what the client’s interest demands, and it resolves it in the direction that costs the incumbent nothing. That is its entire job.

Which is what Kodak’s executives were doing too. They were not villains. They were rational people protecting a profitable structure with a story about why the structure was fine, and the story was internally coherent right up until the market stopped agreeing.

Film was a service to memory. Digital was just an evolution they could manage on their own timeline. They were structured appropriately. Every reassurance was true in its own terms and collectively fatal.

The advice profession’s Kodak risk is not the AI agent. Gregory is right that execution and relationship are hard to displace, and Lloyd is right that the human plus machine model is the likely shape of the next five years.

The risk is the reflex itself: that the profession spends the window defending its margin structures with ever more refined vocabulary, while the consumer, already repricing the strategy layer on her phone, quietly concludes that the relationship she values and the conflicts she is paying for are two separable things, and starts shopping for the first without the second.

Kodak had the camera. It had the patents. It had the engineers and the brand and a thirty-year head start.

What it did not have was the willingness to let the comfortable label go before the market forced it to. The profession has the relationship, the execution, the trust, and a moat that is real.

The only question that matters is whether it can name its conflicts as plainly as it names its strengths, before someone less invested in the vocabulary names them first.

The label is not the thing. The money is the thing. It always was.