Even Calypso Never Sent an Invoice

The Odyssey is in theaters, so the details are fresh for a lot of people right now. We can't stop seeing healthcare in it.

Odysseus spends seven years on Ogygia. He is held there by a goddess who is, by any standard, a generous host. She shelters him. She feeds him. She offers him immortality, which is a better product offering than anything in enterprise software. In seven years, she never hands him a bill.

The man stranded on that beach is the man who thought up the horse. The trick worked by exploiting the one law nobody in that world was willing to break — hospitality. A gift is sacred. You take it inside the walls. By the time anyone reconsidered, the Greeks were already in the city.

He built the horse to end the war and get his men home, and it came with a cost he did not price. The same law he turned against Troy is the one that keeps him on Ogygia: a host too generous to refuse. He is taken care of well enough that leaving looks unreasonable, and he knows he is trapped anyway. He just can't see a path off the island. Until he does.

This is where everyone who has signed a contract for healthcare software finds themselves. Trapped in paradise, aware of how trapped they are, unable to see the path home. Until now.

How a healthcare integration actually works

Two systems that need to share data require one connection. Add a third system and you need three. A fifth, and you need ten. Ten systems, and you have forty-five separate connections that need maintenance and can break at a moment's notice. Everything is built to run from point A to point B. No scalability, no economies of scale, no optimization. Just a bridge with a toll and a single driver. If you want to get anywhere else, you need a new bridge with another toll, built just for you.

That cost curve does not appear in any proposal. It never needs executive approval to balloon, and the only person stressing about it is the CFO.

Each of those connections is a point-to-point arrangement, purpose-built, holding a set of field mappings that somebody has to keep true. Your organization generates the test files. Your organization staffs the person who knows why the mapping breaks when the sending system updates. When it does break — and it breaks on your partner's release schedule, not yours — you open a ticket, you wait in a queue, and the clock runs whether or not anything moves, delaying revenue the entire time. The party best positioned to fix it is the party whose design produced the failure.

Nobody has to be scheming for this to work. The incentive does it unassisted. A business that sells connections has a roadmap requiring more connections next year, and the reliable source of more connections is your continued fragmentation. Ask any integration vendor whether they hope to sell you fewer interfaces next year.

The compounding effect is the part that gets undersold. Every new system you buy makes every existing connection more fragile, which means the cost of adding your eleventh product is not the price of the eleventh product. It is the price of the eleventh product plus the renegotiation of everything already wired. Past a certain count, organizations stop evaluating better software. Not because the software isn't better. Because the wiring makes the answer no before anyone reads the proposal.

That is why integrations are the Trojan horse of healthcare. One arrives as the solution to the problem it creates, you pay to bring it inside your own walls, and there is a celebration on the day it lands. What is inside is a permanent condition. Ten years later you are still trying to get home, and unlike Odysseus, you are being invoiced every month.

The part that is genuinely antiquated

Strip away the pricing and the labor and there is still an antiquated design decision underneath. The format healthcare actually runs on is older than the web, and it moves information the way a telegraph did. One side composes the message. An operator converts it into dots and dashes. An operator on the far end translates it back and hands over a copy. Everyone involved assumes no storm took out a utility pole along the way, and that nobody wrote Tony where the record said Anthony, leaving the whole thing to get lost in a sea of data mismatches.

FHIR modernized the format. It did not change the posture. Most of it in the field is still used to ship copies from one system to another on a schedule, because the systems underneath still assume the record lives in more than one place.

So that is what an interface still does. It makes a copy of your data and sends it somewhere else on a schedule. Every connection you maintain is another credentialed path out of your building, another copy of the record living in a system whose access controls you do not manage, another party you must trust in perpetuity, because the copy is never coming back.

Outside of healthcare, the rest of software stopped doing this. When we pay a vendor, we do not ask them for a copy of their general ledger. They send an invoice through Bill.com, and both sides get scoped access to that one invoice, with a record of who touched it. When it gets paid, the system confirms the money is moving and nobody is left wondering whether a check will arrive in the mail. The invoice is never duplicated, never re-keyed, never typed into a second system incorrectly. The permission is what moves, and what comes with it is trust, faster payment, and a better relationship with the vendor.

That is the actual difference, and it is not a feature — it is a posture. Permissioned access to a record that stays where it belongs, versus copies in flight forever.

Why we built NEXUS to get smaller

We built our integration platform, NEXUS. We would like it to have less to do every year. That sounds like a strange thing for a company to say, but the rationale is simple. We want to offer a better alternative, and a path to reach it that doesn't require paying more to get there.

NEXUS exists for everything outside the GoldenI boundary — labs, payers, pharmacies, public health reporting, and the system you are running today that you would like to leave. Unlike the rest of the category, it scales. We integrate with a laboratory once, and every organization on GoldenI inherits that connection. A practice can send orders to any lab, any pharmacy, or any partner already mapped, without a project, a quote, or a queue. For a partner nobody has mapped yet, the configuration tool reads the partner's own integration documentation and drafts the field mappings for testing. That is the design target: mapping work that doesn't require a data engineer on either side. Everything maps to one standardized data model. The lab maps to GoldenI. The EHR maps to GoldenI. From there, NEXUS is managing credentials.

This is real work and it is not going away soon. But NEXUS is the only product in our suite whose success looks like shrinking, and its scope contracts as more of the network stands on shared ground.

Products built on GoldenI do not integrate with each other. There is nothing between them to bridge — they share one data infrastructure, one identity model, one permission model, one audit trail. A result recorded in the laboratory product is not transmitted to the clinical product. It is the same record, read by someone with the right to read it. The list of connections does not exist, because the premise that produced it does not.

Here is what that means in practice. If a laboratory runs its orders and results on NEXUS, that shared data model means REMITUS, our revenue cycle product, is a decision to turn on rather than another integration to buy. No new mapping, no new project, no new bridge.

Wanting to shrink your reliance on point-to-point integrations is what forces our pricing to be honest. NEXUS is not built to be a profit center, because a profit center that shrinks is just bad business. You have already spent enough money scaling your integrations, and the goal is to help you start recouping that, not to add another line to it. No organization should stay with a vendor because leaving is too expensive. That isn't a partnership. It is a hostage situation with a support contract.

Everyone is still stuck on the island. We are here to help you build the raft.