An essay · May 2026
The CRM era is over.
Dealership Revenue Intelligence is the operating system that replaces it.
Ricardo Gamon · Founder, Gladius · 20-year operator
The old job, and the new one
CRMs were built to store contacts.
DRI is built to create revenue.
When the category we call “auto CRM” was designed in the late 1990s, the job was clerical. Reps made phone calls, scribbled on paper, and a CRM existed so a sales manager could see who had been called and who hadn’t. The data model was a filing cabinet with search. The reporting was a count of rows. For two decades, every subsequent generation — Reynolds, CDK, VinSolutions, Elead, DealerSocket, DriveCentric — shipped a faster filing cabinet with a prettier skin, and called the result innovation.
That job no longer exists. A modern dealership is not short on contact rows. It is short on the four things that actually move revenue: speed-to-lead under a minute, after-hours coverage, accurate desking math at the curb, and a follow-up cadence that runs whether the rep logs in or not. None of those are storage problems. They are orchestration problems. A filing cabinet cannot solve them, no matter how many AI features you bolt onto the side.
Dealership Revenue Intelligence is the category that starts from the new job. It is not a CRM with AI. It is the layer the CRM used to pretend to be — the system that captures every signal across the floor, books the appointment without asking permission, and surfaces the one decision the manager actually needs to make this hour. Storage is incidental. Outcomes are the product.
“The 2003 architecture cannot make the 2026 number. That is the whole essay.”
What broke
Four cracks in the foundation.
None of these are speculative. They are the receipts of the last five years.
01 · Reps don’t enter data
Every CRM on the market depends on the salesperson to type. They don’t. They never have. The data rots inside a week, the reports lie inside a month, and by the end of the quarter the manager is making forecasting calls against a database that bears almost no resemblance to the floor. The vendor blames training. The dealer blames the rep. The truth is the architecture assumes free human labor that nobody is going to give.
02 · DMS lock-in is a tax, not a feature
The CDK outage of June 2024 is the receipt. Three weeks, 15,000 rooftops, no DMS, in some cases no payroll. The same handful of vendors that charge dealers a five-figure monthly fee to integrate with their own data also control whether that data leaves the building. That is not software. That is a toll booth. A category built on top of a toll booth cannot evolve faster than the toll booth lets it.
03 · AI bolted onto a 2003 architecture
You cannot ship a 2026 outcome through a 2003 data model. The incumbents have spent two years adding the word “AI” to product pages while the underlying schema, event bus, and integration model stayed exactly where they were. The reason their AI chat goes off the rails is not the model. It is that the system underneath it does not know what a deal is. It knows what a contact row is.
04 · Compliance debt the incumbents cannot pay down
FTC Safeguards is now the law. GLBA always was. The cost of retrofitting column-level encryption, MFA on every seat, and a real incident-response posture into a monolith that was never designed for any of it is higher than rewriting the product. The incumbents will not rewrite. So they will keep filing extensions and hoping no one is looking. Someone is looking.
What DRI looks like instead
Four contrasts. One operating system.
01 · Auto-capture, not data entry
Calls, texts, web chat, walk-ins, and DMS events flow into a single timeline without a human typing them. Reps do not log into the CRM to update the CRM. The CRM is wherever the conversation happened, watching the conversation, and writing itself.
02 · An open ecosystem, not a toll booth
Standard webhooks, a documented API, and one-click migrations off VinSolutions, Elead, DealerSocket, and DriveCentric. We do not charge integration fees. We do not own the dealer’s data. We do not need to.
03 · AI-native from the schema up
Every entity — lead, deal, appraisal, follow-up — is designed around what a model needs to reason about it. The system knows what a desking pencil is, what an objection is, what an equity opportunity is. Not because we told it. Because the schema makes those first-class.
04 · Defense-native — AWAIS as the category’s moat
Autonomous Web Application Intelligence System — AWAIS — runs underneath every endpoint across every vertical we ship. 66 active rules, 261 catalogued attack patterns, IQ-145 self-grading engine. The incumbents pay a vendor for a WAF. We built ours, it learns, and it watches all five products at once.
“The job is to make the schema match the floor — not the other way around.”
The five fronts
Same operating system. Five verticals.
Gladius ships five products on one shared platform — Auto CRM, BDC, Detail, Turf, and Stone fabrication. That is not a roadmap. That is what is live today, with three pilot rooftops in production. The reason we can ship a full vertical in weeks rather than years is that every vertical sits on the same DRI primitives: the same lead graph, the same desking engine, the same auto-capture pipe, the same AWAIS defense layer.
That mesh is the part the incumbents cannot copy. Their auto product is a separate codebase from anything they sell into adjacent verticals — if they sell into adjacent verticals at all. Ours is one codebase, five brands, one defense system, one playbook. Every rule AWAIS learns inside an automotive tenant gets applied the same hour to a detail tenant and a turf tenant. That is not a marketing line. It is a federation.
We named it Dealership Revenue Intelligence because we started in automotive and we are not going to pretend otherwise. But the “dealership” in DRI is a proxy for any high-ticket, ops-heavy local operator with a sales floor, a service bay, and a back office. The category travels.
The contrarian bet
Not a better Reynolds.
The obvious move, the move every funded startup in this space has made for a decade, is to build a slightly better Reynolds and try to win on UI. That move loses. The incumbents will not be displaced by a prettier CRM; they will be displaced by a category that makes the CRM irrelevant. We are not building a better Reynolds.
We are building the category that makes Reynolds irrelevant. The hard part is not the code — the code is shipping every week. The hard part is convincing a 65-year-old GM that the system he has cursed at for twenty years can actually be replaced. That conversation is what the next two years of this company are.
If we are right, in five years the question on every dealer’s desk is not “which CRM do we use,” it is “what is our DRI score.” If we are wrong, the incumbents will keep charging $11,000 a month for software your grandkids would laugh at, and we will go back to selling cars.
“We are not building a better Reynolds. We are building the category that makes Reynolds irrelevant.”
Receipts
Don’t take our word for it.
Every claim above has a public page behind it. Read the ones that matter to you.
- AWAIS — live federation mesh
The defense layer running underneath every vertical, visualized.
- Compare — Gladius vs. the incumbents
Feature-by-feature against VinSolutions, Elead, DealerSocket, DriveCentric.
- Status — live ecosystem uptime
30-day uptime grid across all 5 verticals.
- Customers — three live rooftops, named
Pilot rooftops in production, not a wall of logos we don’t have.
If this sounds like the next ten years to you,
book fifteen minutes.
The founder runs every first call. Fifteen minutes, calendar-only, no SDR layer between us. If the math doesn’t work in your store, you’ll know on the call and we’ll part friends. If it does, you’ll walk away with a migration plan written by the person who wrote the product.
End of essay · Gladius · 2026