Skip to main content
Rooftop
Rooftop

Investor overview

The intelligence layer dealership software never had.

Three workflow applications create a continuous vehicle record—and a compounding platform advantage across acquisition, merchandising, sales, service, and ownership.

What follows is the company, the architecture, and the current state of the product — including the parts that are not built yet and the numbers that do not exist yet. This page carries no offering terms of any kind.

Request Investor BriefingView Platform Architecture

How the flywheel turns

  1. 01

    Applications

    A dealership adopts one workflow application because it solves one department's problem today.

  2. 02

    Structured Vehicle Events

    Doing that work writes timestamped, structured facts about a specific VIN into the Shared Vehicle Record.

  3. 03

    Better Decisions

    The next appraisal, disclosure, or reconditioning call is made against what the store already observed rather than against a guess.

  4. 04

    Higher Workflow Value

    Each application is worth more inside the record than it would be standing alone, because it reads what the others wrote.

  5. 05

    More Applications

    That makes the second and third application an easier decision than the first — and each one returns more events to the record.

Every step writes to and reads from the Shared Vehicle Record at the centre, and step five returns to step one. Worth pressing on, because it is the claim the company rests on: this is an architecture argument about where the inputs originate, not a machine-learning claim. Rooftop asserts no proprietary model and no exclusive data rights, and the automatic cross-application handoffs that make the loop close on its own are still being built.

The thesis in twenty seconds

Rooftop begins with three valuable dealership applications and compounds its defensibility through the continuous vehicle intelligence record those applications collectively create.

  • Market Problem

    Fragmented tools and siloed data create blind spots, rework, and lost value across the vehicle lifecycle.

    The problem in full

  • Product Entry

    Start with high-impact workflows that deliver fast results and establish the shared vehicle record.

    The entry strategy

  • Business Model

    Application subscription with usage-based components. Structure designed for expansion and retention.

    What is charged

  • Go-to-Market

    Direct sales with implementation support, training, and customer success across dealership groups.

    The current motion

  • Defensibility

    The shared vehicle record gets more valuable as more workflows and dealerships participate.

    And what is not a moat

  • [KEN: two of the five cards above are the approved comp copy and both are contradicted by pages this site already publishes. (1) Business Model says “usage-based components”; /pricing and section 07 say the opposite — not per VIN, not per deal, not per appraisal — and that is a load-bearing differentiator against the incumbent stack. (2) Go-to-Market says “direct sales with implementation support, training, and customer success across dealership groups”; section 08 says the current motion is founder-led selling, and no customer-success function and no dealer-group deployment exists. Approve the comp copy and change the product, or approve the site’s version and change the comp. Both cannot ship.]

Company status at a glance

Product status
LivePilotEvery label resolves through the same source a dealership sees on the product status page, so this page cannot quietly disagree with it. Only Live means a dealership can use it today. Capability-by-capability detail is in section 09.
Traction
Verified onlyOne deployment, disclosed as a related party, with results pending an approved measurement methodology. No customer count, no pipeline, no partnership, and no outcome figure appears anywhere on this page. The full version.
Financials
Available under NDARevenue, unit economics, and the cap table belong in a data room, not on a page reachable by URL. No TAM, no ARR, and no target valuation is published here.

01Company thesis

Sell three things a dealership already needs. Compound the record underneath them.

Rooftop begins with three valuable dealership applications and compounds its defensibility through the continuous vehicle intelligence record those applications collectively create.

Rooftop is the platform

The AutoCurb application, the AutoLabels application, and the AutoFilm application run on it. None of the three is a platform, and none is a module of a suite a dealership has to buy whole.

Each application stands on its own

A dealership buys AutoCurb because it needs to acquire vehicles, AutoLabels because it needs its disclosures to be right, AutoFilm because it needs its video to lead somewhere. None of them requires believing the thesis.

The record is the second-order product

Every one of those workflows writes structured, timestamped facts about a specific VIN. Sold separately, they are three useful tools. Connected, they are a continuous history of a vehicle that no single-workflow vendor can assemble.

The compounding is the company

The applications are how Rooftop gets into a store and how it earns its subscription. The record is what makes leaving expensive and what makes the next application worth more than the last.

02Automotive retail problem

A dealership knows a great deal about every car it touches, and keeps almost none of it.

The failure is not a lack of software. A franchised store runs a DMS, an inventory system, a CRM, an appraisal tool, a merchandising tool, and a service system. The failure is that each one describes the vehicle for its own purpose and then stops.

The vehicle is described three times, by three systems

Acquisition describes a car one way, merchandising another, service a third. The same VIN carries a different truth in each system, and no system is wrong from its own point of view.

Evidence is created and then discarded

A technician puts a car on a lift and learns more about it in twenty minutes than an appraisal tool learns from a VIN and a mileage field. That knowledge stays in the repair order.

Disclosure is produced by hand

The label on the glass, the addendum, and the advertised price on the website are maintained separately. When they disagree, the dealership finds out from a customer or from a regulator.

The vendor stack is priced against the dealer

Per-VIN and per-deal pricing means a dealership's best month is its most expensive month. That is a pricing model the incumbent cannot unwind quickly, because it is load-bearing revenue.
  • [KEN: supply any third-party market sizing you are prepared to defend line by line in diligence. No TAM, SAM, or dealer-count figure appears on this page, because an indefensible market number is the first thing a serious investor pulls on.]

03Why now

The cost of recording a vehicle's history fell. The cost of not recording it did not.

The documentation burden is permanent, not tied to one rule

The FTC CARS Rule was vacated by the Fifth Circuit in NADA v. FTC on 27 January 2025 and never took effect. What binds a dealership is FTC Act Section 5 and state disclosure law — unaffected by that decision, enforced case by case, and satisfied by contemporaneous records rather than by intent. A dealership that cannot reconstruct what it disclosed is exposed regardless of which federal rule is in force this year.

Capture is finally cheap

Structured evidence — video, timestamped findings, signed disclosures, condition capture — now runs on the phone already in the technician's hand. The cost of writing a vehicle's history down has collapsed; the systems that would store it have not caught up.

The buyer has changed shape

A shopper arrives having already read everything a dealership publishes about a car. Vehicle information that is verifiable is worth more than vehicle information that is merely present.

Operators are re-examining the stack

The founder built these applications because he was paying for the alternatives from inside a store. That is an operator's observation from one desk, not a market study.

04Shared Vehicle Record architecture

One record per VIN, written by every application that touches the car.

The Shared Vehicle Record is the common data layer beneath the three applications. Each application writes to it in the course of doing its own job, and each reads what the others wrote. It is not a data warehouse a dealership has to populate; it is the residue of work the store was already doing.

Identity
What the vehicle is — VIN decode, trim, equipment, and the disclosures attached to it.
Evidence
What was seen and when — inspection findings, photographs, video, and the timestamp on each.
Condition
What state the vehicle is in, recorded as it changes rather than as a single snapshot at appraisal.
Decisions
What the dealership chose to do, who approved it, and what they were looking at when they approved it.
Value
What the vehicle was worth at each point, and the inputs behind that number.
Ownership
Who held the vehicle, what was disclosed to them, and what they signed.

Worth stating plainly, because it is the claim an investor should press on: this is an architecture argument, not a machine-learning claim. Rooftop does not assert a proprietary model, a trained system, or a data advantage it has not built. What it asserts is that the inputs arrive from workflows a dealership pays for anyway.

The current delivery status of every connected capability is in section 09. Two are live, one is in pilot, one is being built, and one is planned.

05Three-application entry strategy

Three separate reasons to buy, all of them writing to the same record.

Rooftop does not sell a platform migration. It sells one application that solves one department’s problem, and it is designed so that a store never has to adopt more than one thing at a time.

  • auto(CURB)

    applicationLiveVehicle Acquisition

    Acquire vehicles directly and manage every step through check request.

    Replaces the trade-in widget and the instant-cash-offer form.

    AutoCurb in full

  • auto(LABELS)

    applicationLiveVehicle Identity and Merchandising

    Turn inventory data into verified, customer-facing vehicle experiences.

    Replaces the addendum printer and a generic e-sign.

    AutoLabels in full

  • auto(FILM)

    applicationPilotSales and Service Video

    Turn video into an actionable customer decision.

    Replaces the generic walkaround video and standalone video-MPI tools.

    AutoFilm in full

All three applications

06Expansion model

Land on one application. Expand to the connected account.

Entry pricing is public: From $1,999 for AutoCurb and From $599 for AutoLabels, per rooftop, per month. A rooftop running all three as one connected account moves to a custom platform plan, which is scoped and quoted rather than listed — and the AutoFilm application carries early-access terms rather than a published rate while it is in pilot. Rooftop publishes what it can stand behind and quotes the rest, rather than printing a number for a product that is still moving.

The unit does not change

Pricing stays per rooftop, per month, at every stage. It does not become per VIN, per deal, or per user as a store expands, so expansion is a decision about capability rather than about volume risk.

The second application is an easier sale than the first

The buyer has already been through implementation, already has users trained, and can see the record the first application has been building. That is the designed shape of the motion.

The connected deployment is quoted, not listed

A connected account is scoped into a custom platform plan rather than sold from a rate card, because what it contains differs by dealership and one of the three applications is still in pilot. What that plan includes and how it is quoted is an open executive pricing decision, not something the marketing site resolved. The public version is on /pricing.

Groups are quoted, not published

Two or more rooftops move to a scoped quote. The per-rooftop model stays; what changes is the rate, the implementation sequence, and the administration layer above the stores. No group rate card exists, and none is implied.
  • [KEN: supply observed expansion — any rooftop that has added an application after its initial subscription, which application, and over what period. Until that exists, the expansion model above is a designed path and not a measured one, and this page says so rather than quoting a net-expansion figure.]

07Business model

Subscription, per rooftop, published where it can be stood behind.

What is charged

A monthly or annual subscription per rooftop. Implementation, configuration, training, and product updates during the term are inside the subscription. There is no separate setup fee on the published rates.

What is not charged

Not per VIN, not per deal, not per appraisal. A store that sells 500 vehicles a month pays the same subscription as a store that sells 50, bounded by the limits confirmed in the order form.

Why the entry prices are public

A dealer can qualify Rooftop without talking to anyone, which is the opposite of how the incumbent stack is sold. It also means every number on the price sheet is one a customer can check.

Why the rest is not

An application in pilot is not priced, and a connected deployment is quoted. Both are on /pricing with the reason stated next to the missing number rather than left to be discovered on a call.
  • [KEN: supply unit economics — ACV, gross margin, CAC, payback period, and net revenue retention — into a diligence data room, not onto this page. Revenue is confidential and is deliberately absent here. This placeholder exists so that nobody fills the gap with an invented figure.]

08Go-to-market

Operator to operator, with an evaluation instead of a giveaway.

Founder-led selling

The current motion is the founder talking to people who run stores, in the language of someone who still runs one. It is credible and it does not scale, and both halves of that sentence are true today.

The entry offer is an evaluation

The Structured 60-Day Proof of Value is a scoped sixty-day evaluation on one rooftop with a measured baseline, success metrics agreed in writing before go-live, reviews at day thirty and day sixty, and a deployment decision at the end. Nothing converts automatically. See the full evaluation process.

Published pricing does the qualifying

A dealer who cannot afford the product finds out before a call rather than after three. That shortens the funnel at the top rather than at the bottom.

What has to be proven next

That the motion works with a buyer the founder has never met, in a store the founder has never worked in. That is the first thing to test, and it is not tested yet.
  • [KEN: supply the go-to-market motion beyond founder-led selling — any channel, referral, reseller, OEM, or vendor relationship that actually exists and has been agreed in writing. No partnership is named on this page, because none has been substantiated.]
  • [KEN: supply pipeline only if you are prepared to define the stage criteria behind it. No pipeline, no signed LOIs, and no paid pilots are claimed here.]

09Product status

What is actually running, capability by capability.

Statuses below come from the same source as the customer-facing product status page, so this page cannot quietly disagree with what a dealership is told. Only Live means a dealership can use it today.

Applications

  • AutoCurbLive

    Vehicle Acquisition

  • AutoLabelsLive

    Vehicle Identity and Merchandising

  • AutoFilmPilot

    Sales and Service Video

Connected capabilities

  • Shared accountLive

    One dealership account across the applications the rooftop subscribes to, with central user and role administration rather than a separate login per application.

  • Shared Vehicle RecordLive

    Each application writes to one record per VIN, so identity, evidence, condition, and decisions accumulate on a single timeline instead of in three systems.

  • Cross-application reportingPilot

    Reporting that spans more than one application is running with selected dealerships. It is not generally available and is not part of what the connected price buys today.

  • Automatic service-to-appraisal handoffIn Development

    Being built. Today a service finding reaches an appraisal because a person moves it, not because the platform moves it automatically.

  • Public APIPlanned

    Approved roadmap, not yet in development. No endpoints, credentials, or documentation are published, and no integration should be scoped against it.

The public view of the same information is on the product status page and the changelog.

10Verified traction

One deployment, disclosed as a related party, with results not yet measurable.

What Rooftop can verify today, in full

  • One deployment. Harte INFINITI, in Hartford, Connecticut. It is an active operating environment rather than a demonstration site.
  • It is the founder’s own store. Harte INFINITI is where Rooftop’s founder runs F&I. It is disclosed everywhere on this site as a design partner and a related party, not as an arms-length customer, and it should be read that way in diligence.
  • Results are pending. No outcome figure is published from that deployment, because no measurement methodology has been agreed and approved. A number produced inside the founder’s own store, without a baseline and a defined method, is not evidence.
  • Nothing else is claimed. No second store, no cohort, no customer count, no retention figure, no pipeline, no partnership, no certification, and no uptime number appears anywhere on this page.
  • [KEN: supply any rooftop beyond Harte INFINITI, and its relationship to you.]
  • [KEN: supply the date the AutoCurb application went live at Harte INFINITI. /changelog carries a date, but that file states its own dates are best-effort, so it is a lead rather than a source.]
  • [KEN: supply whether the AutoLabels and AutoFilm applications are deployed at Harte INFINITI today, since when, and which teams use them. Until then they are not counted as deployments here.]
  • [KEN: supply VINs under management, and how many carry more than one application. Both are countable and neither is guessed at here.]
  • [KEN: supply an approved measurement methodology — baseline, metric definitions, and who signs off — before any outcome number from Harte INFINITI is published anywhere.]

11Design partners

A design partner is not a customer, and this page does not blur the two.

What a design partner is here

A dealership that runs the product in real operating conditions and shapes what gets built, with a relationship to Rooftop that is disclosed rather than implied. Its results are not reference results.

Harte INFINITI

Hartford, Connecticut. The founding design partner, and the store where Rooftop's founder runs F&I. The related-party relationship is stated on the customer page, on the homepage, and here, in the same words each time.

Why the distinction matters to an investor

A design partner proves the product can run. Only an arms-length customer proves someone will buy it. Rooftop has the first and is explicit that it has not yet demonstrated the second.

What would change the picture

A dealership with no relationship to the founder, paying for the product, in a store the founder has never worked in. That is the milestone worth watching, and it has not happened yet.
  • [KEN: supply any design partner beyond Harte INFINITI — the store, the relationship, and whether they have consented to being named. No second design partner is listed here.]
  • [KEN: supply written permission before any dealership name or logo appears on this page. None is displayed today.]

12Defensibility

The record is the moat, and it is not built yet.

Defensibility here comes from where the inputs originate, not from a model. The dealership generates the evidence in the ordinary course of running its business, and Rooftop sits at the point where that evidence is created rather than downstream of it.

The inputs come from inside the store

A buyer bidding on a vehicle it has never seen is working from a VIN and a description. A dealership running Rooftop bids on a vehicle its own shop already had on a lift, with the inspection in hand. That input cannot be purchased — the dealership already owns the channel that produces it.

Each application feeds the other two

Service video produces evidence an appraisal cannot generate on its own. Appraisal produces condition an addendum cannot generate on its own. A single-application competitor has no access to the inputs that would make its own product better.

Switching cost rises with the record, not with the contract

The cost of leaving is not a termination clause. It is the accumulated history of every vehicle the store has touched. That grows with time and with the number of applications in use.

What is not a moat, stated plainly

This is an architecture argument. Rooftop claims no proprietary model, no patent, and no exclusive data rights. It also has one deployment, so nothing here has been proven at more than one store, and the automatic cross-application handoffs that make the record compound are still being built.

The single nameable dependency

The AutoFilm application is the capture node for service evidence, and it is in pilot. The automatic service-to-appraisal handoff that makes the record compound is in development. Until both land, Rooftop is three good applications sharing an account and a record — which is a real product, but it is not yet the compounding one described at the top of this page.

13Leadership

One person, named, with nothing else claimed around him.

Ken Criscione — Founder

Twenty-five years in automotive retail before writing a line of Rooftop. He still runs F&I at Harte INFINITI in Hartford, Connecticut, which is why the applications look the way they do and why the first deployment is a related party. Background is on the about page.

What is deliberately not on this page

No executive team, no advisory board, no board of directors, and no named engineers. Leadership is one person today. An investor should treat founder concentration as the central operational risk in this company, and the honest version of that risk is more useful than a slide of advisor headshots.

  • [KEN: supply anyone else who should be named — employee, contractor, advisor, or board member — with their role, their commitment level, and their consent to appear. Nothing is listed until all three exist.]
  • [KEN: supply the specific hires that reduce founder concentration, in the order you would make them. That plan is the answer to the first question every investor will ask about this section.]

14Funding contact

If you invest in vertical software, the fastest version of this is a conversation.

Rooftop is founder-financed. This page carries no offering terms — no amount, no valuation, no instrument, and no statement that the company is or is not raising — and it should not be read as one. It describes the business and invites a conversation.

Request Investor BriefingView Platform Architecture
  • [KEN: have securities counsel review this page before publish. It is reachable by URL, which is why it carries no offering terms at all — a publicly reachable page describing an active private offering raises Regulation D general-solicitation questions.]
  • [KEN: the investors mailbox does not exist yet. CONTACT_READY is false in config/contact.ts, so both Request Investor Briefing links correctly fall back to the monitored inbox. Flip it only once investors@ is live and has been tested end to end.]
  • [KEN: the comp puts Request Investor Briefing in the site header on this route only. components/rooftop/Header.tsx takes no props and hard-codes Request Demo for every page, and it is shared by the whole site, so it was not edited. The action leads the hero here instead. Approve a one-line Header change — an optional CTA prop, or a route check for /investors — and the header will match the comp.]

Forward-looking statements

This page describes Rooftop’s business and product strategy. It is not an offer to sell or a solicitation of an offer to buy any security, and it is not investment advice. Forward-looking statements reflect current expectations and are subject to change. Product statuses, roadmap items, and timing described here may change without notice.

Call usBookCall Ken