TrustLayerLabThe go-to-market · CoverageFit
A $140M layer no product sells.
The plan to take CoverageFit to market: who buys it, what it costs, how it is sold, and the order of the rollout. The market is sized from public data, every assumption is labeled, and the five-year model on this page is live. Change the assumptions and see the impact.
Executive summary
Sell the judgment to the base TrustLayer already owns, starting in construction.
CoverageFit sells the answer no product on the market sells: whether a vendor’s coverage actually fits the work. It launches as the paid tier on top of free verification, priced inside the market’s published band of $13 to $29 per vendor per year.[4] It goes first to existing TrustLayer accounts, where selling costs about half of winning new customers,[12] and the same brokers and channel then carry it to new buyers. Construction opens the rollout because the exposure is clearest there. At the baseline assumptions below, the line reaches operating income in year two and has recovered its full investment in the same year.
The market
How big it is, and why it is open now.
Every tool on the market sells the same confirmation: the certificate matches the requirement. The question that decides a claim, whether the coverage fits the work, is answered by brokers, by hand, when anyone asks at all. No product sells that answer. That matters for two reasons: the first product in names the category and sets the price, and for TrustLayer it is the natural extension of PolicyReview. The reading already exists; the judgment is the missing layer, and what used to be expert manual work now costs cents per read.
| Input | Value | Basis |
|---|---|---|
| Exposure-heavy, vendor-managing US organizations | ~400,000 | Construction ~800K establishments, higher-exposure subset ~250 to 350K[5], plus CRE ~100K, hospitals ~6.1K, and depository lenders ~8.9K[7] |
| Average vendors tracked per account | ~60 | COI tracking is typically adopted at 25 to 50 active vendors; 60 is a working estimate for accounts large enough to buy. |
| Share of vendors that carry real risk | ~40% | Illustrative. One of the numbers only TrustLayer can confirm. |
| Price per read vendor per year | ~$15 | Inside the public verified band of $13 to $29.[4] |
| SAM = organizations x risky vendors x price | ~$140M / yr | Derived from the rows above. |
The field
TrustLayer already publishes tracker-by-tracker comparisons on its own site[15], so the matrix does not repeat them; it asks the one question those pages do not, which layer each product occupies. Every tool tracks the certificate. Only the TrustLayer stack reads the policy, and nothing on the market judges the coverage against the job.
| Tool | Tracks and verifies the certificate | Reads the policy | Judges coverage against the job | Where it focuses |
|---|---|---|---|---|
| myCOI / illumend | ● | ○ | ○ | COI compliance across industries.[2] |
| Jones | ● | ◐ | ○ | Real estate and construction; human auditors plus AI review policy documents against the requirement checklist.[13] |
| Bunker | ● | ○ | ○ | COI tracking that is also a licensed broker, so it can place coverage to close a gap.[14] |
| Certificate Hero | ● | ○ | ○ | The agency side: issuing certificates.[3] |
| TrustLayer · Verification | ● | ○ | ○ | The platform core: the certificate verified against the requirement, at scale.[1] |
| TrustLayer · PolicyReview | ● | ● | ○ | The policy read, page-cited, integrated in the stack. |
| TrustLayer · CoverageFit | ● | ● | ● | The judgment layer. Proposed; the demo shows the reading and the judgment running. |
● does it ◐ partial or different ○ does not. Cells are marked conservatively: Jones earns its partial because its auditors and AI genuinely review policy documents.[13]
The economics
What it costs, and what it returns.
Two tiers, both software. Verification lands the account free.[1] The reading is the paid tier, priced inside the market’s verified band of $13 to $29 per vendor per year.[4] There is no human in TrustLayer’s loop and no services tier: the product informs, the customer’s own broker rules when coverage has to change, and the margin stays software margin.
Per vendor per year is the market’s unit and the unit in the model below. It does not have to be the invoice. For a large account, package its vendors as one flat monthly fee: the sale becomes one known number against one $43,000 claim,[10] which is an easier line for procurement than a per-vendor rate card. The economics underneath do not change.
The five-year P&L
The five-year P&L for the CoverageFit line, down to operating income and the return a capital decision turns on. Revenue counts every vendor under the reading, in existing accounts and in new ones; the sales-cost assumption is low because the selling starts with the base TrustLayer already owns. This is the CoverageFit line only, not a model of TrustLayer as a whole, and it leaves out the retention and deal-win effects argued below.
The four inputs that move the outcome. The rest are held at the baseline assumptions listed below the model. Move any slider and the scenario becomes Custom; Reset returns to the baseline.
| Assumption | Value | Basis | Type |
|---|---|---|---|
| Price per vendor / year | $13 to $29 | Public verified-coverage band; tracking alone is $6 to $8.[4] | Sourced |
| Cost to serve (inference) | cents / read | Measured by the extractor eval in this repo, priced on Anthropic’s published rate card. | Measured |
| Gross margin (burdened) | 85 to 92% | Support, infrastructure, and success drive this, not inference. Set above the ~79% median SaaS subscription gross margin (Benchmarkit / KeyBanc 2025) because the work is mostly automated. | Assumption |
| Sales & marketing | 30% of new ARR | Selling to current customers. Winning a new one runs ~$1 to $2 per $1 of first-year ARR; selling more to a current one runs ~$0.20 to $0.27 (The SaaS CFO). 30% sits a touch above that, to be conservative. | Assumption |
| Engineering | ~$300K / yr | One senior engineer, fully loaded (1.25 to 1.4x base), plus AI and infrastructure, matching how this POC was built. A conventional team runs toward $1M and the slider covers that. Sales and marketing is its own line, above. | Assumption |
| Annual churn (this line) | 12% | Median B2B SaaS gross revenue churn across 939 companies (Optifai). Sets the lifetime behind LTV / CAC. | Benchmark |
| Discount rate | 20% | Hurdle rate for a growth-stage internal initiative. NPV positive at this rate clears the bar. | Assumption |
| Vendors under CoverageFit | Scenario input | The one figure TrustLayer’s customer base decides, not the open web. It is the primary uncertainty. | Assumption |
The type column says where each value comes from. Public data on this market is thin, so several of these are judgment calls tied to the nearest credible benchmark; this table is exactly where TrustLayer’s real numbers would replace them.
Left out of the numbers on purpose
The P&L counts only the subscription, the part public data can size. Two larger effects are real but cannot be put in dollars from outside TrustLayer, so they are argued here, not added to the model.
An account that runs its coverage judgments through CoverageFit builds a record it would lose by leaving. Switching means re-collecting context and giving up the history of what was flagged and ruled, which raises the switching cost on the whole relationship, not just the add-on.
Not counted above: sizing the retention lift needs TrustLayer's churn data.
Against trackers that stop at the certificate, CoverageFit is something the others cannot show in a bake-off. In a competitive evaluation it is a reason to choose TrustLayer, which pulls in the whole account, not only the CoverageFit line.
Not counted above: sizing the win-rate lift needs TrustLayer's pipeline data.
Who buys
Who buys it, and how it is sold.
The buyer is defined by exposure, not company size: any company managing vendors whose work can cause a loss. Below: the order of the segments, the four people in the deal, and how the sale actually runs.
First. The exposure is clearest and the trades are where the judgment runs deepest.
Second. Contractors and service vendors on the property.
Later. On-premises vendors with exposures a generic requirement misses.
Later. Borrowers and their contractors, where coverage protects the collateral.
| Role | Who | What they care about | The message that lands |
|---|---|---|---|
| Economic buyer | VP Risk or CFO | Avoided loss and E and O exposure | A small annual fee against a six-figure denied claim: one average work injury costs about $43,000.[10] |
| Champion | Risk manager | Catching gaps, covering more vendors | Catch the gap a certificate check cannot see. The live demo, then the first broker-confirmed gap in the pilot. |
| User | Risk analyst | No new data entry | Every input is read from documents already in the workflow. |
| The customer's broker | Reviewer and referral path | Staying essential to the client | Decision support, never a determination. A flag routes to the broker the customer already uses, with its citation, so the review gets faster and the judgment stays theirs. No cost to TrustLayer. |
The sale starts inside the base. Free verification lands the account,[1] the product offers the reading when a risky vendor appears, and sales works the existing accounts heaviest in the covered trades and states, at about half the cost of winning a new customer.[12] The brokers who already sell TrustLayer carry it to new buyers. Renewal is built in, because everything in an account expires: policies renew, certificates lapse, endorsements change mid-term, new vendors arrive. The reading runs on every new document, the platform tells the account what it needs and when, and that is why the fee is annual.
The moat
Two of the three inputs already live in TrustLayer.
The fit judgment takes three inputs: the vendor’s policy, the insurance requirement in the contract, and the work the vendor was hired to do. The policy is the easy one, anyone can read a PDF. The requirement and the job are the hard ones, and TrustLayer’s customers already load both onto the platform with every engagement. A competitor starting fresh has to convince customers to hand over contracts and vendor records before its product can say anything at all, and every judgment TrustLayer records, per trade and per state, deepens an advantage a new entrant starts without.
This moat rests on one thing being true: that the requirement and the job enter TrustLayer early, at onboarding or contract signing, not late in some downstream step. If they arrive late or incomplete, the moat softens, and the independent-licensor structure below gets more attractive.
How to hold it
Own the brand, keep the integration. The question is not whether to build CoverageFit but how TrustLayer should hold it: three structures, trading the moat against reach.
Simplest and fastest. All the margin stays home, nothing new to brand or maintain.
Cost: No category, no standalone reach, nothing to point a broker or carrier at. The capability stays invisible, the way Fill and Sign was before it was launched.
TrustLayer owns CoverageFit as a named product with its own page and go-to-market. It ships inside the stack and can be sold on its own to adjacent buyers. This keeps the data-position moat, creates a category to name, and opens a second market without giving up the account base.
Cost: Real brand and go-to-market investment, and a second product surface to carry.
CoverageFit as its own company, licensing the technology to TrustLayer. It can be the neutral analysis layer across competing trackers, earn on both ends, and ring-fence the insurance-adjacent liability.
Cost: Forfeits the integration moat. A standalone reader has to re-collect the job and the requirement it no longer owns, invites channel conflict, and can end up competing with the platform it depends on.
The recommendation follows from the moat: productize CoverageFit as a named layer so it has a category and a second market, but keep it inside TrustLayer’s data position, because that position is the moat. Break it into a separate company only to become the category-neutral analysis layer across every tracker, which is a different bet than winning as TrustLayer’s.
The rollout and the measures
Five stages. Each opens on evidence from the one before.
Each stage opens when the build reaches the milestone it depends on and the prior stage produces the evidence named on its tag: a measured accuracy number, a broker-confirmed gap, adoption beyond one account. No customer relies on the judgment in a trade or state where it has not been proven on real policies first.
| Stage | Opens on | What happens |
|---|---|---|
| Stage 1 | A measured accuracy number | Prove the reading on TrustLayer’s own documents. No customer is involved, so no customer can see a wrong answer. The output is a measured accuracy number and a written list of what the reader cannot handle yet. |
| Stage 2 | Live inputs working | One private pilot: a short vendor list, one state, the trades where the judgment runs deepest, a customer who knows it is a pilot, and the customer’s own broker ruling on every flag. |
| Stage 3 | The first broker-confirmed gap | A small group of design partners: a few more construction accounts, still one state. Enough to see whether adoption grows beyond a single account. |
| Stage 4 | Adoption beyond one account | Open construction broadly. Each new state and trade is proven on real policies before any customer there relies on the judgment. |
| Stage 5 | Construction economics holding | Commercial real estate, healthcare, and lending, in that order, each proven the same way first. |
The measures
Five numbers, and the decision each one drives. Targets belong to TrustLayer once the baseline is known; the framework does not.
| Measure | What it tells you | The decision it drives |
|---|---|---|
| First flag ruled on | The product worked in front of the customer once. | No activation means fix the workflow before selling more. |
| Share of risky vendors read (the one to watch) | Whether accounts adopt the reading across their vendors. | Climbing means invest. Flat in the pilots means stop. |
| Broker-confirmed gaps | Real coverage gaps found by the product and confirmed by the broker. | One confirmed gap clears the pilot. A wrong flag that reaches a decision halts it. |
| Broker confirmation rate | How often the broker agrees with the flag. | Low agreement means fix the judgment, not the sales pitch. |
| Win rate by segment | Which vertical closes fastest. | Point the spend at the segment that closes first. |
The boundary and the risks
What TrustLayer fills in, and the risks that are real.
This plan is built on public data. The table names what TrustLayer adds to make it a budgeted, dated operating plan. Below it, the three risks that matter for a product with this many inputs, each with its mitigation.
| Built here, from public data | TrustLayer's to fill in |
|---|---|
| The size of the market, built from public data. | How many customers TrustLayer has and how many vendors each one manages, which set the obtainable revenue. |
| A price range from the public band, with the model to test any point in it. | The final price. |
| The sales plan: who buys, in what order, through which channel. | The adoption rate and willingness to pay, which only come from customer conversations. |
| The measures and the decisions they drive. | The targets, the budget, the staffing, and who owns it. |
- The product misses a gap.A wrong flag is visible: the broker refutes it and the judgment gets fixed. A miss is invisible: the product says nothing, and the customer assumes the vendor is covered. Mitigation: accuracy measured against TrustLayer's own document library before any customer relies on the judgment, every new trade and state proven on real policies first, and a plain statement of what the product does and does not evaluate.
- The inputs arrive incomplete.The judgment is only as good as the requirement and the job data behind it. Where contracts, exhibits, or vendor records enter the platform late or thin, the product has less to judge. Mitigation: the product shows what it is missing per vendor instead of guessing around it, and the missing-input rate is tracked per account from the pilot forward.
- A customer relies on a judgment and a claim still goes bad.The serious risk for any product that opines on coverage. Mitigation: every flag is decision support with its citation, the ruling belongs to the customer's broker exactly as it does in the manual process today, and the record shows who ruled and why. Same legal posture as today, with a better paper trail.
Sources
Every number, attributed.
Public market and insurance facts are cited. TrustLayer-specific dollar figures are illustrative and tied to the public verified band, because account-level prices are quoted, not published.
- [1]TrustLayer (network of 517,000+ companies; Starter free up to 50 vendors; tiers Starter / Pro / Complete). source
- [2]illumend / myCOI (buyer-side COI compliance platform). source
- [3]Certificate Hero (agent and issuer side certificate issuance). source
- [4]Vertikal RMS, COI tracking pricing guide (self-service about $6 to $8; full-service verified about $13 to $29 per vendor per year). source
- [5]CPWR / US Census County Business Patterns, construction establishments (NAICS 23 ~800K; specialty-trade subcontractors ~511K; building construction ~252K, 2022). source
- [6]US Census / The Roofing Brief, roofing contractors (NAICS 238160 ~24,500 establishments, 2022). source
- [7]American Hospital Association Fast Facts (~6,100 US hospitals); FDIC (~4,487 insured banks); NCUA (~4,455 insured credit unions). source
- [8]2024 KeyBanc Capital Markets & Sapphire Ventures SaaS Survey and High Alpha / OpenView 2024 Benchmarks (NRR median ~101 to 110%, top quartile ~120 to 130%; gross margin ~75 to 85%). source
- [9]Benchmarkit 2025 and Bessemer efficiency benchmarks (CAC payback efficient under 18 months; LTV to CAC floor 3:1). source
- [10]National Safety Council, Injury Facts (average cost per medically consulted work injury ~$43,000, 2023). source
- [11]Carolina Risk Partners and CFMA, general liability height exclusions (work above three stories or 35 feet); The Hartford, monopolistic workers' compensation states (ND, OH, WA, WY); IRMI, CG 20 10 vs CG 20 37; Saxe Doernberger & Vita, New York Labor Law 240/241 action-over exclusions. source
- [12]SaaS capital-efficiency benchmarks: expansion revenue is roughly twice as capital-efficient as new-logo acquisition. source
- [13]Jones (insurance compliance for real estate and construction; human auditors plus AI agents reviewing COIs and policy documents). source
- [14]Bunker (COI tracking; also a licensed broker that can place coverage for a non-compliant vendor). source
- [15]TrustLayer's own published tracker comparisons (e.g. TrustLayer vs Bunker). source