Five firms: the same levers, five different bridges

How to read the archetypes

Each firm is described by the five things that most change its bridge: how it earns revenue today, what work converts first, how much of its work is reserved to a licensed human, which levers carry the bridge, and the trap that has caught firms like it. The dimension with no software equivalent is the licensed-human constraint, and it is the one that most separates the five.

The bridges are described in direction and sequence, not in points of margin, because the honest state of the evidence (hub, "What is missing") does not support a number that is not the firm's own. The case study (chapter 9) works one of them through with a model, and labels every assumption in it.

A. The compliance consultancy (the case study)

What it is. A firm of twenty to sixty people advising manufacturers and brands on product, chemical, supply-chain and sustainability compliance: which regulations apply, what the suppliers must disclose, whether the product passes, what to declare, and what changed this quarter. Revenue is annual retainers (priced, in practice, on expected hours), project fees for assessments and new-market entries, and a growing share of reporting work driven by regulation.

Revenue model today. Retainers and projects; recurring in habit more than in contract; realization under pressure because clients see the hours.

What converts first. Supplier data collection and chasing, substance and regulation screening, declaration and report drafting, change monitoring. These are most of the delivered hours and they are the "ready" units on a compliance platform.

Licensed-human constraint. Low. Few outputs need an accredited signature. What the client is buying is the firm's judgment on what the screening means and what to do about it, which is the judgment share and the premium tier.

Levers that carry the bridge, in order.

1. P1 and P6 confirm the ceiling and the fit; for this archetype the audit is mostly confirmation.
2. P9 and R1: retainers become platform subscriptions with an advisory tier. This is the whole game here, because the hours are visible and the client will otherwise expect the saving.
3. R2 and R8: the first platform-delivered line is continuous monitoring with the annual assessment included, which converts the episodic retainer into a subscription with history.
4. D1, D4, D2: units move; the firm's screening method is written into frameworks; analysts become reviewers; production hiring stops.
5. R3: scope expands across regulations and jurisdictions at marginal cost, which is where NRR comes from.
6. X3: client data on a shared platform needs consent renegotiated at renewal; R10 waits on it.

The trap. The client sees the platform and asks why it still pays the consultancy. R11 (the priced advisory tier) and R8 (the monitoring subscription) must land before D1 is visible to the client, or the conversion is a price negotiation the firm loses. The second trap is institutional: the platform company's own sales team sees the consultancy's clients as prospects, and the consultancy's partners see the platform as a competitor. The structure (P7) and the compensation formula (O7) have to make them one business.

Why this is the case study. It is the archetype where the conversion ceiling is highest, the licensed constraint lowest, and the pricing risk clearest. If the thesis works anywhere, it works here; if the pricing conversion fails here, it fails everywhere.

B. The accounting, bookkeeping and tax firm

What it is. A regional firm with monthly bookkeeping and advisory clients, seasonal tax compliance, and some attest work. Private-equity investment in this sector is already extensive and the structure it forces is known: attest in a CPA-owned firm, everything else in a separately owned one.¹

Revenue model today. Monthly fees plus seasonal compliance; recurring share already high; fixed fees already dominant in the sector.²

What converts first. Bookkeeping and reconciliation, return preparation, notice response, workpaper assembly. The named vendor evidence in the delivery chapter comes from this archetype.³

Licensed-human constraint. High. Returns and attestations require a credentialed signer; ownership rules require the alternative practice structure.

Levers that carry the bridge, in order.

1. P5 and X2 first: the structure and the sign-off design decide what the platform may do and in which entity.
2. D1 and D2: preparation moves to the platform; the pyramid of preparers becomes a smaller team of reviewers under the signers. This archetype has the largest base and therefore the largest D2 effect.
3. D3: sampling before signature, because the signer owns the error.
4. R3 and R8: pricing is already fixed, so the revenue levers are scope (advisory, more entities, more filings) and converting seasonal work into year-round monitoring.
5. R6: the smallest clients, previously unprofitable, served on the base tier.
6. O7: partner compensation in a formalised partnership is the hardest to change and the most necessary; the APS complicates who is paid from where.

The trap. Automating preparation without the review model raises liability faster than margin: the signer is still liable, the preparer who caught errors is gone, and the platform's accuracy is unmeasured by anyone independent. D3 before D1, always, here.

C. The legal-process and contract-services provider

What it is. An alternative legal services provider or a legal-process firm: contract review and abstraction, entity management, filings, immigration paperwork, document review. Not legal advice. The sector's own live example is a platform that bought a 300-person provider of this kind and sells a fixed-fee, platform-plus-people offering to corporate legal departments.⁴

Revenue model today. Hourly and per-matter; the hourly culture is in the people as much as the contracts.

What converts first. Abstraction, first-pass review, standard filings, entity updates. The high-value judgment (what to negotiate, what to advise) stays human and is in many jurisdictions reserved.

Licensed-human constraint. Medium, and jurisdictional. Advice is the practice of law; process is not; the line is drawn by state and country, and P4's review of client outside-counsel guidelines will find AI-use restrictions and disclosure requirements more often here than anywhere else.

Levers that carry the bridge, in order.

1. P4 and P5: what the clients permit and what the law reserves, before anything else.
2. P9 and R1: per-matter fixed fees first, then subscriptions for recurring matter types. The hardest conversion of the five, because of the people.
3. O7 at the same time: the origination-and-hours culture is strongest here; the compensation formula changes or the pricing does not.
4. D1 and D3: units move with 100 percent review for anything that leaves the firm, by rule, because the hazard (fabricated content in legal documents) is documented.⁵
5. R5: speed is sold, because deadlines are real in this work.
6. X1: the E&O conversation happens before the first unit moves.

The trap. Partners resist P9 because it ends the hour as the unit of value, and they frame the resistance as quality. D3's published error rate is the only answer that works, and O7 is the only reason they accept it.

D. The technical and engineering services firm

What it is. Testing support, certification management, environmental and permit reporting, regulatory engineering. Much of the work is physical or accredited; much of the rest is data gathering, report assembly and regulatory cross-referencing.

Revenue model today. Project fees, with recurring permits, renewals and monitoring as the seed of a subscription.

What converts first. Report assembly, regulatory cross-referencing, renewal tracking, data compilation from tests the firm or others performed.

Licensed-human constraint. High in places. Stamped drawings, accredited laboratory results and notified-body steps stay human by rule; the platform prepares beneath them.

Levers that carry the bridge, in order.

1. P1, P5 and P6 decide whether to bid at all. If the reserved and physical share is most of the cost, this is a conventional services firm with a better back office, not a conversion.
2. R8: permits and renewals become subscriptions; this is the clearest recurring seed in the five.
3. D1 and D4: the firm's reporting method, which is often decades of templates, is the asset to capture.
4. X2: the sign-off step is designed into every reserved unit's flow.
5. R3: more sites, more permits, more jurisdictions per client at marginal cost.

The trap. Platform fit overestimated. A deal team that sees "compliance reporting" and assumes a compliance platform fits may find that the platform runs none of the engineering content and the build list exceeds nine months. P6 is the bid decision here, and it should be run by the platform team with the firm's actual reports.

E. The claims, billing and revenue-cycle services company

What it is. A firm that codes, submits, follows up and appeals claims for healthcare providers, paid a percentage of collections or a fee per claim. The sector is already automating (63 percent of medical-group leaders leaning toward more revenue-cycle automation in 2026; a third using AI for coding or charge review) and the independent accuracy evidence is scarce by a regulator's own account.⁶

Revenue model today. Outcome-priced already: a share of what is collected. Conditions two and condition P9 do not apply; the margin lever is pure.

What converts first. Eligibility checks, coding, claim creation and submission, denial drafting, payment posting. Nearly everything is codifiable; the constraint is error cost, because the payer measures it.

Licensed-human constraint. Low to medium. Payer rules, not licensure, govern; some appeals require a clinician's input.

Levers that carry the bridge, in order.

1. D6 first: the firm's denial rate is the manual baseline, and the payer already publishes it in the form of denials.
2. D1 and D3 together: coding moves to the platform with sampling against the denial rate; here D3 and D6 are the same lever.
3. R5: capacity to appeal more claims faster is revenue for the client and for a firm paid on collections, and the reported overturn rate on appealed denials makes the case.⁷
4. D2: the largest base of the five, and the biggest reshape.
5. B3: payer-specific rules make the framework library the roll-up's asset; every add-on in the same payer markets converts faster.
6. R6: the small practice with a hundred claims a month, previously unserved, served on the base tier.

The trap. An error rate the payer punishes. Because pricing is already outcome-based, there is no P9 to hide behind: a platform that codes worse than the people it replaced shows up in collections within a quarter, which is the client's revenue and the firm's. The sampling plan is the business.

The comparison