Exit Boardroom · AI Exposure Audit · Full Sample Report
AI Exposure Audit: Greenfield Environmental
Environmental consulting: contaminated land assessment, EPA compliance reporting, remediation project management, and ecological field surveys for construction, mining and government clients. ~$4M revenue, 12 staff. Owner-operator: Kevin, 61.
Prepared by Paul Hopcraft · 7 July 2026 · Horizon: next 12 months
Private & Confidential (sample). For illustration only — do not circulate.
Executive summary
Greenfield is not in the clear, and it is not in trouble. It sits just above the midpoint, and that is exactly where a buyer's advisor does the most damage to a price. The fieldwork and the relationships that built this business over years are hard to copy. The reporting and monitoring work sitting behind them is not. Kevin has about two years before he wants out. That is enough time to move the number, but only if the two heaviest exposed lines get fixed rather than left as they are.
Biggest single risk: EPA compliance reporting & monitoring, scoring 73 / 100. It carries the highest gross margin in the business at 45%, and it is the single most exposed line on the board. The judgment is mechanical, the data already flows in automatically, and the only thing protecting it today is that no one has built the AI-native version yet. That combination — high margin plus easiest to copy — is exactly what a buyer's advisor prices down hardest.
How the scoring works
This is a rules-based score, not a gut feel. Same inputs, same score, every time. A buyer's advisor could pick up the method, apply it to any business, and land on the same number.
Each margin line is scored 0–100 on five dimensions. Higher means more exposed to a two-person AI-native team rebuilding it in 60–90 days.
| Dimension | Weight | High score means |
|---|---|---|
| DP — Digital vs physical | 30% | Screen-based, copyable without physical presence |
| JS — Judgment standardisation | 25% | Rule-based, repeatable — not deep expertise |
| RL — Relationship lock-in | 15% | Transactional, low switching costs |
| DA — Data accessibility | 15% | Data sits in software, easily piped |
| CR — Credential / regulatory moat | 15% | No licence or accreditation required |
Line score = (DP × 0.30) + (JS × 0.25) + (RL × 0.15) + (DA × 0.15) + (CR × 0.15), rounded. Overall = revenue-weighted mean of line scores.
Exposure heat-map
Every margin line, ranked from most exposed to least.
The single biggest revenue line, contaminated land assessments at 30% of revenue, sits mid-Amber at 53. The line that drags the overall up is EPA compliance reporting — only 22% of revenue but the most exposed line by a wide margin. The two Green lines together carry 28% of revenue and are the most defensible things in the business — and the two things that live almost entirely in Kevin's head.
Line-by-line findings
Each finding shows the five dimension scores, what drove them, how a two-person team would copy the line, and how defensible it actually is.
EPA compliance reporting & monitoring
Revenue: 22% · Gross margin: 45%
What it is. Producing the monthly and quarterly compliance monitoring reports clients need to keep their environmental licences current. The highest-margin line in the business.
What drove the score. Once the monitoring data lands, this is screen-based, template-driven report writing (DP 85, JS 80). The monitoring well data already feeds into Greenfield's database automatically from field loggers (DA 80). What holds it back from a perfect score is that Kevin has to be the nominated qualified person on the licence for several of the biggest clients (CR 45) — a real but soft barrier.
How a two-person team would copy it. Wire the monitoring-well data feed into an AI reporting pipeline that drafts the report against the same regulatory template, flags any exceedance for a human sign-off, and files it. The hard part of this line was never the writing, it was having the structured data — and that data already exists and already flows in on its own.
How defensible it is. Not very, as it stands. The compliance requirement is real, but the actual production of the report is exactly the kind of work software was built to take. The defence is to own the AI version in-house before someone else builds it, and to use the nominated-person requirement to bind clients into longer renewal terms rather than annual ones.
Remediation project management
Revenue: 20% · Gross margin: 28%
What it is. Running live remediation projects end to end. Coordinating contractors, tracking milestones, keeping regulators and clients informed, signing off validation reports. The second-largest revenue line.
What drove the score. A good part of this is coordination and comms that automate reasonably well (DP 65), but every site is different and Kevin is making real judgment calls on remediation approach as the work progresses (JS 50). Clients rarely switch mid-project (RL 45), which softens the exposure, and signing off validation reports needs an environmental consultant registration (CR 55).
How a two-person team would copy it. Automate the milestone tracking, the contractor scheduling glue, and the regulator/client status updates. The routine coordination is genuinely copyable. What a small team would struggle with is the on-the-ground judgment when a site behaves differently than expected, which is where the real value in this line sits.
How defensible it is. Defendable if the routine coordination is systematised and the judgment work is kept and documented as the differentiator. The risk is leaving it half-managed by hand, neither efficient nor protected. Formalise the "how we decide what to do when a site goes off-plan" playbook — that is the part a two-person team cannot buy off the shelf.
Contaminated land assessments (Phase 1 & 2 ESAs)
Revenue: 30% · Gross margin: 38%
What it is. Phase 1 desktop reviews and Phase 2 intrusive site assessments for contaminated land, following the AS 4482.1 protocol. The single biggest revenue line in the business.
What drove the score. Phase 1 is mostly desktop and report writing, but Phase 2 needs a crew out drilling and sampling on site, which pulls the physical dimension to the middle (DP 55). The protocol is standardised, but Kevin is the only one who signs off the exceedance judgment calls (JS 55). Lab results come back digital and land straight in the tracking spreadsheet, but the site-history desktop review is still manual (DA 65). Two senior consultants hold EPA site-assessor accreditation, which takes years to earn (CR 30) — the strongest moat on this line.
How a two-person team would copy it. They could take the Phase 1 desktop-review half fast: an AI workflow that pulls site history, drafts the preliminary assessment, and flags data gaps for a human. Phase 2 is a different story — it needs licensed people physically drilling and sampling on site, which a small team cannot shortcut.
How defensible it is. Middling, and split down the middle of the line itself. The accreditation and the physical fieldwork hold. The desktop review does not. The move is to own the AI-assisted desktop drafting in-house rather than let a competitor build it first, while leaning on the accreditation as the real long-term moat.
Ecological & flora-fauna field surveys
Revenue: 18% · Gross margin: 25%
What it is. Boots-on-the-ground species and habitat surveys — walking transects, identifying flora and fauna, assessing habitat for development and mining approvals.
What drove the score. Almost entirely physical fieldwork (DP 20). Species identification is judgment built over 20 years in the field, and most of it lives in Kevin's head until the report gets written (JS 30, DA 35). A scientific licence is required to survey threatened species, and that is not quick to obtain (CR 15) — the hardest credential moat in the business.
How a two-person team would copy it. They cannot, not in 90 days. Getting the scientific licence alone takes longer than that, before a single transect is walked. This is a genuine moat, not a soft spot.
How defensible it is. Highly, as a moat — but with a real transfer risk sitting underneath it. The expertise that makes this line hard to copy lives almost entirely in Kevin's head and is not written down. That is not an AI-exposure risk, it is a succession risk, and it matters just as much to a buyer: expertise a buyer cannot see documented is expertise a buyer will discount, licence or no licence.
Government panel & key-client relationships
Revenue: 10% · Gross margin: 40%
What it is. Nine years on the regional government's pre-qualified environmental panel, plus the trust and history with the handful of biggest clients that keeps the work coming without a competitive tender every time.
What drove the score. Deep relationship lock-in (RL 10) — getting onto a government panel takes a multi-year track record a two-person team cannot buy its way into (CR 30). Knowing how each agency actually wants things handled is not written down anywhere (JS 25, DA 30). The smallest revenue line, but the highest-margin one, and the hardest to replace.
How a two-person team would copy it. They cannot. Nine years of panel history and agency trust does not transfer by building a tool. The only way in is to actually earn the track record, the same slow way Greenfield did.
How defensible it is. Very, as a moat — with the same transfer-risk caveat as the field-survey line. The panel relationship is real, but it currently rides entirely on Kevin personally. A buyer pays for a relationship they can see will survive the handover, not one that walks out the door with the owner.
What the score means for a sale
51 is the revenue-weighted mean of the five line scores. It sits in the Amber band — labelled Elevated for an overall.
Roughly a third of Greenfield, by revenue, is genuinely hard to copy: the field surveys and the government panel relationships. The other two-thirds carries real exposure, led by the compliance reporting line at 73 — the most profitable line and the easiest for a small AI-native team to rebuild. A buyer's advisor will find that line in diligence. They always do. They will use it to argue the recurring compliance revenue is worth less than it looks on paper, and they will price it down.
Kevin has about two years. That is enough time to move the overall score down before a buyer ever sees the business, but it is tighter than three — there is less room to fix things slowly. The compliance reporting line is the single highest-leverage fix on the board: most exposed, highest margin, and fastest to defend, because the data already flows in on its own. Fix that line and the same recurring revenue a buyer would have discounted becomes revenue a buyer pays full price for.
The 90-day action plan
The point of the plan is simple: move the exposed lines from Red and Amber toward Green before a buyer prices them. Each move either lowers a line's exposure score or makes it harder for a buyer's advisor to discount.
Capture the government panel relationship history out of Kevin's head and into a system. Document the agency contacts, the unwritten preferences, the nine years of panel history in a form a buyer can see and trust. A relationship a buyer cannot see is a relationship a buyer will not pay for.
Document the ecological survey judgment and the site-assessor exceedance protocol. Write down the actual decision rules behind species identification calls and exceedance sign-offs. Convert tacit expertise a buyer would discount into a documented process a buyer can underwrite.
Stand up AI-assisted EPA compliance report drafting in-house. Take the most exposed line in the business (73) and build the thing a two-person team would build — automated draft from the monitoring data feed, with a human sign-off step before it goes out. Once Greenfield owns the efficient version, a copyist has nothing to undercut.
Move the Phase 1 desktop review to an AI-assisted workflow. Take the exposed half of the land-assessment line (53) onto an AI-assisted process — the Phase 2 fieldwork and the exceedance sign-off stay with the accredited assessors. Keep the fast, cheap gain in-house rather than leaving it for a competitor.
Formalise the remediation project-management playbook. Document how Kevin decides what to do when a site goes off-plan mid-remediation — the judgment call that currently drives most of that line's 55 score. A written playbook a buyer's team can follow is worth more than the same knowledge locked in one person's head.
Bundle compliance reporting and land assessments into a single retainer. Combine the now-automated compliance monitoring (73) with the assessment work (53) into one ongoing service clients buy together. Bundled recurring revenue with switching costs is harder for a buyer's advisor to peel apart and discount.
Show the exposure score falling. Re-score after the moves above and document the drop from 51. A score moving in the right direction is a story a buyer pays for.
Your report
Free.
By application. 4 audits a month.
This is what you get — your real margin lines, scored the same way, with a 90-day plan built around your actual business and your actual exit timeline. I do it personally, with you, not a template you fill in alone.
Get yours free — apply for your audit →
If you plan to sell in the next few years, you want this now. Not in the data room.