BUSAC

Applications

How the method applies to real decisions.

These hypothetical engagement patterns show the questions, interventions and evidence designs BUSAC uses. They are not client results. Published impact will appear only when measurement is complete and disclosure has been approved.

Every pattern below runs the same evidence chain, from locked baseline to attributed financial impact. The chain, explained

Hypothetical Engagement Patterns

Four patterns, one structure.

From underwriting an AI thesis to recovering revenue leakage: four examples of how we turn a value claim into an evidence-backed decision. Each pattern is written the way an engagement is designed: the value hypothesis, the data it needs, the bounded intervention, the metrics, the strongest feasible counterfactual and the rule that decides.

Hypothetical · Technology Value Diligence, buy-side

Underwriting an AI story in a tech-enabled services acquisition

Value hypothesis
The seller’s deck attributes a third of the forward EBITDA bridge to AI-driven automation of servicing; the deal team needs to know how much of that is evidenced, modelled or roadmap before it prices the bid.
Who captures the value
The acquirer, through the price: only what is V3 or V4 today enters the base case, and the V2 pools are priced as upside with the test that would prove them.
Required data
Data-room technology and vendor documents, pilot results with their baselines, servicing cost and volume history, staffing by process, platform and contract inventory.
Bounded intervention
Fifteen working days: claims register, data tests on two servicing processes with extracts, management and technology sessions, value pools sized as ranges.
Primary metric
Share of the claimed uplift that is V3 or V4 today, and the V2 range for the remainder with the assumptions stated.
Guardrail metrics
Vendor lock-in and key-person exposure, data quality, disintermediation risk to the core service, integration debt from prior acquisitions.
Strongest feasible counterfactual
Pilot results re-tested against a locked baseline where data allows; otherwise classified V2 with the counterfactual named as a post-close test.
Decision rule
The committee prices only V3 and V4 into the base case; V2 pools enter the upside case with the evidence design that would prove them in the first 100 days.
Hypothetical · Contract-to-Cash Leakage

Recovering unbilled and under-billed revenue at a PE-backed B2B services company

Value hypothesis
Contracts, rate cards, timesheets and invoices live in different systems, so work is delivered that is never billed, billed below the agreed rate, or billed without the indexation and surcharges the contract allows.
Who captures the value
The portfolio company, as recovered and future revenue on the same contracts; the record separates one-off recovery from the run-rate effect of the corrected billing rules.
Required data
Signed contracts and amendments, rate cards, time and delivery records, invoice lines, credit notes, twelve to twenty-four months of history.
Bounded intervention
AI extraction of commercial terms from contracts, matched line by line against delivery and invoicing for one business unit; a leakage register with root causes, and corrected billing rules in the invoicing workflow.
Primary metric
Leakage rate (delivered but unbilled or under-billed value as a share of contracted revenue) for the treated unit, against the locked baseline.
Guardrail metrics
Customer disputes and credit notes, days sales outstanding, invoice accuracy, account-manager time spent on billing.
Strongest feasible counterfactual
Phased rollout by business unit; untreated units serve as the comparison, with contract mix and volume reconciled to finance.
Decision rule
Scale to the remaining units if the attributed leakage reduction clears the agreed threshold with disputes within tolerance; otherwise narrow the rules or stop.
Hypothetical · Cost-to-Serve

Scaling operations without scaling costs at the same rate

Value hypothesis
As a PE-backed distributor grows, manual work in sales support, order processing and customer operations grows with volume, so cost-to-serve rises in step with revenue and caps margin expansion. A share of that work can be automated or routed better without loss of accuracy or service.
Who captures the value
The portfolio company itself, and the record says how: hires avoided at the plan’s growth rate, or hours redeployed to revenue work. Which of the two counts is agreed with the CFO before the build.
Required data
Order and ticket volumes by channel, handling time per step, headcount and cost by team, error and rework rates, re-contact rates, service levels, revenue by segment.
Bounded intervention
Automated intake and routing for one order channel and one team: document extraction, validation against master data, an exception queue for what the rules cannot settle, inside the existing order-management tool. People keep every exception.
Primary metric
Cost-to-serve per order for the treated channel, fully loaded, against the locked baseline; orders handled per FTE as the secondary.
Guardrail metrics
Order accuracy and first-time-right, lead time and service levels, customer complaints, overtime in the treated team.
Strongest feasible counterfactual
Phased rollout by channel, with untreated channels as the comparison and volume mix reconciled to finance; interrupted time series where channels differ too much to compare.
Decision rule
Extend to further channels and teams if the attributed reduction in cost-to-serve clears the agreed threshold with accuracy and service levels intact; otherwise modify the rules or stop.
Hypothetical · Platform & Integration

Post-acquisition platform integration at a services roll-up

Value hypothesis
Fragmented systems from successive acquisitions force manual re-keying and reconciliation, capping the capacity automation could release.
Who captures the value
The portfolio company, as capacity released and rework avoided; the record states whether that is counted as cost avoided at the plan’s growth rate or as hours redeployed.
Required data
System inventory, process cycle times, error and rework logs, staffing by process step.
Bounded intervention
An integration layer for one process across two acquired entities, with automation on the unified data.
Primary metric
End-to-end cycle time and capacity released for the treated process.
Guardrail metrics
Data quality, exception rate, control effectiveness, user adoption.
Strongest feasible counterfactual
Interrupted time series on the treated process, with the same process at untreated entities as a reference.
Decision rule
Extend to further processes or entities if the observed change is attributed above threshold; otherwise modify the design or stop.

Our Operating Standard

The process standards we hold ourselves to.

Standards for how every engagement is run. Not impact figures; there are none to publish yet.

How to read any impact claim

The standard we apply to our own measurement, and the questions we would ask of anyone’s.

  1. Locked baselineagreed before deployment, never restated
  2. Evidence design statedholdout, phased rollout, matched control, time series or reconciled before/after
  3. Evidence level statedV1 to V4 on every figure, never one blended number
  4. Guardrail metrics reportedthe material side effects, not only the benefit
  5. Reconciled to financethe figure ties to the P&L, with confounders named
  6. Retraceable to sourcewho can reproduce it, from which data, without the author in the room
Agreed Baseline and evidence plan

Before deployment, with the client, and recorded.

Weeks To a bounded deployment

From agreed hypothesis to a working intervention in a bounded part of the business.

6-step Evidence chain

Baseline, intervention, performance, operational impact, attribution, financial impact.

Stated Evidence level on every claim

V1 to V4, never collapsed into one number.

Where might value be hiding in your portfolio?

Bring one workflow, cost line or initiative. We will say where we would look first, and how we would test it.

Pressure-test one hypothesis Our evidence standard
Let’s Talk