10 — Professional & Business Services
Pipeline, staffing and delivery margin modelled as one engine — where a proposal, a bench week and a write-off all resolve to margin per delivered hour.
North Star
Margin per delivered hour
The value engine, end to end
The same six moves we run everywhere, expressed in the physics of this industry.
Where does delivery margin actually leak?
Margin leakage map by engagement type
What drives win rate, utilisation and write-off?
Causal delivery model
What happens to margin if we change staffing or scope policy?
Engagement and capacity twin
Who drafts, staffs and monitors day to day?
Delivery support agents
Is client confidentiality absolute across every hop?
Confidentiality and independence control file
Does the margin hold on the next engagement?
Outcome contract + engagement monitor
Where the value is
Before a twin is built, we agree where the money sits and what moves it.
Value pool
Pipeline conversion
Driver
Proposal quality and qualification discipline
Measured as
Won value per proposal hour
Value pool
Utilisation
Driver
Bench time and staffing fit
Measured as
Billable ratio
Value pool
Delivery margin
Driver
Scope creep, rework, write-off
Measured as
Realisation rate
Value pool
Knowledge reuse
Driver
Whether prior work is findable and reusable
Measured as
Hours saved per engagement
Use cases
Each one is a number somebody already owns, a twin that models it, and an agent that moves it inside a defined authority. Filter by function.
22 use cases
Extracts obligations, dates and liabilities from the contract estate and monitors whether the organisation is actually meeting them.
Answers internal policy questions with the paragraph attached, and says plainly when the policy does not cover the situation.
Finds the data, redacts what must be redacted, and produces the response pack inside the statutory clock.
Correlates signal across identity, endpoint and network, promotes the alerts that matter and writes the incident timeline for the analyst.
Learns the expected shape of every feed and raises the alarm when a pipeline goes quiet — before the report is wrong.
Keeps every model, prompt and agent in one register with its owner, validation evidence and drift status — the artefact the regulator asks for.
Forecasts departure risk by team and role, identifies the drivers that are actually within management control, and prices the intervention.
Matches candidates to roles on capability evidence, with bias testing on the shortlist and a human owning every rejection.
Plans shifts against forecast demand, skills and fatigue rules so service levels hold without overtime absorbing the gain.
Finds duplicate payments, missed discounts, off-contract spend and price creep across the ledger, and closes each at source.
Automates reconciliation, flags the entries that need judgement, and drafts the variance commentary with its supporting figures.
Replaces the quarterly spreadsheet with a rolling forecast that updates on live signal and shows what changed since last week.
Scores leads on what closed before, retires the pipeline that will not, and gives sellers a shorter list they trust.
Segments on behaviour and value rather than on demographics, and keeps the segments current as behaviour moves.
Picks the next touch that most changes the outcome for this customer, and suppresses the contact that would only cost money.
Matches content to intent at the moment of intent, inside brand and compliance guardrails that are enforced, not advisory.
Reaches the right audience from context and first-party signal, so performance does not depend on identifiers that are disappearing.
Sends the expensive channel only where it pays, using a profile built on response rather than on list size.
Maps the intent, routes it to the cheapest competent resolver, and gives that resolver the case already summarised.
Predicts which tickets will breach, pre-drafts the resolution, and escalates the ones that need a human before the SLA runs out.
One grounded answer across documents, tickets, wikis and recordings — with the source, the date and the owner attached.
Answers a complex planning question end to end — options, budget, constraints — and hands over a plan rather than a list of links.
Outcome twins
Each twin owns a set of levers. Each feature moves one lever. Each agent executes inside a defined authority.
Qualify harder and propose faster on the work that actually closes.
Levers
Features
Multi-agent layer
Proposal Agent
Drafts from precedent with sources cited internally.
Match the right people to the right work before the bench forms.
Levers
Features
Multi-agent layer
Staffing Agent
Proposes staffing that fits both the client and the career path.
See scope creep and rework while they are still cheap to correct.
Levers
Features
Multi-agent layer
Margin Agent
Flags engagements trending to write-off with the cause.
Make prior work findable, reusable and safe to reuse.
Levers
Features
Multi-agent layer
Precedent Agent
Retrieves usable precedent inside client-confidentiality walls.
Unit economics
Set your volume, switch features on or off, and move each impact to your own evidence. Everything is expressed per delivered hour.
Delivered hours per year
900,000
Contribution per delivered hour
$47
Per-hour economics; substitute per-engagement or per-seat for managed-service models.
Annual program investment
$3.5m
Pipeline Twin
Volume — Proposals per partner-hour
4.0% — Precedent found and reused instead of rewritten.
$1.7m
annual value
Capacity Twin
Unit cost — Cost per delivered hour
4.0% — Bench weeks converted into billable fit.
$5.0m
annual value
Delivery Margin Twin
Leakage — Write-off and rework leakage
24.0% — The change order is raised while the client still expects it.
$4.8m
annual value
Knowledge Twin
Unit cost — Research hours per engagement
3.0% — Answers with the source document attached.
$3.7m
annual value
Pipeline Twin
Unit revenue — Revenue per hour
2.0% — Price set against outcome evidence, not against a rate card.
$3.3m
annual value
Value by feature
Annual value contributed by each active feature at current settings.
Value and ROI vs volume
How engineered value and ROI move as delivered hours per year change. The marker is your current setting.
Cumulative value vs investment
Where the curve crosses the investment line is payback.
Value mix by lever
Which lever the engineered value is actually coming from.
Baseline contribution
$42.3m
Engineered annual value
$18.5m
ROI on program
429%
Payback
2.3 months
Directional model. In an engagement every number here is replaced by your measured baseline, signed off by finance, and written into the outcome contract before any agent goes live.
Scale
Prove realisation and utilisation on a bounded portfolio.
Same twins, new engagement types; confidentiality walls carried across.
Shared runtime, per-client isolation, one margin-per-hour number.
Integrations
Vendor-agnostic by design. Connectors are added per engagement — no platform lock-in.
Practice management
CRM & pipeline
Knowledge & collaboration
Data & AI
Regulations and standards
Compliance is a design input, not a review gate. Every agent action is logged, attributable and reversible.
Client confidentiality and privilege
Hard isolation between client corpora across every agent hop.
Independence rules
Automated conflict checking before any engagement is accepted.
GDPR
Lawful basis for personnel and client personal data in any model context.
EU AI Act
Transparency for automated staffing and performance-affecting decisions.
Records retention
Engagement evidence retained and disposed to schedule.
Standards supported
Next industry
Health & Life Sciences
