NSAI commercial steering SETUP Scene 01 of 21 · Cover
Project Signal · lifecycle frame

Commercial steering across the services lifecycle.

A neuro-symbolic AI engine that surfaces evidence at six points of decision — improving services revenue and gross profit through better decisions, not new tools.

Scope Three services lines · $600M revenue
Frame 3 phases × 2 dimensions · six cells
Audience CFO, CRO, services GM, practice leadership
Executive summary · the headline in 60 seconds SETUP Scene 02 of 21 · Executive summary
What this is, what it produces, what we're asking for

Six decisions in the services lifecycle. One engine. Auditable margin uplift.

The bottom-line claim +$14–26M annualized GP +1.0 to +1.9 percentage points blended margin lift on a $600M services book at realistic capture. Reaches +$18–35M at mature steady state — ramping over 18-24 months as managed services renewals roll through.
Calibrated against
SPI · Service Leadership Index · A&E benchmarks
Honest commitment:
your firm's number
produced by the GP First Instance
01 · The mechanism Six points of intervention, calibrated against the firm's own data. Three lifecycle phases — lead, proposal, delivery — each with a revenue and a cost dimension. Six cells, one substrate. The engine surfaces evidence at the moment of decision — inside the operator's existing tool, not as a new system.
02 · The deployment sequence Where the heavy levers concentrate determines where to start. Managed services first — largest pocket, recurring revenue. Project services second — highest dispersion, multiple dominant cells. Staff augmentation third — smaller pocket, operational discipline. Single-sequence rollout, not big-bang.
03 · The attribution architecture Every recovered dollar is traceable down to specific decisions. Six primary KPIs feed three intermediates — pipeline quality, won-deal margin, delivered margin realization — which feed services revenue and gross profit. Per-decision logging plus holdout cohorts produce audit-grade receipts.
04 · The ask Run the GP First Instance: 10 weeks, your data, your number. Resale-first scope. Engine runs on actual operational data. By Week 10 you have your specific capture rate — not industry estimates — plus a data-readiness assessment for services rollout. The framework is ours; the number is yours.
Read this slide alone and you have the headline. Read the rest of the deck and you have the receipts.
NSAI commercial steering · structural frame MECHANISM Scene 03 of 21 · The frame
The commercial lifecycle · six decision surfaces · one substrate

Three phases. Two dimensions. Six places where the engine intervenes.

Every services engagement moves through lead, proposal, delivery. At each phase, decisions get made about revenue (what we expect to make) and cost (what we expect to consume). The cells are surfaces of intervention. The substrate is what makes them one system.

Phase →
Dimension ↓
Phase 01Lead
Phase 02Proposal
Phase 03Delivery
Dimension 01 Revenue what we expect to make
Cell 01 Demand-shape signal Which accounts and markets are showing buying signals; which existing customers are about to expand. Calibrated against actual conversion patterns.
Cell 03 Capacity-aware pricing Margin floors that move with utilization, bench, and forward commitments. Static templates replaced with calibrated functions.
Cell 05 Booked-to-recognized control Recognition trajectory + scope-stability + customer-health signals per engagement. Daily forecast built from active engagement reality.
Dimension 02 Cost what we expect to consume
Cell 02 Pursuit allocation Which opportunities deserve pre-sales investment given portfolio capacity and opportunity-cost across pursuits.
Cell 04 Cost-to-serve estimation Role-mix, ramp, risk premium calibrated against the corpus of comparable past engagements. Estimation queries actuals, not memory.
Cell 06 In-flight margin steering Daily predicted-delivered-margin per engagement; drift signals at week 14, not week 26; scope-leakage candidates surfaced for change-request conversion.
The cells aren't six tools. They're six surfaces of one substratethe same engine, calibrated differently, surfacing evidence at the moment of decision in each phase.
Cell 01 · lead × revenue MECHANISM Scene 04 of 21 · Demand-shape signal
Cell 01 · lead phase · revenue dimension

Where is demand actually about to materialize?

Today: pipeline reflects what sellers have qualified, not what the market has signaled. The engine watches signals continuously across both the existing customer base and the addressable market — surfacing where to focus, before the deal is in CRM.

Coordinates
Lead × Revenue
CELL 01
+ The lever · what the engine does
The engine maintains a continuous signal model that tracks buying-readiness indicators across two surfaces: existing customers (environment growth, contract end-dates, sponsor changes, escalation patterns, expansion footprints) and target markets (publicly available signals: funding events, leadership changes, technology decisions, regulatory triggers). The output is a ranked surface of where demand is most likely to materialize next quarter and the next two quarters out — scored against historical conversion patterns for similar signals at similar accounts.
Frontline experience
In the operator's tool
What should I be working on this week?
CRM · AE morning view · book of business
AccountNorthwind Health · existing customer
Last QBR47 days ago
Active deals0 (renewal in 6 months)
Demand-shape signals · ranked
Northwind Health · environment +38%, contract due in 6mo, new VP started 3wks ago
Acme Manufacturing · ITSM volume +62%, two cloud workloads added Q2
Globex Logistics · sponsor change, support tier escalation pattern
"3 accounts most likely to need a scope-up conversation in the next 90 days. Open any to see signal evidence."
Open Northwind See all signals Snooze
Leadership experience
In the executive's view
"Where is the BU's demand actually heading?"

Sales leadership sees the demand-shape map for the BU — which segments are heating up, which are cooling, where the existing customer base is generating expansion signals vs. where new-logo work needs to fill the gap. Strategic resource decisions — hiring patterns, marketing focus, vertical investment — informed by signal density, not by last quarter's pipeline.
Primary KPI
Demand-signal precision
Of the accounts the engine flagged as having buying signals, what percentage actually generated qualified opportunities in the predicted window? Trended over time. Decomposed by signal type. Holdout cohort of un-flagged accounts provides the counterfactual baseline.
Attribution Every flagged account logged with timestamp + signal evidence. Every actual opportunity logged in CRM. Match is mechanical. Holdout cohort isolates causation.
Data needed
CRM accounts & opportunitiesExisting-customer telemetry (RMM/ITSM)Contract end-dates & sponsor recordsExternal market signals (intent, news, hires)Historical signal-to-opportunity conversion
● ready● cleanup● build
Cell 02 · lead × cost MECHANISM Scene 05 of 21 · Pursuit allocation
Cell 02 · lead phase · cost dimension

Which pursuits deserve sales investment?

Pre-sales architects, deal desk, executive sponsorship are scarce capacity. Today they get spread across all qualified pursuits. The engine surfaces opportunity cost per pursuit so the sales team allocates against expected portfolio-net value.

Coordinates
Lead × Cost
CELL 02
+ The lever · what the engine does
The engine tracks pursuit cost in real time — pre-sales hours, deal-desk attention, executive sponsorship time, demo engineering — and computes portfolio-net contribution per pursuit: cost vs probability-weighted return, accounting for capacity displaced from other pursuits in the same capability pool. Pursuits are ranked not by raw deal size but by net portfolio value under current capacity — surfaced at qualification gates, in pursuit reviews, and at every escalation request.
Frontline experience
In the operator's tool
Should I really be spending pre-sales time on all of these?
CRM · active pursuits · ranked by net contribution
Active pursuits14 in qualified or later
Pre-sales hours320 committed this month
Bench-to-pursuit ratio1.8 (tight)
Portfolio-net ranking
O-7218 Acme cloud migration · net +$240k · pursue
O-7245 Globex platform · net +$180k · pursue
O-7301 Initech replatform · net −$45k · de-prioritize (displaces 2 stronger pursuits)
"Bottom 4 pursuits in your queue net-negative under current capacity. Recommend deprioritize or qualify-out."
Review bottom 4 See math Override
Leadership experience
In the executive's view
"Are we chasing the right deals with the right intensity?"

Sales leadership sees the pursuit pipeline ranked by portfolio-net contribution under current capacity, not by raw pipeline value. Pursuits in the bottom quartile by net contribution flagged for de-prioritization or qualify-out. Sales-resource allocation steered weekly against the ranking — not reactively against whoever shouts loudest.
Primary KPI
Pursuit efficiency
Pursuit hours invested per dollar of won engagement. Sub-decomposed: win rate on pursuits scored above the engine's recommended threshold vs. win rate on pursuits scored below. Two complementary measures showing both efficiency and selectivity gains.
Attribution PSA pursuit time-tracking + CRM win/loss. Per-pursuit logging of engine recommendation + operator decision (pursue / de-prioritize / qualify-out) + outcome. Holdout cohort optional but powerful.
Data needed
CRM opportunities + stages + probabilitiesPre-sales time-tracking (PSA)Capability requirements per pursuitHistorical pursuit-to-win patternsCapacity state by capability pool
● ready● cleanup● build
Cell 03 · proposal × revenue MECHANISM Scene 06 of 21 · Capacity-aware pricing
Cell 03 · proposal phase · revenue dimension

What price protects margin given who's available?

Today: static margin floor regardless of capacity state. The firm over-prices when capacity is available (loses winnable deals) and under-prices when capacity is scarce (commits scarce people to thin-margin work). The engine makes pricing capacity-state-aware.

Coordinates
Proposal × Revenue
CELL 03
+ The lever · what the engine does
Bid pricing reflects current utilization in the capability pool, bench state by skill, and forward commitments already made. Margin floor lifts when seniors are scarce, relaxes when capacity is available — calibrated against the historical price-vs-win-probability curve for this customer profile and deal shape. The architect sees the floor with calibration visible. They can override; the override is logged with reason.
Frontline experience
In the operator's tool
What's a defensible floor for this bid right now?
PSA · bid — opportunity O-7218
CustomerAcme Health · cloud migration
Scope$4.8M · 14-week delivery
Static floor26% margin
Quoted26% · $1.25M GP
Capacity-aware pricing
Senior architect util 91% · bench thin in cloud
Forward commitments load Q2-Q3 senior pool
"Recommend floor 28% (+2pts). Win rate at this floor: 64%. Foregone margin if held at 26%: $96k."
Accept Override Escalate
Leadership experience
In the executive's view
"Are we leaving margin on the table or losing winnable deals?"

Pricing exec sees firmwide pricing patterns: which segments price aggressively, which leave margin on the table, where the floor needs recalibration. Override patterns visible — if architects systematically override down by 3+ points in a segment, the floor needs review. The pricing template is no longer a static document — it's a continuously calibrated function with operator behavior as a feedback signal.
Primary KPI
Capacity-aware pricing impact
Two paired measures: win rate at engine-recommended floor vs. win rate at static floor (where override happens, comparison is direct). And margin captured per won deal vs. counterfactual at static floor. Both produce auditable receipts of pricing-decision quality.
Attribution Both recommended price and accepted price logged. Counterfactual is direct (static floor is known). Win/loss recorded in CRM. Each closed deal produces a clean attribution record.
Data needed
Historical bids: priced + won/lost + marginCustomer segmentation + deal-shape taxonomyLive capacity state (PSA + HR + ATS)Forward-commitment graphStatic-floor reference (current pricing template)
● ready● cleanup● build
Cell 04 · proposal × cost MECHANISM Scene 07 of 21 · Cost-to-serve estimation
Cell 04 · proposal phase · cost dimension

What will this actually cost to deliver?

Today: estimation by template + memory. Architects bid on a role-mix that's 5-10 points off what comparable engagements actually delivered at — baking in margin compression before the deal is signed. The engine puts the corpus inside the estimation tool.

Coordinates
Proposal × Cost
CELL 04
+ The lever · what the engine does
The estimation template runs an inline corpus query at the moment of bid construction: comparable past engagements (same scope shape, geography, customer profile, technology stack), with median role-mix, range, sample size, and the actuals delivered. When the architect's draft estimate is meaningfully off from the corpus pattern, the gap is quantified in dollars. Risk premium calibrated against historical variance for engagements with this profile — not a flat 10%.
Frontline experience
In the operator's tool
Is my estimate really right, or am I working from memory?
PSA · estimate — opportunity O-7218
EngagementCloud migration · 14-week
Your role mix18% architect · 60% senior · 22% jr
Risk premium10% (template default)
Corpus query · 14 comparable engagements
Comparable delivered at 26% architect / 52% senior / 22% jr
Variance band · risk premium fitted at 14% for this profile
"Your mix likely under-estimates by $266k. Adjust to corpus, or document why this engagement is structurally different."
Apply corpus Override Document why
Leadership experience
In the executive's view
"Is our estimation discipline closing the quoted-vs-delivered gap?"

Practice lead sees the firmwide estimation accuracy trend: quoted-vs-delivered margin gap closing or widening, by deal shape, by architect, by customer segment. Outlier architects identified. Outlier deal-shapes flagged for template recalibration. The corpus is no longer a passive archive — it's a live calibration loop, with each closed engagement updating next quarter's estimates.
Primary KPI
Estimation accuracy
Delivered margin variance vs. quoted margin, measured per engagement, trended over time. Sub-decomposition: variance attributable to role-mix drift, scope leakage, ramp inefficiency, risk under-pricing. The decomposition itself is auditable evidence of which mechanism is improving.
Attribution Engine estimation log preserves each recommendation + override. PSA records actuals. Per-engagement variance attribution mechanical from logs. Counterfactual: static-template estimates reconstructable from historical patterns.
Data needed
PSA engagement history with role-level actualsEngagement metadata (scope, geo, tech stack)Loaded role-rate cards by geographyOutcome attribution (why variance occurred)Risk-variance corpus by engagement profile
● ready● cleanup● build
Cell 05 · delivery × revenue MECHANISM Scene 08 of 21 · Booked-to-recognized control
Cell 05 · delivery phase · revenue dimension

Will the booking actually recognize?

Bookings aren't revenue. Engagements get paused, scope reduces, customers churn mid-delivery, recognition slips. Today these surface in finance reports after the fact. The engine builds a daily recognition trajectory from active engagement reality — control loop and forecast in one signal.

Coordinates
Delivery × Revenue
CELL 05
+ The lever · what the engine does
Two interlocked sub-mechanisms. Control loop: engagements at risk of not fully delivering get flagged early through pattern signals — scope-reduction conversations, customer-pause indicators, sponsor-stability signals, ticket-sentiment trends. Forecast: recognition timing predicted per engagement from actual delivery trajectory rather than original SOW timeline. Daily recognition forecast is the rollup. Quarterly forecast becomes defensible because every component traces to a specific engagement state.
Frontline experience
In the operator's tool
Which of my engagements need attention this week?
PSA · delivery manager · portfolio view
Active engagements23 in delivery
Booked revenue$48.2M committed
At-risk this week3 flagged
Recognition risk · flagged engagements
E-4912 Globex · sponsor changed last week, ticket volume −40%
E-5034 Initech · customer email referenced "tightening scope"
E-5108 Acme · recognition trajectory 30 days behind plan
"Proactive intervention candidates. $1.4M of recognition at risk if no action this week."
Open E-4912 See all 3 Acknowledge
Leadership experience
In the executive's view
"What's our forecast actually saying, and why?"

CFO receives the daily recognition forecast built from active engagement trajectories. Forecast moves day-to-day with attribution: which engagements moved the forecast vs. yesterday and why. Quarterly forecast is no longer a sales-leadership judgment call — it's an artifact built from active reality, traceable to specific engagements, defensible to the board. Variance commentary writes itself.
Primary KPI
Recognition realization rate + forecast accuracy
(a) Recognized revenue as % of booked revenue, by cohort — trended over time. (b) Forecast accuracy: how close the engine's daily recognition forecast was to actual quarterly recognition. Both auditable against finance + PSA truth.
Attribution Daily forecast snapshots logged. Engagement-level recognition events logged. Per-engagement attribution of forecast contribution and realization. Pre-engine forecast as pre-period baseline.
Data needed
Active engagement state from PSACustomer-health telemetry (CSAT, ticket trends)Sponsor & org-stability signalsFinance recognition eventsScope-change requests & pauses log
● ready● cleanup● build
Cell 06 · delivery × cost MECHANISM Scene 09 of 21 · In-flight margin steering
Cell 06 · delivery phase · cost dimension

Catch the drift at week 14, not week 26.

Today: project margin gets discovered at quarter-end variance review. By then the trajectory is locked. The engine produces a daily predicted-delivered-margin per engagement, with drift drivers attributed and intervention candidates surfaced — while intervention still works.

Coordinates
Delivery × Cost
CELL 06
+ The lever · what the engine does
The engine maintains a daily predicted-final-margin per active engagement by joining time-tracking actuals against the engaged plan. When the trajectory drifts (senior hours up, junior hours down, scope expanding outside the engaged SoW), the drivers are attributed and surfaced to the PM as intervention candidates: file a change request, escalate staffing, or accept with documented rationale. Three weeks of drift becomes a decision moment, not a discovery.
Frontline experience
In the operator's tool
Is this engagement on track financially, or am I about to find out it isn't?
PSA · PM workspace · engagement E-7218
EngagementAcme cloud migration
Week14 of 28
Engaged margin16.4%
Predicted final12.4%
Margin trajectory · drift detected
Trajectory: declining 0.3 pts/week
Top driver: senior architect hours +22% vs. plan
Scope-change candidates: 3 customer activities outside engaged SoW
"Three weeks of drift becomes a decision moment, not a discovery. File change request, escalate staffing, or accept with rationale."
File CR Escalate Accept & document
Leadership experience
In the executive's view
"Where in the portfolio is margin actually compressing?"

Practice lead sees the engagement portfolio with margin trajectories: which engagements are on plan, which are drifting, which need escalation. Drift patterns surface across engagements (e.g., "senior-architect substitution rising in cloud practice"). Practice margin steered during the quarter — through intervention on specific engagements, not by tightening templates next quarter.
Primary KPI
Margin trajectory accuracy + intervention rate
(a) Accuracy of engine's daily predicted-delivered-margin vs. actual-delivered-margin at engagement close. (b) Intervention rate: % of engagements with engine-flagged drift that had documented PM intervention — and margin recovery on intervened-vs-not.
Attribution Daily prediction snapshots logged. Final delivered margin recorded. Intervention events logged with reason. Counterfactual: comparable un-flagged or un-intervened engagements provide control population.
Data needed
Time-tracking by role per engagement (daily)Activity logs vs engaged SoWEngagement plan baseline (role-mix, milestones)Corpus of past engagement trajectoriesChange-request log + reason codes
● ready● cleanup● build
The substrate · one closed loop MECHANISM Scene 10 of 21 · The loop
Six surfaces. One substrate. Continuous calibration.

Outcomes from each phase calibrate the next quarter's decisions.

The cells aren't independent interventions. Delivery outcomes feed the corpus that calibrates next quarter's lead-scoring and proposal estimation. Lead-stage signals feed proposal-stage pricing context. Proposal-stage decisions set the baseline for delivery-stage drift detection. The engine learns continuously because the substrate is shared.

PHASE 01 Lead PHASE 02 Proposal PHASE 03 Delivery CELL 01 Demand-shape signal CELL 03 Capacity-aware pricing CELL 05 Booked-to-recognized CELL 02 Pursuit allocation CELL 04 Cost-to-serve estimation CELL 06 In-flight margin steering CORPUS · CONTINUOUS CALIBRATION Outcomes from delivery feed back into lead scoring & proposal estimation
Feedback path 01
Delivery actuals → proposal estimation
Every closed engagement recalibrates the corpus that informs the next quarter's role-mix benchmarks, risk-variance models, and ramp-cost expectations. The estimation accuracy KPI improves as a structural property of the system.
Feedback path 02
Delivery health → lead scoring
Customer-health trajectories from delivery feed back into the lead-stage signal model: which signals predicted expansion, which predicted churn, which predicted scope reduction. Demand prediction gets sharper each cycle.
Feedback path 03
Bid outcomes → pricing calibration
Win/loss outcomes at each price point calibrate the price-vs-win-probability curve per customer profile. The pricing function isn't static — it's a continuously fit model with every closed deal as new evidence.
Six cells. Three phases. One substrate, learning continuously. The engine doesn't just intervene at decision moments — it gets better at intervening with every cycle.
Attribution architecture · the auditable chain MECHANISM Scene 11 of 21 · The KPI tree
From cell-level KPIs to financial outcomes · the receipts chain

Six primary KPIs. Three intermediates. Two outcomes the CFO already tracks.

The engine's primary KPIs aren't financial — they're decision-quality measurements. They aggregate into intermediate metrics finance and sales already track, which aggregate into services revenue and services gross profit. Every dollar attributed to the engine is traceable down to specific rule firings on specific decisions.

CELL 01 · KPI Demand-signal precision CELL 02 · KPI Pursuit efficiency CELL 03 · KPI Capacity-aware pricing impact CELL 04 · KPI Estimation accuracy CELL 05 · KPI Recognition realization rate CELL 06 · KPI Margin trajectory accuracy PRIMARY KPIs INTERMEDIATE 01 Pipeline quality qualified deal flow INTERMEDIATE 02 Won-deal margin quoted at signing INTERMEDIATE 03 Delivered margin realization recognized + protected in flight INTERMEDIATES FINANCIAL OUTCOME 01 Services Revenue recognized in period FINANCIAL OUTCOME 02 Services Gross Profit recognized revenue minus delivered cost OUTCOMES
Every recovered dollar is traceable down the tree — from financial outcome to intermediate to specific KPI to specific rule firing on a specific decision. This is what high-rigor attribution looks like as a continuous artifact, not a quarterly review.
The workflow in motion · three service lines, three rhythms APPLICATION Scene 12 of 21 · Daily operations
Deal flow per service line + the UI the operator sees at the moment of decision

Three service lines have three different rhythms. The engine intervenes inside each.

Project services runs linear engagement cycles. Managed services runs a recurring renewal cycle. Staff augmentation runs a fast iterative loop. Same engine, three timeshapes — surfacing evidence in the operator's existing tool at the moment of decision.

Service line 01 Project services linear · 60-180 day sales cycle · 3-12 month delivery
Deal flow
LEAD qualify in PROPOSAL price + scope DELIVERY execute CLOSE recognize CELL 02 CELL 03 CELL 04 CELL 05 CELL 06 UI BELOW ↓
Architect's bid pricing screen · Cell 03 firing at the proposal moment
PSA · bid — opportunity O-7218
CustomerAcme Health · cloud migration
Scope$4.8M · 14-week delivery
Role mix26% architect · 52% senior · 22% jr
Static floor26% margin
Quoted26% · $1.25M GP
Engine recommendation
Senior architect util 91% · bench thin in cloud
"Recommend floor 28% (+2pts). Win rate at this floor: 64%. Foregone margin if held at 26%: $96k."
Accept Override Escalate
Service line 02 Managed services cyclical · continuous SLA · renewals every 12-36 months
Renewal cycle
WATCH telemetry TRIGGER renewal due SCOPE-UP audit + propose NEW TERM deliver SLA CELL 01 CELL 05 recurring UI BELOW ↓
CSM's renewal workspace · Cell 05 scope-true-up audit
Renewal workspace · customer C-2204
CustomerNorthwind · healthcare vertical
Current MRR$48k · signed Q1 2023
Renewal duein 6 weeks
Defaultrenew at $48k + CPI
Scope-true-up audit
Endpoints +38% · cloud workloads +62% · ticket volume +44%
"Cost-to-serve grew $17k/mo since signing. Recommended scope-true MRR: $65k. Recoverable: $209k annualized, defensibility brief attached."
Open brief Adjust Defer
Service line 03 Staff augmentation fast loop · days-to-weeks · extension every 30-90 days
Placement loop
REQ IN intake SCAN match bench PRICE pay-rate × bill PLACE on assignment EXTEND 30-90 days re-match next req CELL 06 CELL 03
Recruiter's req intake screen · Cell 06 bench-match firing
ATS · req Q-8412 intake
ClientFinServ Corp · Sr. Cloud Engineer
SkillsAWS · Terraform · Python · SOC2
Bill rate$155/hr target · 6-month
Defaultpost to external sourcing → 3-4 wks
Bench-match scan
Internal candidate found · 91% skill match · rolling off in 8 days
"Marisol K. clears bench cost in 1-2 weeks vs. 3-4 weeks external. Margin advantage on this placement: $59k."
Open profile Compare Skip
Three different deal shapes. Three different operator workflows. One engine, three intervention rhythmsthe architect sees their bid floor, the CSM sees their renewal audit, the recruiter sees their bench match. Inline, in the tool they already use.
Service-line deep-dive · project services APPLICATION Scene 13 of 21 · Project services walkthrough
Service line 01 · from qualify to close-out, end-to-end

Project services: how a deal moves — and where the engine intervenes.

The highest-dispersion service line. Four dominant cells: estimation accuracy and capacity-aware pricing at the front, in-flight steering and demand-shape signals at the back. Multiple intervention points, multiple compounding margin protections.

The shape $100M revenue 20% margin $20M GP
Project services lifecycle · 8 stages · cells firing
STAGE 01 Qualify STAGE 02 Discovery STAGE 03 Estimate STAGE 04 Price STAGE 05 Sign & mobilize STAGE 06 Deliver STAGE 07 Close-out STAGE 08 Phase-2 watch CELL 02 CELL 04 CELL 03 CELL 06 CELL 05 CELL 01 pursue decision corpus query capacity floor daily margin + risk signals expansion signals
Today · how it runs Five points where margin leaks before anyone notices
  • Estimation by template + memory. Architects bid on role-mix 5-10 pts off what comparable engagements actually delivered — margin compression baked in at bid stage.
  • Risk premium is a flat 10%. Aggressive timelines and first-time customers systematically under-priced; low-risk engagements over-priced and lost.
  • Pricing is capacity-blind. Same margin floor regardless of whether seniors are 91% utilized or sitting on bench. Capacity tightness doesn't translate to rate discipline.
  • In-flight margin invisible until quarter-close. By the time variance is discovered, the trajectory is locked. Recovery costs more than prevention would have.
  • Phase-2 follow-on missed. No tracking of expansion signals at close-out. Customer relationships warm enough to pitch are left to cold outreach later.
With NSAI · what changes Four cells firing across the lifecycle, each protecting a specific margin pocket
  • CELL 04Estimate stage. Inline corpus query at the moment of bid construction: comparable engagements, median role mix, variance band. Quantifies the gap in dollars.
  • CELL 03Price stage. Margin floor calibrated against current utilization, bench state, forward commitments. Capacity-aware function, not static template.
  • CELL 06Delivery stage. Daily predicted-final-margin trajectory. Drift drivers attributed. Scope-change candidates surfaced for change-request conversion at week 14, not week 26.
  • CELL 05Delivery stage. Customer-health signals: sponsor changes, ticket sentiment, recognition trajectory. Engagement-risk surfaces before recovery is needed.
  • CELL 01Close-out / Phase-2. Expansion signals tracked at engagement close. Successful engagements re-enter the demand-shape signal model.
Service-line deep-dive · managed services APPLICATION Scene 14 of 21 · Managed services walkthrough
Service line 02 · continuous SLA delivery + cyclical renewal

Managed services: the recurring-revenue cycle — protected and grown.

The largest absolute pocket. Cell 01 and Cell 05 dominate: continuous demand-shape signal + scope-true-up audit at renewal. Recurring revenue compounds — once captured, rebased MRR persists for the next term.

The shape $300M revenue 30% margin $90M GP
Managed services cycle · continuous + cyclical · cells firing
CONTINUOUS · WATCH (CELL 01) · environment growth, sponsor changes, ticket sentiment STAGE 01 Renewal trigger STAGE 02 Scope-true-up audit STAGE 03 Renewal negotiation STAGE 04 New term begins STAGE 05 Steady-state SLA STAGE 06 Health monitoring CELL 05 CELL 05 CELL 03 CELL 06 CELL 05 operational audit + brief defensible MRR · capacity-aware cost-to-serve telemetry customer-health signals
Today · how it runs Five places where recurring margin gets quietly eroded
  • Renewals default to "original + CPI." Customer environment growth invisible at renewal moment. 3-year-old MRR carrying 3 years of expanded scope at original pricing.
  • RMM/ITSM/observability data exists but isn't connected to the renewal workspace. CSM negotiates on relationship, not on operational evidence the customer's own dashboards confirm.
  • Cost-to-serve creep tracked at quarterly margin review. Engineer-tier escalations, tooling pass-through inflation discovered after the next renewal has already happened.
  • Customer-health signals noticed reactively. Sponsor changes, ticket sentiment, escalation patterns surface as problems — not as opportunities for proactive intervention. Mid-term scope expansions absorbed without economic impact analysis.
With NSAI · what changes Two dominant cells running on a recurring rhythm, compounding value over multiple terms
  • CELL 01Continuous watch. Account-level signal model tracks environment growth, sponsor changes, escalation patterns. Renewal-window detection 6 months out.
  • CELL 05Renewal trigger + audit. Scope-true-up audit pre-attached: endpoints +X%, workloads +Y%, complexity +Z%. Defensibility brief anchored on the customer's own operational data.
  • CELL 03Renewal negotiation. Recommended scope-true MRR with capacity-aware floor. CSM still negotiates — but from a calibrated starting point, not a static template.
  • CELL 06CELL 05Steady-state delivery + health monitoring. Cost-to-serve telemetry (tier-mix drift, complexity, pass-through) + customer-health leading indicators. Drift caught mid-term; pause/churn intercepted before recognition erodes.
Service-line deep-dive · staff augmentation APPLICATION Scene 15 of 21 · Staff augmentation walkthrough
Service line 03 · fast iterative placement loop, high-volume

Staff augmentation: thin margins, high transaction volume, many small calibration events.

Smallest pocket as a percentage. Two dominant cells: bench-match scanning + capacity-aware bill-rate discipline. Per-placement margins are structurally thinner; the engine produces value through frequency, not magnitude.

The shape $200M revenue 15% margin $30M GP
Staff aug placement loop · 7 stages · cells firing
STAGE 01 Req intake STAGE 02 Bench scan STAGE 03 Match decision STAGE 04 Pay × bill rate STAGE 05 Place STAGE 06 On-assignment STAGE 07 Extension ↻ CELL 06 CELL 02 CELL 03 CELL 05 CELL 06 internal-match scan cost-vs-time tradeoff rate distribution + floor engagement health 30-day signal · re-match
Today · how it runs Five places where small leaks compound across high transaction volume
  • Requisitions go straight to external sourcing. Bench register exists in ATS but isn't scanned at intake. Internal candidates with strong matches sit on bench while external recruiters work 3-4 weeks.
  • Pay-rate negotiations happen on feel. No visibility into where the offer sits on current market distribution for that skill profile. Inconsistent rates across the same skill, same geography.
  • Bill-rate floors are static. Same minimum regardless of capacity tightness or skill scarcity. Margin discipline doesn't tighten when consultants are scarce.
  • Extension decisions happen at the last minute, with bench treated as drag. No leading indicators; roll-offs surprise both sides, bench profile mis-matched to forward demand.
With NSAI · what changes High-volume cycle · small interventions, frequent firing, compounding margin discipline
  • CELL 06Bench scan at intake. Auto-scan against bench register at every req. Skill-match score with economic comparison: internal cost vs external sourcing cost.
  • CELL 02Match decision. Internal vs external surfaced with tradeoff math: margin advantage on this placement, time-to-fill, bench cost cleared.
  • CELL 03Pay × bill rate pricing. Pay rate shown against current market distribution; bill rate floor adjusts with capacity state. Rate discipline becomes a function, not a feeling.
  • CELL 05CELL 06On-assignment monitoring + bench steering. Engagement health signals (satisfaction, escalations, scope-fit) predict extension 30+ days out; forward-demand-shaped bench hires into gaps. Bench becomes managed inventory.
The operator's workspace · one screen, six cells, ambient steering APPLICATION Scene 16 of 21 · The operator UI
A senior architect's Monday morning · what the engine looks like as a daily companion

The engine isn't a dashboard. It's ambient guidance in the tools operators already use.

One workspace, four panels, multiple cells firing concurrently. The architect doesn't switch contexts to get engine input — the PSA they already work in surfaces evidence inline. Bid pricing, engagement queue, notifications, decision log: every panel anchored to a specific cell, every recommendation logged with counterfactual.

PSA · senior architect workspace · Mon Q2W3
S. Patel · cloud practice
Active bid · opportunity O-7218 CELL 03 · CELL 04 firing
CustomerAcme Health · cloud migration Scope$4.8M · 14-week delivery Your draft mix18% arch / 60% senior / 22% jr Corpus says26% arch / 52% senior / 22% jr (n=14) Static floor26% margin Quoted26% · $1.25M GP
Engine recommendation · senior util 91%, bench thin
Mix gap: your draft under-estimates by $266k (Cell 04). Floor lift: recommend 28% (+2pts), win rate 64% · foregone margin if held at 26%: $96k (Cell 03).
Apply both Apply mix only Override See math
My engagements · margin trajectory CELL 06 · 7 active
E-7042 Initech replatform 22.8%
E-7218 Acme migration (wk 14) 12.4%
E-7156 Globex platform v2 18.6%
E-7089 Sentinel data lake 28.4%
E-7193 Northwind retainer 31.2%
E-7204 Initrode integration 19.8%
E-7251 Vandelay phase-2 23.1%
Engine recommendations · last 24h 4 unread
CELL 06 · drift detected 07:42
E-7218 Acme: senior architect hours +22% vs plan, predicted final margin 12.4%. Three weeks of drift; intervention window open.
CELL 04 · estimation gap 06:15
O-7301 Initech: corpus query suggests role mix 10pts off from comparable, risk premium under-set at 10%. Recommended adjustment: $187k.
CELL 03 · capacity tight Sun 22:10
Senior architect pool at 91% through Q3. Margin floors lifting +1.5pts on new cloud bids.
CELL 01 · account heating Sun 18:33
Northwind: environment +38%, new VP started 3wks ago, renewal in 6mo. Phase-2 conversation candidate.
Your decisions this week · logged 14 decisions · 79% accept rate
accept O-7245 floor lift to 30% (+2pts) +$58k
accept E-7218 file change request (scope creep) +$144k
override O-7256 floor held at 26% (strategic) −$34k
accept O-7193 mix correction (+2% architect) +$92k
escalate O-7301 floor variance >5pts: practice exec pending
accept E-7156 in-flight intervention +$71k
override E-7204 drift flag dismissed (rationale: customer ask) logged
Each panel anchors to a specific cell — but the architect doesn't have to know that. They see their bid, their engagements, their notifications, their decisions. The engine is the substrate; the operator's workflow is the interface.
The executive view · rolled-up KPIs, attribution receipts, decisions APPLICATION Scene 17 of 21 · The executive UI
The CFO's quarterly steering pack · what leadership sees, with receipts

The executive view is not a different system. It's the rollup with attribution attached.

Six primary KPIs, three intermediates, two outcomes. Every number on the screen traces to specific decisions logged in the operator UI. Confidence intervals are real (from holdout cohorts and counterfactuals), variance commentary writes itself, and override patterns surface where the engine needs recalibration.

Steering pack · firmwide · Q2 close, week 3 review
R. Chen · services GM
Primary KPIs · firmwide rollup · q-to-date vs trailing 4Q baseline Refreshed 06:00 today · CI from 15% holdout cohort
Demand-signal precision 78.4%
+6.2pp
CI ±3.1pp · n=438
Pursuit efficiency 2.1×
+0.8×
CI ±0.2× · n=212
Cap-aware pricing impact +1.8pp
+1.1pp QoQ
CI ±0.4pp · n=174
Estimation accuracy ±4.2pp
gap closes 3.4pp
CI ±0.8pp · n=174
Recognition realization 94.6%
+2.7pp
CI ±0.9pp · n=312
Margin trajectory acc 87.2%
14wk earlier
CI ±2.1pp · n=89
Practice margin trajectories · QTD vs baseline attribution: matched-pair, holdout adjusted
Cloud & data 24.8%
+2.4pp
App modernization 21.6%
+1.7pp
Security & compliance 26.2%
+1.2pp
Managed services 33.1%
+3.4pp
Data engineering 19.2%
−0.6pp
Staff augmentation 16.1%
+1.4pp
AI & analytics (new) 18.8%
baseline
Override patterns · needs review 3 flagged
Cell 03 · cloud bids 38% override
Sustained 38% override on cloud bids floor lifts. Pattern: architects holding at 26% despite engine recommendation 28+%. Likely model under-fit for emerging customer segment.
Cell 04 · security practice 26% override
Estimation gap dismissed on security engagements. Architects citing "compliance complexity not captured in corpus." Recommend corpus enrichment with compliance tagging.
Cell 06 · staff aug 31% override
Drift flags dismissed for extensions. Recruiters citing "client relationship not measured." Investigate; staff aug margin variance still tightening.
Pending decisions · you 4 actions
Cell expansion EOW
Approve Cell 05 (recognition control) rollout to managed services. Cell 01 ramp at 88% adoption; cell ready. Estimated lift on $300M book: +$4-7M Q3.
Rule recalibration 10 days
Approve cloud-segment floor model retrain. Override pattern (38%) above 25% threshold. Retrain on Q1-Q2 outcomes; expected to bring override to ~15%.
Attribution report Q-close
Sign quarterly attribution report. Q2 engine contribution: +$4.8M GP firmwide (CI $4.1-5.4M). Board version drafted. Audit trail attached.
Same data, different audience. The architect sees decisions; the executive sees rollups with confidence intervals. Both views are continuous artifacts of the same per-decision logging substrate — one screen at a time, one quarter at a time.
Where each cell matters most · service-line impact APPLICATION Scene 13 of 21 · Service-line relevance
Six cells × three service lines · deployment sequencing logic

Each cell isn't equally valuable everywhere. Where it lands hardest determines where to deploy first.

Project services, managed services, and staff augmentation have different economic shapes — different cells dominate in each motion. The matrix below shows which cells are dominant, important, or minor for each service line. The deployment sequence follows from where the heavy levers concentrate.

Cell ↓ · Service line →
Service line 01Project services
Service line 02Managed services
Service line 03Staff augmentation
Cell 01 Demand-shape signal lead × revenue
Important
Account expansion + Phase-2 pursuits from delivered engagements. Less about cold demand, more about follow-on signal from existing customers.
Dominant
Renewal-window detection + scope-growth signals. The single most valuable lead-stage signal in any IT services book. Drives every renewal conversation.
Minor
Requisition prediction from existing client environments. Useful but staff aug demand tends to be reactive, not predictive.
Cell 02 Pursuit allocation lead × cost
Dominant
Pre-sales architects are scarce. Pursuit cost compounds; opportunity cost across pursuits is real. Clearest portfolio-net contribution math.
Important
New-MSA pursuits consume pre-sales heavily. Renewal pursuits are lower-effort but volume matters. Mix-management value.
Minor
Recruiter time is a real cost but pursuit dynamics differ. Bench-match scanning (cell 06) carries more weight here.
Cell 03 Capacity-aware pricing proposal × revenue
Dominant
The largest single lever in project services. Margin floors that flex with capacity = 1-3 points of practice margin in most firms. Highest-leverage cell.
Important
New MSAs benefit from capacity-aware pricing. Renewals dominated by scope-true-up dynamics (cell 05) more than fresh pricing.
Important
Pay-rate × bill-rate spread management. Capacity-aware bill-rate floors materially improve margin on every placement.
Cell 04 Cost-to-serve estimation proposal × cost
Dominant
The estimation-drift mechanism. Corpus-informed role-mix + risk premium calibration = 3-5 points of margin defended at bid stage. The heavy lever here.
Important
Cost-to-serve modeling for new MSA pricing matters significantly. Comparable-customer corpus helps avoid underwriting losses at signing.
Minor
Standard placements have well-known cost structures. Cost-to-serve estimation matters less than pay-rate discipline (cell 03).
Cell 05 Booked-to-recognized control delivery × revenue
Important
Engagement-pause + scope-reduction risk on long migrations. Forecasting matters; daily trajectory provides defensible quarterly forecasts.
Dominant
Customer-health signals + scope-true-up audit. The recurring-revenue defense mechanism. $2-4M annualized recurring on a $30M MS book is on the table here.
Important
Extension-vs-rollout dynamics. Consultant-engagement health predicts extension probability. Forecasting placements 30-60 days out.
Cell 06 In-flight margin steering delivery × cost
Dominant
Catch the drift at week 14. Daily margin trajectory + scope-leakage candidates = 2-4 points of margin recovered + change-request capture. Heavy lever in delivery.
Important
Cost-to-serve creep monitoring. Engineer-tier escalations + tooling pass-through visibility. Material but slower-moving than project services.
Dominant
Bench-match + opportunity-cost visibility. The dominant operational pattern in staff aug. Pay-rate distribution + bench scanning at requisition intake.
Deploy first
Managed services
Largest pocket, recurring. Cells 01, 05 dominate; cell 03 supports new-MSA work. Single-mechanism wins ramp over 12-24 months as renewals come up.
Deploy second
Project services
Highest dispersion, multiple levers. Cells 02, 03, 04, 06 all dominant. Sequenced rollout: estimation discipline first, then in-flight steering.
Deploy third
Staff augmentation
Smaller pocket, operational mechanisms. Cells 03, 06 dominant. Important for portfolio capacity visibility; weakest standalone case for the engine.
The bottom line · estimated impact by service line APPLICATION Scene 14 of 21 · What this is worth
From cell-level levers to GP uplift · what we estimate, calibrated against industry benchmarks

Three service lines. Different cells dominate. Different bottom-line shape.

The six cells don't move every service line equally. Where the heavy levers concentrate determines the addressable pocket. What the engine actually captures is bounded by adoption velocity, data discipline, and customer-acceptance dynamics — so we frame realistic capture (50%) and mature steady state (65%) separately.

Firm shape
Project services $100m · 20% margin · $20m GP | Staff augmentation $200m · 15% margin · $30m GP | Managed services $300m · 30% margin · $90m GP | Total $600m revenue · 23.3% blended · $140m GP
Service line 01
Project services
$100m revenue × 20% margin = $20m GP
Dominant Cell 02 Cell 03 Cell 04 Cell 06
Addressable pocket
5–9 margin points · $5–9m GP
Estimation accuracy (3-5 pts) + in-flight steering (2-4 pts) + capacity-aware pricing (1-3 pts), net of overlap between cells.
Realistic capture 50% · deployment plan
+$2.5–4.5m GP
Margin moves 20% → 22.5–24.5%
Mature steady state 65% · long-term
+$3.3–5.9m GP
Margin moves 20% → 23.3–25.9%
What to know
The highest-dispersion service line — multiple cells dominant. Sequenced rollout: estimation discipline first, then in-flight steering, then capacity-aware pricing as bid templates calibrate.
Service line 02
Staff augmentation
$200m revenue × 15% margin = $30m GP
Dominant Cell 03 Cell 06 + Cell 05
Addressable pocket
1.5–3 margin points · $3–6m GP
Bill-rate floor discipline + bench-match optimization + pay-rate distribution inline. Smaller per-placement physics, but high transaction volume.
Realistic capture 50% · deployment plan
+$1.5–3m GP
Margin moves 15% → 15.75–16.5%
Mature steady state 65% · long-term
+$2–3.9m GP
Margin moves 15% → 16–16.95%
What to know
Smallest pocket as a percentage — staff aug margins are structurally thinner. Operationally clean to deploy: high transaction volume gives the engine many small calibration events per day, not per quarter.
Service line 03
Managed services
$300m revenue × 30% margin = $90m GP
Dominant Cell 01 Cell 05 + Cells 03, 06
Addressable pocket
8–15% of recurring · $20–38m GP
Scope-true-up audits + cost-to-serve creep monitoring on the recurring portion ($250m), as renewals come up over 12-24 months.
Realistic capture 50% · at full ramp
+$10–19m GP
Margin moves 30% → 33.3–36.3%
Mature steady state 65% · long-term
+$13–25m GP
Margin moves 30% → 34.3–38.3%
What to know
Largest absolute pocket — recurring revenue compounds. Most value lands on a 12-24 month delay as the renewal book rolls through. Once captured, the rebased MRR persists for the next term.
Firmwide
Estimated GP uplift
Realistic (50%) +$14–26m Annualized. +1.0 to +1.9 pts blended margin — reaches this level over 18-24 months as managed services renewal book rolls through.
Mature (65%) +$18–35m Annualized at steady state. +1.3 to +2.5 pts blended margin. Engine fully calibrated; corpus mature; operator adoption reflexive.
Anchored on
SPI Research 2026 PS Maturity Benchmark · HPO estimating accuracy +15%, change control +13%, resource management +14% | Service Leadership Index (ConnectWise) · top-quartile service GM 42.1% vs median ~28% | SPI 2024 PSA gap data · PSA users vs non-users: utilization +10%, project margin +24%, EBITDA +28%
The P&L view · engine impact, finance language APPLICATION Scene 15 of 21 · The P&L view
From cell-level levers to specific income-statement lines

Where the engine shows up on the P&L finance already tracks.

The engine doesn't change the P&L's structure. It changes the lines — at realistic capture, on the firm's $600M services book. Each impacted line traces back to specific cells, which trace back to specific operator decisions.

Services P&L · current vs. realistic capture
Services revenue $600.0M +$2–5M
  · recognized in period $600.0M +$2–5M
COGS (delivered services cost) $460.0M −$12–21M
  · direct labor $365.2M −$8–14M
  · bench / unutilized $48.4M −$3–5M
  · tooling pass-through $46.4M −$1–2M
Services gross profit $140.0M (23.3%) +$14–26M
SG&A — pursuit cost $36.0M −$2–4M
Services contribution $104.0M +$16–30M
Forecast accuracy (variance) ±6–9% ±2–3%
DSO / working capital −3–7 days
How each line moves · cells responsible
Revenue
recognized
CELL 01CELL 05
Demand-shape signals drive renewal scope-ups + expansion proposals; recognition control protects booked revenue from scope-reduction and customer-pause leakage.
Direct labor
CELL 04CELL 06
Estimation accuracy reduces role-mix drift at bid stage; in-flight steering catches drift at week 14, not week 26.
Bench /
unutilized
CELL 02CELL 06
Bench-match at requisition intake, pursuit prioritization against capacity. Bench becomes managed inventory, not residual cost.
Gross profit
(realized)
CELL 03CELL 04CELL 06
Capacity-aware pricing lifts floors when capacity is scarce; estimation + in-flight steering protect quoted margin through delivery.
Pursuit cost
(SG&A)
CELL 02
Pursuit allocation against portfolio-net contribution: fewer wasted pre-sales hours on portfolio-negative pursuits.
Forecast
accuracy
CELL 05
Recognition trajectory built from active engagement reality, not SOW-timeline aspiration. Variance commentary writes itself.
Working
capital
CELL 05
Earlier detection of at-risk recognition + faster scope-true-up = revenue recognized closer to delivered. DSO compresses; cash arrives sooner.
The engine doesn't show up as a new line. It shows up across the lines that already existsame statement, different shape, with attribution down to the decisions that produced each move.
What to ask us next · conversation pathways CLOSE Scene 16 of 21 · What to ask us next
Four conversations, four levels of commitment, four kinds of evidence

Wherever you are on this argument, here's the next conversation.

No single next step. Four pathways — from internal pressure-test to engineering deep-dive to deployment. Pick the one that matches what you need to learn next.

Pathway 01 The GP First Instance
10 weeksresale-first scopeyour data
The engine running on your actual data, against your actual baseline, with rule firing rates measured. Produces your firm's specific capture rate — not estimates from comparable firms. Includes a data-readiness assessment for services rollout. Output: your number, with receipts · calibrated deployment plan
Pathway 02 Service-line prioritization workshop
Half-dayoperating leadership+ pre-read
Working session with practice leadership, finance, and CRO to commit to the deployment sequence — managed first, then project, then staff aug. Walks firm-specific impact estimates and confirms cells matching your profile. Output: deployment-sequence decision · quarterly milestones
Pathway 03 Data-readiness assessment
2 weeksyour systemsstructured walk
Focused engagement with IT and operations to map your PSA, CRM, RMM, ITSM, ATS against each cell's data needs. Produces a readiness scorecard (ready / cleanup / build) and integration sequencing plan anchoring deployment timeline. Output: per-cell data-readiness scorecard · integration timeline
Pathway 04 Architecture deep-dive
3 hoursengineering + opstechnical audience
For technical leaders who need to pressure-test the substrate — ontology, rules, reasoning, governance, attribution mechanics. Walks the attribution architecture, holdout-cohort design, and per-decision logging producing audit-grade receipts. Output: technical confidence · integration architecture sketch
Different audiences need different evidence. The deck makes the case; the next conversation calibrates it to your firm. Tell us which pathway is the right next step.
Appendix · the questions you'll ask anyway APPENDIX Scene A1 · Implementation timeline
Appendix 1 · implementation timeline

What does the first 90 days actually look like?

Three phases · foundation, deployment, expansion. The GP First Instance covers the first 10 weeks — everything after ramps based on what the first instance reveals about data readiness and adoption.

Days 1-30 · foundation Data integration + corpus build Engine connects to PSA, CRM, RMM, ITSM, ATS. Historical engagement data ingested and normalized into a queryable corpus. Skill taxonomy reconciliation typically the largest data-cleanup task.
  • System integrations (5-7 sources)
  • Corpus build from 18-24 months
  • Skill taxonomy work — often the gating item
  • Baseline measurements established
Days 31-60 · deployment First cells live + rule calibration Cells deploy in priority order. Resale + managed-services renewal-window detection (Cell 01) typically first; rule firing rates monitored against operator response patterns.
  • 2-3 cells live in initial scope
  • Operator training (inline, in-tool)
  • Rule firing rates calibrated
  • Per-decision logging operational
Days 61-90 · expansion Adoption tightens + first attribution report Operator adoption deepens; corpus enriches with logged decisions. First quarterly attribution report produced — initial KPI movements, capture rate, next-cell expansion plan.
  • Quarterly attribution report (audit-grade)
  • Capture rate observed — not estimated
  • Next-cell deployment committed
  • GP First Instance complete
See also
Pathway 03 (data-readiness assessment) compresses days 1-30 work into 2-week pre-deployment scoping. Most timeline risk lives in skill-taxonomy reconciliation — assess this early.
Appendix · the questions you'll ask anyway APPENDIX Scene A2 · Change management
Appendix 2 · change management

How do operators actually adopt this?

The engine is inline, in the tools they already use. No new system to log into, no separate dashboard to remember to check. Adoption is a function of relevance and trust, not training.

Design principle 01 No new tool · no new login The engine appears as panels inside the operator's existing PSA, CRM, ATS. Architects see capacity-aware floors in their bid template. CSMs see scope-true-up audits in their renewal workspace. Operators don't change tools; the tools get smarter.
  • No separate dashboard
  • No daily standalone check-in
  • Surface where decisions already happen
Design principle 02 Override is always available Every recommendation has an Override option with a reason field. Operators retain authority; the engine surfaces evidence. Override patterns themselves become signals — if architects systematically override down 3+ points, the floor needs recalibration.
  • Operator authority preserved
  • Override reason captured
  • Override patterns feed rule calibration
Adoption pattern Early skeptics become strongest advocates Senior architects with strong intuition find corpus confirms what they'd have done — lowest resistance. Junior practitioners catch errors they would have made unaided — highest direct value. Resistance comes from middle-tier operators with strong opinions and weak data; this is where leadership cover matters most.
See also
Scenes 04-09 show each cell's UI mock — this is what "inline in the operator's tool" actually means. Scene 12 shows the same pattern instantiated for three service lines.
Appendix · the questions you'll ask anyway APPENDIX Scene A3 · Differentiation
Appendix 3 · how this differs from what you have

How is this different from our BI tools, our PSA, our existing AI?

Different in three specific ways: where it sits in the workflow, what it operates on, and what kind of attribution it produces. NSAI is not a replacement for any of these; it's a layer that connects them.

vs. BI / dashboards BI shows past; engine shapes present BI dashboards are read tools — the executive interprets them after the fact. The engine surfaces evidence at the moment of decision — the architect sees the floor when pricing, not in next week's review. BI tells you what happened. The engine influences what happens next.
vs. PSA / CRM / ATS PSA records; engine reasons PSA captures what was estimated, scoped, billed. The engine queries that history at the moment a new estimate is being built, applies rules calibrated against actuals, and surfaces the corpus-informed recommendation inline. The PSA is the system of record; the engine is the system of reasoning on top.
vs. generative AI / LLMs LLMs generate; NSAI attributes Generative AI produces plausible answers; the reasoning isn't auditable. NSAI's recommendations trace to specific rules firing on specific evidence, with corpus subset, calibration, and counterfactual all logged. Audit-grade attribution requires a substrate LLMs structurally cannot provide.
See also
Scene 11 (the KPI tree) shows the attribution architecture concretely — this is what "audit-grade" produces, on a continuous basis, automatically.
Appendix · the questions you'll ask anyway APPENDIX Scene A4 · Data privacy & governance
Appendix 4 · data privacy & governance

Who sees what, and where does the data live?

All operational data stays inside your firm's data perimeter. The engine deploys as a layer on top of your existing systems, not as an external service that exfiltrates data. Customer data, operator decisions, financial actuals: all in your environment.

Data residency Your environment, your perimeter The corpus, rule library, and reasoning engine deploy inside the firm's existing data perimeter — on-prem, private cloud, or your VPC in a public cloud. Customer data does not leave your environment.
  • On-prem · private cloud · your VPC
  • No data exfiltration to vendor
  • Same perimeter as PSA / CRM / RMM
Access controls Inherits your existing identity and roles The engine respects existing access controls of host systems. An architect sees recommendations only on opportunities they have access to; a CSM sees scope-true-up audits only for their accounts; a practice exec sees rolled-up steering for their practice.
  • SSO via your IdP
  • Role-based access mirrored from PSA / CRM
  • Audit log of every recommendation surfaced
Customer-data handling Telemetry, not customer-private content For managed-services scope-true-up: the engine reads operational telemetry from RMM, ITSM, observability — endpoint counts, ticket volumes, complexity. It does not read customer-private content (emails, files, application data).
  • Telemetry only, not content
  • Customer-side governance unchanged
  • Standard data-processing agreements apply
See also
Pre-deployment governance review folds into Pathway 03 (data-readiness assessment). Most enterprise security teams have approved equivalent architectures — patterns are well-understood.
Appendix · the questions you'll ask anyway APPENDIX Scene A5 · Internal ownership
Appendix 5 · who owns this internally

Who owns this on our side — and how heavily?

Three roles matter: an executive sponsor, an operations lead, and a technical lead. The engine is operationally light to run once deployed; the work concentrates at deployment and at quarterly recalibration.

Role 01 · executive sponsor Practice exec, BU GM, or COO Senior executive providing cover for the deployment and steering at quarterly reviews. Effort: ~2 hours/month at steady state; more during initial deployment for cell-by-cell prioritization.
  • Cell deployment sequencing
  • Override pattern review (quarterly)
  • Attribution report sign-off
  • ~2 hrs/month at steady state
Role 02 · operations lead Services ops director or practice ops lead Operational owner driving day-to-day adoption and feedback. Engages with operator overrides, surfaces friction, coordinates with domain experts. Effort: ~5-10 hours/week at steady state; full-time during initial deployment.
  • Operator adoption coaching
  • Override pattern triage
  • Domain-expert coordination
  • ~5-10 hrs/week at steady state
Role 03 · technical lead Data engineering lead or BI architect Technical owner managing the integration surface — PSA, CRM, RMM, ATS connections; corpus refresh; data quality monitoring. Effort: ~1-2 days/month at steady state; substantially more during deployment.
  • Integration health monitoring
  • Corpus refresh validation
  • Data-quality issue triage
  • ~1-2 days/month at steady state
See also
During GP First Instance, our team carries most operations and technical work. Internal ownership ramps as the engine moves toward firmwide steady state — typically months 6-12.
Use ← → arrow keys