A client's engagement sentiment has been declining for three weeks. The trend is invisible in any single meeting or message, but the system composed it from Slack tone, meeting transcript patterns, and declining response times. It surfaces the finding, cites the evidence, and routes a recommendation to the relationship owner before the client escalates.
INTELLIGENCE LAYER
The questions you've been asking, answered.
The intelligence layer reads from your operational systems and composes what it finds into a structured, legible picture of what is actually happening across your organization. Leadership gets real visibility, the kind that's usually locked in the heads of the people closest to the work. On top of that visibility, the system generates strategic recommendations and routes them to the people who can act. Every claim traces to a primary source. Every recommendation carries a hypothesis and a way to verify it.
WHAT IT DOES
Composed situational awareness on a fixed cadence.
Most organizations collect operational data and never compose it. Project status lives in one tool, financials in another, sentiment in a third, and the picture that matters lives only in the head of whoever happens to be in the most meetings that week. Leadership reads symptoms and infers cause from memory.
The intelligence layer changes this. It reads from a knowledge graph that connects all of your organization's operational data into a single queryable surface. Sources are pulled on a fixed cadence, normalized, and composed into structured artifacts that name the cause as well as the symptom. The artifacts are versioned so that week-over-week change is itself a first-class signal.
The richer the knowledge graph, the more powerful the intelligence layer becomes. With more sources connected and more relationships mapped, the system can surface patterns and connections that would be impossible with fragmented data. The intelligence layer can start before the full knowledge graph is built, but it reaches its full potential on top of it.
Slack Meeting recordings CRM Project management Time tracking Knowledge bases
SYSTEM OUTPUT
What the output actually looks like.
AE.STUDIO · PROJECT INTELLIGENCE
PORTFOLIO OVERVIEW
Weekly project status
Acme Corp Needs Attention
PLATFORM MODERNIZATION Client: Positive Team: Concerned
Load testing blocker unresolved for 3 weeks. UAT on document pipeline passed at 97.2%.
$342K / $500K
UPDATED MAY 23, 2026 Meridian Health At Risk
DATA PLATFORM Client: Neutral Team: Concerned 2 RISKS
Budget at 88% with 6 weeks remaining. Scope addition in Sprint 11 not reflected in estimate.
$176K / $200K
UPDATED MAY 23, 2026 Atlas Logistics On Track
DOCUMENT INTELLIGENCE Client: Positive Team: Positive
Document pipeline deployed. Client NPS +18 points. USPS modernization RFP creates expansion opportunity.
$135K / $300K
UPDATED MAY 23, 2026 ← All Projects
Acme Corp
Needs Attention Overview Health Check Market Signals Politics Recommendations
STAKEHOLDERS
NameRoleFocus
Jordan ParkTechnical LeadInfrastructure, API layer
Lisa HuangVP ProductRoadmap, UAT sign-off
David ChenCTOArchitecture decisions
AE SIDE
Sarah ChenPMDelivery, client comms
Troy LarssenPDRelationship, commercial
Priya MehtaTech LeadArchitecture, code review
CURRENT ENGAGEMENT
Platform modernization with AI document processing pipeline. Phase 1 (document ingestion) complete and in UAT. Phase 2 (API layer + integrations) in progress. Load testing and July rollout are the critical path items.
BUDGET
$158KRemaining
$22KMonthly rate
7.2Runway (weeks)
MILESTONES
COMPLETEDocument ingestion pipelineApr 18
COMPLETEUAT environment setupMay 2
IN PROGRESSAPI layer v2Jun 6
AT RISKLoad testingJun 20
UPCOMINGJuly rolloutJul 14
WEEKLY TIMELINE
WEEK 14 · MAY 19–23
UAT passed at 97.2% on real data. Load testing raised 3rd consecutive standup. SOW still unsigned.
WEEK 13 · MAY 12–16
Client CEO mentioned board pressure on AI timeline. Document pipeline entered UAT. Budget burn 8% ahead.
WEEK 12 · MAY 5–9
Sprint 12 scope addition: API integration layer not in original estimate. First load testing concern raised.
On Track: No
Load testing concerns raised by client technical lead in three consecutive standups without resolution. Budget at 68% with 10 weeks remaining, tracking ahead due to unscoped integration work. UAT on document pipeline strong at 97.2%.
CLIENT SENTIMENT
Jordan ParkTECHNICAL LEAD
Engaged. Raising infrastructure concerns consistently, which indicates investment in the project's success.
Lisa HuangVP PRODUCT
Described document pipeline as "the first thing that actually works on our messy data."
RISKS
HIGHLoad testing unplanned. July rollout depends on it.Schedule planning session with Jordan Park.
MEDIUMBudget 8% ahead of plan from unscoped Sprint 12 work.Re-estimate remaining sprints.
IMMEDIATE Put out fires
POLITICAL COMPLEXITYYELLOW
Jordan Park is the primary champion and technical gatekeeper. David Chen (CTO) is disengaged but retains veto power.
POWER MAP
Jordan ParkTECHNICAL LEAD
ControlsTechnical sign-off, team adoption
WantsInfrastructure done right
ThreatLOW
David ChenCTO
ControlsArchitecture veto
WantsLow-risk, proven approaches
ThreatMEDIUM
ACT
Zenith Corp announced AI document processing at their annual conference.
Relevance: Validates Acme's investment. Creates urgency to ship before Zenith reaches market.
← All Projects
Meridian Health
At Risk Overview Health Check Politics Recommendations
STAKEHOLDERS
NameRoleFocus
Dr. Rachel OseiChief Data OfficerStrategy, budget approval
Marcus WebbCFOBudget oversight
AE SIDE
Alex RiveraPMDelivery, team management
Troy LarssenPDRelationship, commercial
CURRENT ENGAGEMENT
HIPAA-compliant data platform for patient record processing. Fixed-price SOW. HIPAA compliance layer added mid-project at client request without a change order. Budget pressure is the primary risk.
BUDGET
$24KRemaining
$29KMonthly rate
3.3Runway (weeks)
On Track: No
Budget at 88% with 6 weeks of work remaining. The fixed-price SOW did not anticipate the HIPAA compliance layer added in Sprint 11 at the client's request. The team has been absorbing the scope addition without a change order. Client relationship is stable but the commercial situation is unsustainable without renegotiation.
CLIENT SENTIMENT
Dr. Rachel OseiCHIEF DATA OFFICER
Neutral. Satisfied with delivery quality but aware the compliance addition strained the engagement. Open to discussing scope adjustment.
TEAM SENTIMENT
Alex RiveraPM
Concerned about team morale. Engineers are working unpaid overtime to absorb the compliance scope. Has flagged this in two internal retros.
RISKS
CRITICALBudget exhausts 3 weeks before projected completion. No change order for compliance scope.Renegotiate SOW scope or timeline.
HIGHTeam morale declining. Two senior engineers have mentioned looking for reassignment.Address workload immediately. Acknowledge the scope absorption.
ROOT CAUSE ANALYSIS
Issue: Fixed-price engagement absorbing unscoped compliance work.
Why 1: HIPAA compliance layer was added mid-project at client request.
Why 2: Team said yes without a change order because the client relationship felt fragile.
Why 3: No scope gate protocol. Direction changes absorbed without impact assessment.
Root cause: Fixed-price contract with no mechanism for scope changes creates a dynamic where saying "yes" to the client means saying "absorb it" to the team.
IMMEDIATE Put out fires
POLITICAL COMPLEXITYYELLOW
Dr. Osei is supportive and open to renegotiation. The risk is her CFO (Marcus Webb), who approved the original fixed-price SOW and may view an amendment as a cost overrun rather than a scope change.
RELATIONSHIP MAINTENANCE
Before the renegotiation meeting, send Dr. Osei a summary of what the compliance layer actually delivered. Frame the amendment conversation as a success story (we built more than planned) rather than a failure story (we ran over budget).
← All Projects
Atlas Logistics
On Track Overview Health Check Market Signals Recommendations
STAKEHOLDERS
NameRoleFocus
Maria SantosCOOOperations strategy, expansion
Kevin ChoHead of OperationsDay-to-day, feature requests
AE SIDE
Priya MehtaTech LeadPlatform architecture
Troy LarssenPDRelationship, expansion
CURRENT ENGAGEMENT
Document intelligence platform for logistics operations. Processing 2,400 docs/day at 95.8% accuracy. Core build complete, now in optimization and feature expansion phase. USPS RFP creates potential new engagement opportunity.
BUDGET
$165KRemaining
$18KMonthly rate
36.7Runway (weeks)
On Track: Yes
Document pipeline deployed to production and processing 2,400 documents/day at 95.8% accuracy. Client NPS shifted +18 points since deployment. The team is now in optimization and expansion phase. A USPS modernization RFP creates a significant new business opportunity that leverages the existing platform.
CLIENT SENTIMENT
Maria SantosCOO
Highly positive. Referenced the project in an all-hands as "the model for how we should work with technology partners." Actively exploring expansion.
Kevin ChoHEAD OF OPERATIONS
Using the dashboard daily. Submitted 6 feature requests in the last two weeks, all reasonable and well-scoped.
STRUCTURAL Build on what's working
ACT
USPS Postal Modernization RFP published May 10. Includes document classification and intelligent routing requirements.
Relevance: Direct fit for Atlas's document intelligence platform. See strategic recommendation for co-bid opportunity.
Source: SAM.gov, May 10 2026WATCH
Amazon announced a logistics document automation pilot in partnership with Anthropic.
Relevance: Validates the category. Atlas should position their proven production system against the "pilot" framing. Existing scale (2,400 docs/day) is a differentiator.
Source: Reuters, May 18 2026ALL ACTIONS
Across the portfolio
CRITICAL
NEEDS FLAG
Scope renegotiation meeting with Dr. Osei. Budget exhausts 3 weeks before completion.
CRITICAL
NEEDS FLAG
Ensure Acme SOW is signed before Sprint 15 kickoff.
HIGH
NEEDS FLAG
Schedule load testing planning session with Jordan Park.
HIGH
NEEDS FLAG
Acknowledge scope absorption to Meridian team. Two seniors considering reassignment.
HIGH
NEEDS FLAGBET
Investigate USPS Postal Modernization RFP. $2-5M co-bid opportunity with Atlas.
MEDIUM
TEAM HANDLING
Share caching solution from Atlas Sprint 8 with Meridian and Pinnacle teams.
PORTFOLIO STRATEGY
Strategic Analysis
Portfolio meta-analysis and recommendations
$1.2MBilled YTD
3Active Projects
42%Top Client Concentration
4Unnoticed Actions
PORTFOLIO HEALTH
REVENUE CONCENTRATION
Acme Corp represents 42% of YTD revenue. Loss of this single engagement would create a significant revenue gap. Diversification is a priority.
RESOURCING & ALLOCATION
Staffing signals surfaced from project briefs, political diagnostics, and team sentiment data
CAPACITY ISSUES
Priya MehtaAcme Corp
Carrying tech lead responsibilities on Acme while also providing caching consultation to Meridian and Pinnacle. At 100% allocation with cross-project requests adding 10-15 hrs/wk.
Flagged in two retros. Declining PR review response times (48h → 72h over last month).
Assign a second senior engineer to Acme for Sprint 15 to absorb code review load. Priya's cross-project consultation should be formally allocated, not absorbed.
PERFORMANCE SIGNALS
Meridian dev teamMeridian Health
Two senior engineers have mentioned reassignment interest to Alex Rivera. Morale declining from sustained scope absorption without acknowledgment.
Alex raised in PD channel 5/18. No response from leadership as of 5/23.
Acknowledge the situation directly (execution kit drafted). Resolve the commercial issue before morale degrades further.
CROSS-PROJECT OPPORTUNITIES
Caching architectureAtlas → Meridian, Pinnacle, Coastal
Atlas's TTL-based caching solution from Sprint 8 directly applies to three other engagements. No knowledge transfer has happened.
15-minute walkthrough from Priya to affected teams this week. Estimated savings: 1-2 sprint weeks per project.
COMMON PAIN POINTS
Recurring problems across multiple projects with suggested fixes
Scope instability on verbal agreements
Acme Corp · Meridian Health
Two engagements are operating without adequate scope boundaries. Acme has an unsigned SOW after 14 weeks. Meridian absorbed a HIPAA compliance layer on a fixed-price contract without a change order. Both teams are absorbing the cost of scope drift.
Implement a scope gate protocol across all engagements: when a client sends a new direction, respond with a 1-line impact note (what gets delayed, hours involved, timeline effect) before pivoting. For fixed-price contracts, require a change order before work begins on out-of-scope items.
Shared technical blocker: caching architecture
Meridian Health · Pinnacle Finance · Coastal Energy
Three engagements hit the same caching invalidation problem on document-heavy pipelines. Atlas Logistics solved this in Sprint 8 with a TTL-based approach. No knowledge transfer has occurred.
Share Atlas's caching implementation with the three affected teams. Estimated savings: 1-2 sprint weeks per project. Assign the Atlas tech lead (Priya Mehta) to run a 15-minute walkthrough for affected teams this week.
Untapped expansion: government RFP pipeline
Atlas Logistics
Atlas's document intelligence platform has proven production performance (2,400 docs/day, 95.8% accuracy) that maps directly to federal RFP requirements. The USPS Postal Modernization RFP is a $2-5M opportunity. No systematic process exists for scanning government RFPs against active project capabilities.
Add a quarterly government RFP scan to the market signals process. Match published RFPs against the capability profiles of active projects. Start with Atlas as the pilot and expand to other engagements with production-proven platforms.
MARKET SCAN
Trends affecting AI development studios and the consulting market
ACTEnterprise AI budget consolidation accelerating
Gartner reports 60% of enterprise AI buyers are consolidating from 5+ vendors to 2-3 strategic partners by end of 2026. Studios that can demonstrate breadth (strategy through production through governance) win consolidation decisions. Point-solution vendors are being cut.
Source: Gartner AI Vendor Consolidation Report, May 2026
WATCHFederal AI procurement expanding to mid-size contractors
GSA announced simplified AI procurement vehicles for contracts under $5M. Previously required large prime contractor relationships. Opens federal work to studios of AE's size directly. Atlas USPS opportunity is an early example.
Source: GSA.gov, April 2026
OUTSIDE PERSPECTIVE
What a candid board advisor would tell AE leadership
You're under-pricing scope risk on fixed-price contracts
Meridian is the second fixed-price engagement in six months where unscoped compliance work ate the margin. The pattern is clear: clients request regulatory additions mid-project, the team says yes to preserve the relationship, and AE absorbs the cost. The relationship is preserved. The margin is not.
Add a compliance risk buffer (15-20%) to every fixed-price SOW that touches regulated data. Or switch to T&M with a not-to-exceed cap, which gives the client budget certainty while giving AE scope flexibility.
Your best engineer is your biggest single point of failure
Priya Mehta is tech lead on Acme, caching consultant to three other projects, and the person everyone calls when something is hard. If she takes a two-week vacation, four projects feel it. That's not a staffing plan, it's a dependency.
Pair a second senior engineer with Priya on Acme for the remaining sprints. Document the caching architecture so the knowledge transfers. The goal is that Priya is valuable because she's excellent, not because she's irreplaceable.
SALES INTELLIGENCE
Sales Pipeline
Active deals from BANT Qualified through Contract Review
NEXGEN ROBOTICS
Computer Vision PlatformAccelerating
Champion secured budget approval. Technical validation complete. SOW in legal review.
BRIGHTPATH HEALTH
Patient Data PipelineSlowing
Champion went quiet after internal reorg announcement. No response to last two follow-ups.
FLUX LOGISTICS
Route OptimizationStalled
Budget holder changed. New VP re-evaluating all AI vendors. 6 weeks since substantive contact.
← Sales Pipeline
NexGen Robotics
Accelerating Recommendations Overview Health Check Market Signals
CONTACTS
NameRoleFocus
Sam TorresVP EngineeringChampion, budget holder
Dr. Rita PhanCTOArchitecture sign-off
AE SIDE
Troy LarssenPDRelationship, commercial
Sarah ChenPMScoping, staffing
DEAL DETAILS
$420KDeal value
Contract ReviewStage
Jun 6Expected close
Troy LarssenOwner
CONTEXT
Computer vision platform for quality inspection on the manufacturing floor. NexGen builds industrial robots and needs real-time defect detection. Technical validation showed our approach outperformed their existing system by 3.2x on false positive rate. Sam Torres is the internal champion, secured budget approval from CFO last week.
Momentum: Accelerating
Champion (VP Engineering, Sam Torres) secured budget approval from the CFO last week. Technical validation passed with strong results. SOW in legal review, expected turnaround 5-7 business days. No blockers identified.
HEALTH META
AcceleratingMomentum
HighClient engagement
StrongCompetitive position
RISKS
LOWLegal review could extend past expected 5-7 days if NDA terms require negotiation.Troy to check in with Sam on legal timeline by Wednesday.
TIMELINE
MAY 22
CFO budget approval confirmed. Sam Torres emailed Troy directly.
MAY 15
Technical validation complete. 3.2x improvement on false positive rate vs. existing system.
MAY 8
Second technical deep-dive with Dr. Rita Phan (CTO). Architecture approach approved.
WATCH
Cognex (NexGen's current vision vendor) announced a new AI-powered inspection module at Automate 2026.
Relevance: Validates the category. NexGen may compare our approach against Cognex's new offering. Our technical validation results (3.2x improvement) are the strongest counter.
Source: Automate 2026 press releases, May 20 ← Sales Pipeline
BrightPath Health
Slowing Recommendations Overview Health Check Politics
CONTACTS
NameRoleFocus
Jamie TorresVP DataChampion (status uncertain post-reorg)
Dr. Amy LinCTOTechnical evaluation, architecture
AE SIDE
Troy LarssenPDRelationship, commercial
DEAL DETAILS
$350KDeal value
SOW SentStage
Jul 1 (uncertain)Expected close
CONTEXT
HIPAA-compliant patient data pipeline. SOW sent April 28. Strong technical alignment during the evaluation phase. Jamie Torres (VP Data) was the active champion driving the process. Internal reorg announced May 5. No substantive contact since.
NEXT STEPS
Re-engage through CTO (Dr. Amy Lin) side channel to determine whether Jamie's role/authority changed in the reorg. If deal is alive, update SOW if priorities have shifted. If champion is displaced, re-qualify with new stakeholder.
Momentum: Slowing
Champion silent for 3 weeks post-reorg. SOW sent April 28 with no response. Two follow-up emails (May 8, May 15) unanswered. No public information about changes to the VP Data role. The deal is not dead, but the silence pattern after a reorg historically correlates with a significant probability reduction.
HEALTH META
SlowingMomentum
LowClient engagement
UnknownCompetitive position
RISKS
HIGHChampion may have lost authority or been reassigned in the reorg.Engage CTO as alternate path to determine champion status.
MEDIUMReorg may have reprioritized the data pipeline project entirely.Frame re-engagement as a check-in, not a push. Allow for scope changes.
TIMELINE
MAY 15
Second follow-up email to Jamie. No response.
MAY 8
First follow-up email to Jamie. No response.
MAY 5
BrightPath internal reorg announced.
APR 28
SOW sent to Jamie Torres. Strong alignment on technical approach.
POLITICAL COMPLEXITYRED
Internal reorg announced May 5. Champion (VP Data) has gone quiet. Unclear whether their role, budget, or authority has changed. CTO attended original demo and expressed interest but is not the direct buyer.
← Sales Pipeline
Flux Logistics
Stalled Recommendations Overview Health Check
MEDIUM
NEEDS FLAG
Investigate: deprioritize or re-approach. New VP is re-evaluating all AI vendors. 6 weeks since contact. Either find a path to the new decision-maker or move to backlog and revisit in Q4.
ALEX RIVERACONTACTS
NameRoleFocus
Previous: Mark DanielsVP Operations (departed)Original champion
Current: UnknownNew VP OperationsRe-evaluating all AI vendors
AE SIDE
Alex RiveraPMRelationship, scoping
DEAL DETAILS
$500KDeal value
BANT QualifiedStage
Q4 (uncertain)Expected close
CONTEXT
Route optimization for a mid-size logistics company. Original champion (Mark Daniels, VP Ops) was replaced 6 weeks ago. New VP is reportedly re-evaluating all AI vendor relationships. No direct contact with the new decision-maker has been established.
Momentum: Stalled
Budget holder changed. New VP re-evaluating all AI vendors. 6 weeks since last substantive contact. The deal was BANT-qualified with the previous VP, but qualification does not transfer to the new decision-maker. Effectively needs to be re-qualified from scratch.
HEALTH META
StalledMomentum
NoneClient engagement
UnknownCompetitive position
TIMELINE
APR 14
Last contact with Mark Daniels. Discussed next steps on SOW.
APR 21
Mark Daniels replaced. New VP announced. No introduction to AE.
MAY 1
Alex sent intro email to Flux general inbox. No response.
← Sales Pipeline
DEAL ACTIONS
Action items across all active deals
HIGH
TEAM HANDLING
Pre-stage engineering team for NexGen before close.
HIGH
NEEDS FLAG
Re-engage BrightPath champion through CTO side channel.
MEDIUM
NEEDS FLAG
Deprioritize or re-approach Flux. New VP re-evaluating vendors.
SALES STRATEGY
Pipeline Analysis
Strategic analysis of the sales pipeline
$1.27MPipeline Value
3Active Deals
$420KWeighted (Human)
$276KWeighted (Intel)
PIPELINE MOMENTUM
PIPELINE PATTERNS
Champion dependency across the pipeline
BrightPath Health · Flux Logistics
Two of three active deals depend on a single champion who is either silent (BrightPath) or displaced (Flux). When the champion goes dark, the deal has no fallback relationship. The pipeline is structurally fragile to organizational changes at target companies.
For every deal past BANT, establish a secondary relationship (the champion's manager or a peer stakeholder) before the deal reaches SOW stage. The cost is one extra meeting. The insurance is worth it.
Human probability estimates are systematically optimistic
BrightPath Health (60% human vs 35% intel) · Flux Logistics (40% human vs 15% intel)
Intel probability estimates diverge significantly from human estimates on the two stalled deals. In both cases, the intel model identifies concrete signals (champion silence, decision-maker change) that the human estimates do not fully weight. Historical data supports the intel model: deals with these signals close at approximately half the human-estimated rate.
Update pipeline forecasting to use intel-adjusted probabilities for deals showing stall signals. This shifts August pipeline revenue projection from $285K to $198K, making the revenue gap alert more accurate and the response more urgent.
NexGen close timing affects capacity planning
NexGen Robotics
NexGen is the highest-probability deal in the pipeline (85-90%) and closes into the same week as the capacity cliff from Acme and Atlas roll-offs. If NexGen closes on time, 80 of the 120-hour gap is filled. If it slips two weeks, the gap persists and the team sits idle.
Pre-stage the NexGen team (staffing memo already drafted). Ask legal for an expedited SOW review. Every day of acceleration reduces the capacity gap risk.
OUTSIDE PERSPECTIVE
What a candid advisor would tell AE about the pipeline
Your pipeline is one deal deep at every stage
One deal in Contract Review, one in SOW Sent, one in BANT. If NexGen slips, there's nothing behind it. If BrightPath dies, there's no other SOW-stage deal. The pipeline has zero redundancy at any stage.
Increase top-of-funnel activity to maintain at least two deals at each stage. The current pipeline supports the next quarter. It does not survive a single unexpected loss.
The intel probability divergence is a feature, use it
Intel sees BrightPath at 35% where the human estimate is 60%. Flux at 15% vs 40%. The pattern is consistent: human estimates don't fully weight negative signals. If the intel model is right (and historical data supports it), August revenue is $87K lower than the human forecast.
Adopt a policy: when intel diverges from human by more than 15%, flag the deal for review and require the owner to either update the human estimate or document why the intel model is wrong. This creates accountability without removing human judgment.
PROJECTIONS
Revenue and capacity outlook
12-week horizon · May 26 – Aug 18, 2026 · Generated May 23
Timeline Overview Gap Analysis Capacity Pipeline
ENGAGEMENT TIMELINE — NEXT 12 WEEKS
ProjectJunJulAug
Acme Corp
Meridian Health
Atlas Logistics
PIPELINE TIMELINE
DealJunJulAug
NexGen
BrightPath
Flux
HEADCOUNT CHANGES
WeekEventImpact
Jun 2Marcus Webb available (Atlas → 0.5 FTE)+32 hrs/wk ML
Jun 9NexGen start (if closed)-80 hrs/wk allocated
Jul 14Priya, James, Anika roll off Acme + Atlas-120 hrs/wk capacity cliff
MONTHLY REVENUE PROJECTION
JuneJulyAugust
Total$482K$445K$380K
From engagements$412K$350K$280K
From pipeline (weighted)$70K$95K$100K
Human predicted Intel predicted
GAP ALERTS
CRITICALWeek 8 (Jul 14)
Capacity cliff: 3 senior engineers rolling off Acme and Atlas in the same week. No pipeline deals close in time to absorb. 120 hrs/wk gap.
HIGHWeek 6 (Jun 30)
Revenue gap: August projected at $380K against $450K target. Pipeline probability-weighted revenue covers only 60% of the shortfall.
MEDIUMWeek 10 (Jul 28)
Skill mismatch: pipeline deals require ML engineering. Current bench weighted toward frontend.
WEEKLY SUPPLY VS. DEMAND
WeekCapacity (hrs)Required (hrs)Gap
Week 1 (May 26)480440+40
Week 2 (Jun 2)480460+20
Week 4 (Jun 16)4404400
Week 6 (Jun 30)400440-40
Week 8 (Jul 14)280400-120
Week 10 (Jul 28)280360-80
Week 12 (Aug 11)280320-40
GAP BY ROLE (WEEKLY HEATMAP)
W1W3W5W7W9W11
Sr. Engineer+40+200-40-120-80
ML Engineer+8+800-48-48
Tech Lead0000-40-40
PM000000
Designer+8+8+8+8+8+8
GAP ALERTS
CRITICALWeek 9 (Jul 14)
Capacity cliff: 3 senior engineers rolling off. 120 hrs/wk gap in Sr. Engineering alone.
HIGHWeek 9 (Jul 14)
Tech Lead gap: Priya rolls off with no replacement. 40 hrs/wk gap.
MEDIUMWeek 9 (Jul 14)
Skill mismatch: pipeline deals require ML engineering. Current bench weighted toward frontend.
HEADCOUNT OVER TIME
Human (allocated) Intel (predicted)
TEAM DETAIL
NameRoleProjectHrs/wkHuman EndIntel EndDiverge
Priya MehtaTech LeadAcme40Jul 14Jul 28+2 wks
James LiuSr. EngAcme40Jul 14Jul 28+2 wks
Anika SharmaSr. EngAtlas40Jul 14Jul 140
Marcus WebbML EngAtlas32Aug 8Aug 80
Alex RiveraPMMeridian20Jul 7Jun 16-3 wks
Sarah ChenPMAcme20Jul 28Jul 280
Rachel TorresDesignerAtlas24Sep 1Sep 10
Dev Team (3)EngineersMeridian120Jul 7Jun 16-3 wks
PROJECT OUTLOOKS
Acme CorpNeeds Attention
Engagement likely extends 2-4 weeks past current estimate if load testing scope is added to Sprint 15. Budget runway supports the extension at current burn rate if SOW amendment covers the delta.
TEAM PROJECTIONS
NameRoleHrs/wkHuman EndIntel End
Priya MehtaTech Lead40Jul 14Jul 28 (+2 wks)
James LiuSr. Eng40Jul 14Jul 28 (+2 wks)
Sarah ChenPM20Jul 28Jul 28
Intel reasoning: Load testing scope addition and unsigned SOW create timeline risk. Historical pattern: scope additions at this stage extend engagements by 2-4 weeks in 80% of cases.
Meridian HealthAt Risk
Budget exhausts before completion without amendment. If renegotiation succeeds (Option C, $24K), engagement completes on time. If it fails, team rolls off 3 weeks early with incomplete deliverables.
TEAM PROJECTIONS
NameRoleHrs/wkHuman EndIntel End
Alex RiveraPM20Jul 7Jun 16 (-3 wks)
Dev Team (3)Engineers120Jul 7Jun 16 (-3 wks)
Intel reasoning: Without SOW amendment, budget runs out Jun 16. Team morale issues increase probability of early disengagement. Confidence: high.
Atlas LogisticsOn Track
Core engagement completing on schedule. Optimization phase transitioning to 0.5 FTE. USPS RFP opportunity could trigger a new engagement starting August if Atlas decides to bid.
SKILLS RETAINED
WEIGHTED REVENUE BY WEEK ($K)
Human Intel
DEAL PROJECTIONS — HUMAN VS. INTEL
NexGen RoboticsContract Review
HUMAN 85% Weighted: $357K Close: Jun 6
INTEL 90% Weighted: $378K Close: Jun 4
DIVERGENCE +5%
Champion has budget approval and legal is reviewing. Historical close rate for deals at this stage with budget approval: 92%. Intel sees slightly higher probability than human estimate.
BrightPath HealthSOW Sent
HUMAN 60% Weighted: $210K Close: Jul 1
INTEL 35% Weighted: $123K Close: Aug 15
DIVERGENCE -25%
Champion silent for 3 weeks post-reorg. Historical pattern: 70% of deals where champion goes dark for 3+ weeks during reorg close at less than half the original probability. Intel recommends updating human estimate.
Flux LogisticsBANT Qualified
HUMAN 40% Weighted: $200K Close: Q4
INTEL 15% Weighted: $75K Close: --
DIVERGENCE -25%
Budget holder changed. New VP re-evaluating vendors. 6 weeks since contact. Deals stalled at BANT with a decision-maker change close at 15% within 6 months historically.
SKILLS NEEDED FROM PIPELINE
1x ML Engineer (computer vision) · NexGen · Start +2 weeks, 16 weeks duration
Rationale: Core platform work. Computer vision specialization required. Marcus Webb is available and has relevant experience.
1x Sr. Engineer (API/infra) · NexGen · Start +2 weeks, 16 weeks duration
Rationale: API layer and cloud infrastructure. Rachel Torres available Jun 9 if Atlas shifts to 0.5 FTE.
2x Data Engineers · BrightPath · Start +6 weeks, 12 weeks duration (if deal closes)
Rationale: HIPAA-compliant data pipeline. No current bench capacity. Would need to hire or reassign from Meridian post-completion.
Illustrative output. Names and data are fictional. Structure and format reflect the actual system.
THE PATTERN
The same loop, at every scale.
The intelligence layer runs the same core process at every level of the organization. For each entity it monitors, the system:
01 Assess what success looks like
Define the goals, KPIs, and health indicators for the entity.
02 Read the current state
Pull from every connected source to build a composed picture of where things stand.
03 Derive recommendations
Scaffolded analysis methods compare current state against success criteria and generate actionable findings.
04 Route to the right humans
Recommendations go to specific people based on mapped roles and responsibilities. The person who can act is the person who sees it.
This pattern is fractal. It runs for a single project and for the department running all the projects. For a single deal and for the sales organization. For a business unit and for the whole company. Each level composes the levels below it. A department view synthesizes its projects. A company view synthesizes its departments. The same architecture, the same analysis methods, at every scale.
WHAT THIS SURFACES
The kinds of things leadership actually sees.
Two departments are working on similar problems with different vendors. Neither knows about the other. The intelligence layer spots the overlap because both are in the knowledge graph, surfaces the connection, and recommends consolidation with a projected cost savings.
A regulatory change in a client's industry appears in a market scan. The system maps it to two active projects whose scope will be affected, and flags it for the project leads with a summary of the implications and suggested next steps.
Budget burn on a project is tracking 20% ahead of plan at the halfway point. The system catches the trajectory, traces it to scope additions that weren't reflected in the estimate, and recommends a re-scoping conversation with the specific numbers and timeline impact.
A key stakeholder has gone quiet. They were the primary advocate for the project, but meeting attendance dropped and their messages shifted in tone. The system runs a relationship diagnostic, identifies the pattern, and surfaces it to leadership with context on what may have changed and how to re-engage.
Across twelve active engagements, three share a common technical blocker that was solved on a fourth project last quarter. The system surfaces the pattern, points to the existing solution, and recommends knowledge transfer to the teams still stuck.
DESIGN APPROACH
What becomes possible when AI can process everything.
People and AI are both jagged frontiers. People are excellent at judgment, taste, relationships, and context. AI is excellent at processing volume, holding connections across large data sets, and running analysis that no person has the bandwidth for. The intelligence layer is designed around this complementarity: specify the analysis and research you know would provide value but isn't feasible for people to do, and let the system run it continuously.
Analysis at infeasible scale.
Cross-functional pattern detection across every source in the organization, running continuously. Connections between departments, projects, and signals that no individual has the bandwidth to track. The kind of synthesis leadership always wanted but never had the hours for.
External context that provides edge.
Market scans, competitive signals, regulatory changes, and industry patterns pulled from outside the company and mapped to internal context. The system surfaces external factors that could affect your business before someone stumbles across them.
The things people struggle with.
Relationship dynamics, organizational patterns, incentive misalignment. People see pieces of this but rarely have the full picture. The system runs a dedicated diagnostic across all available signals and names what it finds. In practice, this looks like power mapping, stakeholder analysis, and political context on every recommendation.
Auditability by design.
Every claim traces to a primary source. When the system makes a recommendation, the reasoning chain and the evidence are visible. The human can disagree with the conclusion and still act on the underlying data. Trust is built through transparency, and the system is architected to be challenged.
Customizable to your priorities.
The analysis rules, the signal priorities, the thresholds for action, the routing logic: all of these are configured to how your organization works and what your leadership cares about. The architecture is general. The implementation is yours.
Designed to improve over time.
The system tracks whether its own recommendations led to the predicted outcomes. Failed predictions feed back into future analysis. The intelligence layer gets better the longer it runs, calibrated by its own results.
THE ACTION MODEL
Actions that carry their own accountability.
Actions are the system's output channel into human attention. Each one carries enough structure that the recipient can act, disagree, or verify.
action concrete, low-cognitive-load instruction
problem the specific issue, with cited evidence
hypothesis what we expect to happen if it is taken
observable how we will know the hypothesis was right
owner routed to the person whose option space is constrained
awareness handled · unnoticed · deprioritized
Execution kits.
After an action is finalized, the system drafts the artifacts the owner needs: email text, meeting agendas, talking points, message drafts. Pre-filled with names, dates, and references. The aim is that the owner opens the kit and the action is mostly done.
Outcome tracking.
Past actions are checked against their observables on the next cycle. Confirmed outcomes get noted. Failed outcomes get noted. A project where past actions consistently failed their observables triggers more conservative recommendation behavior. The system learns from its own track record.
ALIGNMENT
Alignment is how the system earns trust.
An intelligence layer that curates information for leadership is exactly the deployment context where AI alignment matters most. The system is designed with specific failure modes in mind.
Anti-hallucination by design
Mandatory citations on every claim. When the system cannot cite a source, it says so. The auditability is structural, built into the output format, and verifiable by anyone who reads the brief.
Information governance
The system produces two artifacts per engagement: a full brief for leadership and a sanitized version for the wider team. Praise in public, correct in private. The sanitization is judgment-based, calibrated to protect individuals while preserving actionable information.
Self-correction through outcome tracking
The system tracks whether its recommendations actually worked. Failed hypotheses feed back into future recommendation quality. This is a concrete implementation of the alignment principle that autonomous systems should be calibrated by their own results.
Continuous recalibration
The governance posture updates as capabilities advance. New failure modes identified in alignment research become new evaluations and guardrails in the production system. The alignment architecture is a living layer, designed to keep pace with what the system can do.
WHERE THIS IS GOING
The trajectory toward more autonomy.
Right now, the intelligence layer recommends and humans decide. That's the right posture today. But the trajectory is clear: as the system proves itself, the handoff points shift. Some recommendations become automatic. Some decisions the system makes on its own, with humans reviewing after the fact. The boundary between recommendation and action moves as trust is earned.
That trajectory is both the opportunity and the risk. More autonomy means more leverage. It also means the failure modes become more consequential. An intelligence layer that curates information for leadership is exactly the deployment context where alignment research matters most: sycophancy, reality distortion, selective disclosure, goal drift. These aren't theoretical concerns. They're the specific failure modes of this kind of system.
This is why our alignment research and our production work are the same practice. The failure modes we study in our research are the failure modes of this system. Production informs the research. Research informs the governance. The system gets more autonomous and more trustworthy at the same time.
BEYOND AE
The pattern is general. The implementation is yours.
AE runs this system across its own engagements. The architecture, the cadence, the action model, the alignment properties: these transfer. Any organization with parallel activities, multiple operational tools, and a need to compose those sources into situational awareness can use this shape. The specifics get tailored to your organization. The principles stay.
Talk to us about building yours