Good morning, Sarah
Monday, March 2, 2026 · 8:14 AM CST

8 decisions across 4 clients · 1 bias alert

2 critical actions requiring immediate attention. 1 compliance deadline. 1 bias drift alert from Compliance Guardian. Signal Scout discovered 5 new intelligence signals since yesterday.

Active Decisions
8
2 critical · 3 high · 3 standard
Bias Alerts
1
Engineering pipeline · Gender disparity
Compliance Actions
2
LL 144 notice · IL audit prep
Agents Active
4 / 4
All systems operational
18
Open Reqs
24
Avg Days to Fill
87%
Quality of Hire
2
SLA At Risk
1
Bias Alert
4/4
Agents Online
Today's Priority Decisions
Critical · Top candidate at risk of competing offer
Action needed within 48 hours
Signal Scout Acme Corp · ML Engineer (Sr)

Candidate Flight Risk - Marcus Rivera

Signal Scout detected: LinkedIn headline changed to "Open to opportunities" 2 days ago. 3 new recruiter connections this week. Zero backup candidates at this seniority level for Acme Corp's critical ML role.

⬤ Flight Risk Signal Intelligence
SS
Signal Scout
AI Agent · Detected
DE
Decision Enrichment
AI Agent · Confirmed
Confidence
87%
🕐 48hr window
⬤ Critical priority
Pipeline Monitor
SLA Breach Risk - DataVault Data Analyst
Interview Scheduled stage for 12 days (SLA: 10). Client satisfaction at risk. Recommend client check-in and backup candidate activation.
● 91% confidence
PM
Pipeline Monitor
Compliance Guardian
LL 144 Notice Required - Meridian Health RN
3 candidates screened via AI scoring for NYC role. 10-day candidate notice window starts now. Failure to notify: $500–$1,500/day/applicant.
⚠ Compliance
CG
Compliance Guardian
Compliance Guardian · Bias Alert
Selection Rate Disparity - Engineering Pipeline
Female candidates advancing at 62% the rate of male candidates across 3 Engineering pipelines. Approaching 4/5ths rule threshold. Review recommended before advancing next batch.
⬡ Bias Alert
CG
Compliance Guardian
Pipeline Monitor
Bill Rate Below Market - TechFlow Sr DevOps
Current bill rate at $70/hr is 8% below market median ($76/hr). 3 candidates declined similar placements this quarter. Rate gap likely contributing to pipeline stall.
● 89% confidence
PM
Pipeline Monitor
Signal Scout + Decision Enrichment
Reclassify Marcus Rivera → ML Engineer Pool
4 new ML GitHub repos (TensorFlow, PyTorch). ML Specialization completed Feb 2026. LinkedIn headline: "Software Engineer | ML Enthusiast."
● 82% confidence
SS
Signal Scout
Recent Agent Activity
Signal Scout detected Marcus Rivera LinkedIn change2h
Compliance Guardian triggered LL 144 notice alert4h
Compliance Guardian detected gender disparity in Engineering6h
Pipeline Monitor flagged DataVault SLA breach risk8h
Pipeline Monitor flagged TechFlow rate below market median12h
Client Overview
Acme Corp3 reqs
Meridian Health2 reqs
TechFlow Inc.3 reqs
DataVault Inc.2 reqs
Upcoming Compliance
LL 144 Annual AuditMar 15
IL HB 3773 ReviewMar 20
CO AI Act PrepApr 1
Decision Queue
8 decisions awaiting your review. Sorted by priority and time sensitivity.
DecisionDeadlineClientAgentTierFlagsConfidence
Candidate Flight Risk - Marcus RiveraLinkedIn activity surge, 0 backup candidates, Acme Corp ML Engineer
⚡ Affects 3 areas		 48 hrs Acme Corp Signal Scout Tier 2 - 87%
LL 144 Notice Required - Meridian Health RN3 AI-screened candidates, NYC jurisdiction, 10-day window
 Today Meridian Compliance Tier 3 ⚠ Compliance 99%
Selection Rate Disparity - Engineering PipelineFemale candidates at 62% advancement rate vs. male, 3 pipelines
 Review Multiple Compliance Tier 3 ⬡ Bias 94%
SLA Breach Risk - DataVault Data Analyst12 days in Interview Scheduled (SLA: 10), client escalation risk
 Mar 4 DataVault Pipeline Tier 2 - 91%
Bill Rate Below Market - TechFlow Sr DevOps$70/hr vs $76/hr median, 34% acceptance probability
 Mar 5 TechFlow Pipeline Tier 3 - 89%
Reclassify Marcus Rivera → ML Engineer Pool4 ML repos, ML cert, LinkedIn headline shift
 Mar 7 Internal Match + Signal Tier 2 - 82%
Pipeline Stall - TechFlow Product Manager3 candidates in screening for 8 days, no interview scheduled
 Mar 6 TechFlow Pipeline Tier 2 - 78%
IL HB 3773 Audit PreparationIllinois jurisdiction roles need documentation review by Mar 20
 Mar 20 Multiple Compliance Tier 3 ⚠ Compliance 95%
← Back to Decision Queue
✅
Decision Approved - Candidate Submitted to Client
Approved by Sarah Chen · Just now · Submittal package sent to Acme Corp
Signal Scout ◆ Tier 2 · Recommended DEC-2026-0847 ● Critical

Candidate Flight Risk - Marcus Rivera

Signal Scout detected significant flight risk indicators for your top ML Engineer candidate. Zero backup candidates at this seniority for Acme Corp's critical role. Recommend immediate outreach and expedited interview scheduling.

Pipeline Stage
✓
Sourced
TalentMatch
✓
Screened
Bullhorn
●
Ready to Submit
TalentPilot
Interview
Client
Offer
Bullhorn
Placed
TalentPilot monitors
Client: Acme Corp Role: Sr. ML Engineer Req Age: 18 days
Where this data comes from 4 sources ▾
SS
Signal Scout
AI Agent · Detected ✓
→
DE
Decision Enrichment
AI Agent · Confirmed ✓
→
SC
Sarah Chen
Sr. Recruiter · Awaiting Decision
Action Window
48 hours
AI Match Score
87%
Top candidate for this role
Flight Risk
High
3 indicators detected
Backup Candidates
0
Critical vulnerability

Candidate Profile - Beyond the Resume

Glass Box
Multi-Dimensional Assessment AI scores · Human-adjustable weights
Technical Skills Resume + GitHub + Assessment
91%
30%
Growth Trajectory Certifications + Learning + GitHub trend
94%
20%
Collaboration Signals References + Team projects + PR reviews
78%
20%
Cultural Alignment Client work style analysis + Interview signals
72%
15%
Stability Indicators Tenure patterns + Career progression logic
65%
15%
Weighted Overall Score 83.4%
⬡ Compliance Guardian: Current weights pass proxy variable check. No demographic correlation detected in active factors.
MR
Marcus Rivera
Software Engineer → ML Engineer
Austin, TX · Remote OK
Experience6 years (4 SWE + 2 ML transition)
EducationB.S. Computer Science, UT Austin
Key SkillsTensorFlow, PyTorch, Python, AWS SageMaker
CertificationsAWS ML Specialty (2026), DeepLearning.AI (2025)
🔍 Signal Scout Insight
Marcus's GitHub shows 4 new ML repositories in 90 days (TensorFlow, PyTorch projects). ML Specialization cert completed Feb 2026. LinkedIn headline changed to "Open to opportunities" 2 days ago. 3 new recruiter connections this week. Pattern indicates active job search acceleration.
📊 Placement Outcome Predictor
Similar Profile Success Rate79%
Avg Tenure (similar hires)18 months
Key Success FactorGrowth trajectory (strongest predictor)
Key Risk FactorCareer transition period (stability)
Based on 847 similar placements over 3 years. Model accuracy: 84%.

Counter-Factual Analysis

What If?
What if we remove "years of ML experience" as a factor?
Marcus's score changes from 87% to 91%. His growth trajectory and project quality compensate. 3 additional candidates enter top-5 pool - all career transitioners with strong GitHub portfolios.
⬡ "Years of ML experience" correlates with age at r=0.72. Removing it improves diversity of submittal pool without reducing predicted placement success.
What if Acme Corp's "culture fit" feedback from last 2 rejections is weighted higher?
Marcus's cultural alignment score drops from 72% to 64%. However, analysis shows Acme's "culture fit" rejections correlate with communication style preference, not job performance.
⬡ Acme Corp has rejected 3 of last 5 submittals citing "culture fit." Historical data shows 2 of those 3 rejection reasons were later filled by similar profiles. Possible preference pattern - flagged for review.
What if we factor in Marcus's flight risk signals?
Without intervention within 48 hours, probability of losing Marcus to competing offer: 67%. Cost of losing this candidate and restarting search: estimated $18,500 (3-4 weeks delay + sourcing costs).
🔍 Flight risk calculation based on Signal Scout pattern matching: LinkedIn activity change + new recruiter connections + market demand for ML Engineers in Austin (1.4:1 role-to-candidate ratio).
⚠ Why This Transparency Matters - Amazon, 2018 ▾ Read more
⚠
Why This Transparency Matters - Amazon, 2018
Amazon's internal hiring AI was scrapped after it was discovered the system penalized resumes containing the word "women's" and downranked graduates of all-women's colleges. The system had no glass-box transparency - nobody could see what factors drove recommendations until investigative reporting exposed the bias. TalentPilot's counter-factual panel and proxy variable checks exist precisely to prevent this. Every factor is visible. Every weight is adjustable. Every adjustment is bias-checked.
📋 Evidence Chain - 3 Sources ▾ Show sources
Signal Intelligence● High Confidence

LinkedIn Activity Surge - Marcus Rivera

Headline changed to "Open to opportunities" Feb 28. 3 new recruiter connections from Meta, Stripe, OpenAI. Profile views up 340% week-over-week.
Source: Signal Scout · LinkedIn Monitoring
Detection timestamp: 2026-03-01T06:14:22Z
Confidence: 87% (3/3 indicators positive)
Pattern: "Active search acceleration" - matches 89% of candidates who accept competing offers within 14 days
View source: Signal Scout Agent Log →
Market Intelligence● High Confidence

Austin ML Engineer Market - Supply/Demand

Austin ML Engineer market: 1.4 roles per candidate. Average time-to-fill: 34 days. Median bill rate: $80/hr. Marcus's profile is in the top 15% of available candidates.
Source: Pipeline Monitor · Market Rate Analysis
Data freshness: Updated daily
Comparable roles analyzed: 127 active postings
Bill rate range (P25-P75): $72/hr - $86/hr
Client History◆ Pattern Flag

Acme Corp Hiring History - ML Roles

Acme Corp has taken an average of 42 days to fill ML roles (industry avg: 34). Last 2 submittals rejected for "culture fit." Previous successful placement stayed 22 months.
Source: Pipeline Monitor · Historical Analysis
Total ML placements with Acme: 4 (past 2 years)
Success rate: 75% (3/4 placed 12+ months)
Rejection patterns: "Culture fit" cited 60% of rejections - above agency average of 23%

Recommended Action Plan

AI Generated
⚡ 4-step action plan · Estimated completion: 3 days
1

Immediate outreach - retention call with Marcus

Express strong interest, share timeline acceleration, gauge competing offer status

Sarah Chen
Today
2

Expedite Acme Corp interview scheduling

Request fast-track to final round. Share candidate flight risk context with hiring manager

Sarah Chen
Mar 3
3

Prepare client-specific interview coaching

Brief Marcus on Acme's "culture fit" evaluation patterns. Focus on communication style alignment

Sarah Chen
Mar 3
4

Activate backup sourcing for ML Engineer pipeline

Even if Marcus advances, pipeline has zero backup. Signal Scout to prioritize ML candidate identification

Decision Enrichment
Mar 5

This decision affects 3 other areas

⚡ Cross-Impact
⚡ Cross-functional impact detected 4 stakeholders · 2 clients affected
Client Health

Acme Corp Placement at Risk

Marcus Rivera departure triggers 90-day replacement SLA — client health score drops from 92 to 74

GP
Gavin Parker
-18 pts
Pipeline

ML Engineer Pipeline Needs Immediate Sourcing

0 backup candidates for Acme Corp role — Signal Scout flagged 3 passive candidates requiring outreach

SC
Sarah Chen
0 backups
Compliance

Selection Rate Review Triggered

Engineering pipeline already flagged for gender disparity — replacement hire enters a monitored pipeline with 4/5ths rule tracking active

CG
Compliance Guardian
4/5ths watch
Team Comp

Acme Corp Engineering Team Gap

Team drops from 4 to 3 engineers — sprint velocity impact estimated at 25% until backfill, affecting Q2 deliverables

DP
David Park
-25% velocity

Compliance Checkpoint

✓ Clear
Bias Scan
✓ No proxy variables detected in recommendation factors
LL 144 (NYC)
Not applicable - Austin, TX jurisdiction
IL HB 3773
Not applicable - TX jurisdiction
CO AI Act
Not applicable - TX jurisdiction
EU AI Act
Not applicable - US-only role
EEOC Guidelines
✓ Selection criteria comply with disparate impact analysis
Audit Trail ID
SP-AUDIT-2026-0302-0847-A
Awaiting your decision · Candidate Flight Risk - Marcus Rivera
Viewing as:
Good morning, Gavin
Monday, March 2, 2026 · 8:14 AM CST

4 active clients · 1 at-risk placement · $2.4M portfolio

Your portfolio health score is 78/100. Acme Corp requires attention - dual-signal analysis detected alignment concerns between placed worker and client. TechFlow contract renewal due in 21 days.

Portfolio Health
78
2 healthy · 1 watch · 1 action needed
Active Placements
13
Across 4 clients · 3 renewals pending
Revenue at Risk
$340K
Acme Corp alignment · TechFlow renewal
Client NPS
72
↑ 4 points from last quarter

Client Portfolio

Click any client to expand details
ClientHealthWorkersRevenueKey Signal
AC
Acme Corp
Enterprise SaaS · Austin, TX
 Action Needed 3 · 1 at-risk $680K ⚠ Alignment gap - rebalancing available ▴
Overview
Placements (3)
Intelligence
Actions
67%
Fill Rate
↓ 12%
42d
Avg Time-to-Fill
vs 34d avg
60%
Submittal Accept
3/5
$680K
Annual Revenue
3 active roles
⚠ Placement Pulse - Signal Gap Detected
Worker reports positive (78% sentiment). Client language includes "revisit scope" and "whether current approach aligns" - 62% alignment. 38% correlation with scope reduction within 60 days.
🔄 Rebalancing: Marcus → DataVault (82% match). $165K preserved, $18.5K saved.
⚠ At stake: $165K–$340K revenue · Intervention: ~2 hours · ROI: 85x–170x
MR
Marcus Rivera
ML Engineer · Month 5 of 12
62%
⚠ At Risk ▾
SK
Sarah Kim
UX Designer · Month 8
88%
✓ Aligned ▾
DO
David Okafor
Frontend Dev · Month 4
85%
✓ Aligned ▾
📊 Intelligence You Can Share With Acme
Acme's ML Engineer time-to-fill (42 days) is 24% above Austin average (34 days). Competitors offering remote-first ML roles fill 31% faster. Recommendation: Suggest expanding remote flexibility or adjusting bill rate to $80–$86/hr range in your next QBR.
📈 Signal Scout Insight
Client language shifted from "expanding team" (Q4) to "revisit scope" (Q1). Historical pattern across your portfolio: this language shift precedes scope reduction 38% of the time within 60 days. Proactive check-in recommended before Q2 planning cycle.
✓
Schedule check-in with Jennifer Park (Engineering Director) to assess Q3 roadmap impact on ML workstream. High priority · This week
🔄
Evaluate Marcus → DataVault rebalance - 82% skill match, $165K preserved, zero revenue gap. High priority · Decision needed
◆
Brief Marcus on potential scope evolution to align expectations and prevent surprise. Medium · After client check-in
◆
Prepare QBR deck with time-to-fill benchmarking and remote flexibility recommendation. Medium · Before March 15
TF
TechFlow Systems
FinTech · Dallas, TX
 Renewal Watch 4 · all aligned $520K ◆ Renewal in 21 days + upsell opportunity ▾
MH
Meridian Health
Healthcare Tech · Houston, TX
 Healthy 5 · all aligned $840K ✓ Highest performing · HIPAA trend detected ▾
DV
DataVault Inc.
Cloud Security · Plano, TX
 New Client 1 · onboarding $340K ● SOC 2 expansion opportunity detected ▾

Your Performance

Gavin Parker vs. agency benchmarks
94%
Placement Retention Rate
Top 8% · vs 87% agency avg
$2.4M
Portfolio Revenue
Top quartile · 4 clients
2.1x
Upsell Conversion
vs 1.4x agency avg
72
Client NPS (avg)
↑ 4pts from Q4 · vs 64 agency avg

Cross-Portfolio Intelligence

Trends across all 13 placed workers
🔄

Talent Liquidity

1 at-risk worker matches 1 open role at another client. Proactive transition preserves $165K/yr with zero gap days.

📈

Emerging Skill Signal

7 of 13 workers mentioned AI/ML tools this month (up from 2 six months ago). Your clients are adopting AI faster than expected.

⏱

Overtime Pattern Alert

3 workers at Acme averaging 46+ hrs/week for 3 weeks. 28% higher flight risk within 90 days.

System Control Panel · David Park, VP Recruiting Ops
Monday, March 2, 2026 · Last system check: 2 min ago

4 agents active · 0 circuit breakers · All systems operational

Configure AI agent behavior, set guardrails, manage autonomy levels, and monitor system health. Changes take effect immediately across all active decisions.

Agents Online
4 / 4
All operational · No degradation
Actions Today
128
78 scans · 35 alerts · 15 recommendations
Circuit Breakers
0
No agents paused · All thresholds healthy
System Uptime
99.7%
30-day rolling · 4.2 hrs planned maintenance

Agent Fleet

Configure individual agent behavior, autonomy, and thresholds
DE
Decision Enrichment
Adds beyond-the-resume intelligence to AI matches - growth trajectory, cultural alignment, and stability signals
Active
23
Actions today
92%
Accuracy (30d)
1.2s
Avg response
Autonomy Level
Confidence Threshold
75%
Kill Switch
Running
Last action: Enriched Marcus Rivera profile - growth trajectory +18%, stability risk flagged from 3 job changes in 4 years · 4 min ago
PM
Pipeline Monitor
Tracks candidate flow, detects bottlenecks, monitors rate competitiveness, and flags time-to-fill anomalies
Active
38
Actions today
88%
Accuracy (30d)
0.8s
Avg response
Autonomy Level
Confidence Threshold
70%
Kill Switch
Running
Last action: Flagged TechFlow pipeline stall - 3 candidates idle >5 days · 12 min ago
CG
Compliance Guardian
Monitors for bias, regulatory violations, proxy variable correlations, and audit readiness
Active
28
Actions today
96%
Accuracy (30d)
2.1s
Avg response
Autonomy Level
Confidence Threshold
85%
Kill Switch
Running
Last action: Bias drift alert - Engineering pipeline gender disparity widening · 1 hr ago
SS
Signal Scout
Monitors LinkedIn, GitHub, and market signals for flight risk and opportunity detection
Active
24
Actions today
87%
Accuracy (30d)
3.4s
Avg response
Autonomy Level
Confidence Threshold
70%
Kill Switch
Running
Last action: Detected Marcus Rivera flight risk - LinkedIn headline change + recruiter connections · 2 hr ago

Global Guardrails

System-wide rules applied across all agents
⚖
4/5ths Rule Monitoring
Alert when selection rate for any group falls below 80% of the highest group's rate (EEOC guideline)
Enforced
🔍
Proxy Variable Detection
Flag when weight adjustments create >35% correlation with protected characteristics
⬆
Auto-Escalation Threshold
Decisions above this risk score automatically escalate to VP review
85%
🔒
Human Approval Requirement
Which decisions require explicit human sign-off before action
📋
Audit Trail Retention
How long decision logs, weight changes, and bias checks are preserved
7 years NYC LL 144 · Illinois AIPA · EU AI Act
🔄
Learning Loop Circuit Breaker
Auto-pause model retraining if accuracy drops below threshold in 7-day window
80%

Circuit Breakers

Automatic safety stops - these trigger when system boundaries are crossed
Bias Drift Breaker
Triggers if 4/5ths rule violation sustained for 48+ hours
Healthy
Last checked: 4 min ago
Accuracy Degradation Breaker
Triggers if any agent accuracy drops below 75% in 7-day window
Healthy
Last checked: 4 min ago
Data Source Breaker
Triggers if any upstream API (LinkedIn, ATS, GitHub) unavailable >15 min
Healthy
All 4 sources responding
Volume Anomaly Breaker
Triggers if rejection rate exceeds 3x normal rate in a 4-hour window
Healthy
Current rate: 0.8x normal

Recent Agent Activity

Last 24 hours across all agents
8:10 AM
DE
Decision Enrichment enriched Marcus Rivera profile for Acme ML Engineer role
Growth trajectory 94% · Cultural alignment 72% · Stability risk flagged · Career arc: upward
Score
8:06 AM
CG
Compliance Guardian cleared Marcus Rivera decision - no bias flags
4/5ths rule: Pass · Proxy variables: None detected · Counter-factual: Within bounds
Clear
7:48 AM
SS
Signal Scout detected flight risk - Marcus Rivera LinkedIn activity surge
Headline change + 3 recruiter connections + 340% profile view increase
Alert
7:32 AM
PM
Pipeline Monitor flagged TechFlow pipeline stall - 3 candidates idle for 5+ days
Escalated to Sarah Chen for review · Priority: Medium
Flag
7:15 AM
CG
Compliance Guardian issued bias drift alert - Engineering pipeline gender disparity
Selection rate: Women 62% vs Men 84% · Approaching 4/5ths threshold (74%)
Bias Alert
6:50 AM
PM
Pipeline Monitor flagged rate competitiveness gap - TechFlow Sr DevOps at $70/hr vs $76/hr market median
34% acceptance probability at current rate · Recommend $73–76/hr range · 3 candidates declined similar placements this quarter
Rate Alert
6:30 AM
PM
Pipeline Monitor detected Acme Corp scope change language in client communication
Flagged for Gavin Parker · Cross-referenced with dual-signal Placement Pulse
Flag
6:02 AM
SS
Signal Scout completed daily market scan - Austin ML Engineer market update
127 active postings · 1.4 roles per candidate · Median $80/hr · Supply tightening
Scan
Bias Monitoring & Compliance · David Park, VP Recruiting Ops
Monday, March 2, 2026

1 active alert · Next audit: March 15

Every decision that flows through the Glass Box generates audit data. This dashboard aggregates that data into the evidence base regulators and auditors need - selection rates, proxy variable correlations, and compliance readiness.

Glass Box Coverage
100%
Every AI recommendation passes through Glass Box
Active Alerts
1
Engineering pipeline · Gender disparity watch
Regulatory Frameworks
3
NYC LL 144 · Illinois AIPA · EU AI Act
Circuit Breakers
0
No agents paused · All thresholds healthy
POC Note: This dashboard shows what TalentPilot monitors in production. Selection rate data requires integration with the ATS (Bullhorn) for demographic categories. Proxy variable correlations are computed from decision patterns flowing through the Glass Box. The example below illustrates a real scenario - a gender disparity flagged in an engineering pipeline - using representative data.

Selection Rate Monitor

4/5ths Rule - Selection rate for any group must be ≥ 80% of the highest group's rate
Gender ⚠ Watch
Men
84% Highest
Women
62% 74% ratio
Non-binary
78% 93% ratio
⚠ Women selection rate at 74% of men's rate - approaching 4/5ths threshold (80%). Engineering pipeline most impacted. Compliance Guardian monitoring.
Race / Ethnicity ✓ Monitored
Tracks selection rates across EEOC categories using demographic data from the ATS. Alerts when any group's rate falls below 80% of the highest group.
Data source: Bullhorn ATS (EEOC reporting categories)
Methodology: 4/5ths adverse impact ratio, rolling 90-day window
Categories: White, Black, Hispanic, Asian, Two or More, Not Disclosed
Populated once ATS integration is active and sufficient decision volume is reached.
Age Group ✓ Monitored
Tracks selection rates across age bands. Particularly important because "Stability Indicators" scoring factor correlates with age at 42% - see proxy variable analysis below.
Data source: Bullhorn ATS (candidate profile data)
Methodology: 4/5ths adverse impact ratio, rolling 90-day window
Bands: 25-34, 35-44, 45-54, 55+
Populated once ATS integration is active. Proactive monitoring given Stability → Age proxy correlation.

Proxy Variable Analysis

Factors that correlate with protected characteristics - flagged above 35% correlation
Scoring Factor Gender Race Age Status
Technical Skills (30%) 8% 12% 15% ✓ Clean
Growth Trajectory (20%) 5% 7% 22% ✓ Clean
Experience Relevance (20%) 11% 9% 28% ✓ Monitor
Cultural Alignment (15%) 24% 31% 14% ⚠ Approaching
Stability Indicators (15%) 9% 6% 42% ⚠ Flagged
⚠ Stability Indicators → Age: 42% correlation detected. When recruiters increase "Stability" weight above 25%, it functions as an age proxy. Counter-factual panel warns recruiters in real-time. This is the exact pattern the Glass Box bias check catches on the decision page.

Counter-Factual Analysis

Aggregated from individual Glass Box counter-factual panels
◈
What it measures
"What would change if we removed this factor?"
📊
Data source
Every Glass Box decision generates counter-factuals
⚠
Flag triggers
Score shifts >5% when a proxy variable is removed
✓
Closes the loop
Tracks whether recruiters adjust after seeing flags
How it works: On the Glass Box decision page, the counter-factual panel shows what happens to a candidate's score when individual factors are removed. This dashboard aggregates those results across all decisions to identify systemic patterns.
Key pattern detected: When recruiters set "Stability Indicators" weight above 25%, it functions as an age proxy - scores shift meaningfully for 45+ candidates. The Glass Box warns recruiters in real-time.
Why it matters: Individual counter-factuals catch bias at the decision point. Aggregated counter-factuals reveal whether bias is systemic across the organization - the exact evidence auditors need.

Learning Loop Health

Model accuracy and retraining status
Model Accuracy (30-Day Trend)
Week 1
91%
Week 2
89%
Week 3
92%
Week 4
90%
Circuit breaker threshold: 75% · Current: 90% - healthy
Retraining Status
Decision EnrichmentLast: Feb 28 · Next: Mar 14
Pipeline MonitorLast: Feb 25 · Next: Mar 11
Signal ScoutLast: Mar 1 · Next: Mar 15
Compliance GuardianRules-based · No retraining needed
All models within accuracy thresholds · No circuit breakers triggered

Audit Readiness

Compliance status by regulatory framework
NYC Local Law 144
96%
✓ Annual bias audit conducted
✓ Summary of results published on website
✓ Notice to candidates (10 business days)
✓ Selection rate data by race/ethnicity and gender
◆ Independent auditor review - scheduled Mar 15
Next audit: March 15, 2026 · 13 days
Illinois AI Policy Act
92%
✓ Candidate notice of AI use in hiring
✓ Opt-out mechanism documented
✓ Data retention policy (7 years)
◆ Annual impact assessment - due April 1
Next assessment: April 1, 2026 · 30 days
EU AI Act (High-Risk AI)
88%
✓ Risk classification documented (High-Risk)
✓ Human oversight measures in place
✓ Transparency requirements met (Glass Box)
◆ Conformity assessment - Q3 2026
◆ Technical documentation update needed
Conformity assessment: Q3 2026
TP
Welcome to TalentPilot
Decision intelligence for recruiting.
Not a replacement - a multiplier.
AI matching platforms tell recruiters who to consider. TalentPilot tells them why - and checks for bias, flags risks, and makes every decision transparent and auditable. It sits between AI matching and human judgment, making the space in between intelligent.
The 10-80-10 framework
Recruiters configure matching weights, set guardrails, and define what "good" looks like (first 10%). 4 AI agents scan signals, enrich candidate profiles, monitor pipelines, and check compliance (the 80%). Recruiters review through a Glass Box - full evidence chain, counter-factual analysis, and bias check - then decide (final 10%). Every decision is traceable and auditable.
◎
Recruiter View
Sarah's daily command center.
8 decisions, zero guesswork.
The Today View shows a senior recruiter's active decisions - each with a confidence score, urgency flag, and recommended action generated by specialized AI agents. The recruiter sees what needs attention, why it matters, and what to do - all before opening a single candidate profile.
What the AI agents surface
Flight risk detection with competing offer signals. SLA breach warnings before deadlines hit. Rate competitiveness analysis against real-time market data. Placement pulse monitoring for at-risk engagements.
Go to Today View →
◈
Glass Box Transparency
See exactly why the AI recommends it.
This is what nobody else does.
Open any decision and the Glass Box shows the complete evidence chain - which data sources informed the recommendation, what each dimension scored, and what would change if the inputs were different. The recruiter can adjust weights, override scores, and the system learns from every decision.
Glass Box features
Multi-source evidence chain with confidence scoring. Counter-factual analysis ("what if we weighted stability higher?"). Real-time bias check at point of decision. Weight calibration sliders with live score recalculation.
Go to Glass Box →
📡
Multi-Role Intelligence
Different roles, same intelligence layer.
Everyone sees what they need.
Recruiters see candidate decisions. Account managers see portfolio health and dual-signal monitoring — when a placed worker's self-assessment diverges from client feedback, TalentPilot catches it before the client calls. VP Ops sees aggregate pipeline health and compliance exposure. 4 specialized AI agents power all three views: Signal Scout detects risks, Decision Enrichment confirms matches, Pipeline Monitor tracks SLAs, and Compliance Guardian watches regulations.
4 AI agents, 3 stakeholder views
Signal Scout: Detects flight risk, competing offers, market shifts. Decision Enrichment: Confirms AI matches with beyond-the-resume intelligence. Pipeline Monitor: Tracks SLA breaches, stalled pipelines, velocity drops. Compliance Guardian: Watches LL 144, 4/5ths rule, IL HB 3773 deadlines. Each agent feeds all three stakeholder dashboards — recruiter, account manager, and VP.
Go to Client Health →
◆
Architecture That Transfers
Same pattern. Different domain.
Bias monitoring that nobody else does.
TalentPilot and its sister product TradePilot (supply chain decision intelligence) share the same core architecture: domain-agnostic decision framework, specialized AI agents, evidence-based transparency, and human-in-the-loop controls. What makes TalentPilot unique is native bias monitoring — not a quarterly report, but real-time checking at point of decision. Counter-factual analysis, proxy variable detection, and 4/5ths rule enforcement built into every recommendation.
What transfers + what's unique
Shared architecture: Decision queue with urgency ranking. Evidence chain with expandable detail. Cross-impact analysis. Agent fleet with independent configuration. Human override with learning loops. TalentPilot adds: Real-time bias monitoring at point of decision. Counter-factual analysis for every candidate. OFCCP compliance as a feature, not an afterthought. Glass Box transparency that no matching platform offers.