Hours back, every cycle
What used to take a manager half a day across spreadsheets now takes minutes. Same numbers, none of the copy-paste.
StackSense AI
Geocoding Operations Performance Evaluation Ranking (PER) Portal
We turn weeks of spreadsheet shuffling into a calm, single-source ranking workflow. AI drafts coaching plans, surfaces performance insights, lets auditors chat directly with the data, and gives managers forward-looking team predictions.
Replaces a manual workflow stitched across
ExcelQuipPython scriptsDataCentralIQTQuickSight
Why we built it
Performance ranking shouldn't be a monthly fire drill.
Today, ranking 300+ associates each cycle costs hours of manual Excel work, fragile copy-paste between tools, and inconsistent coaching narratives. StackSense AI turns the whole loop into one auditable system, so managers spend their time on people, not pivot tables.
One source of truth
Every score is traceable to its inputs and approvers. Versioned V1 → Final snapshots make tier changes obvious.
Consistent, fair scoring
The calc engine is pure TypeScript with locked formulas, so no two managers will compute a quality score differently.
Built around the workflow
Boosts, LP overrides, exclusions, and approvals model how teams actually work, not a generic HR tool retrofitted.
SLA-aware approvals
Pending approvals surface before they slip. Missed deadlines escalate so cycles never stall silently.
AI across the workflow
Coaching plans, performance insights, auditor chat, and team predictions. Scoring and tiering stay deterministic.
How AI shows up across the portal
Coaching, insights, conversations, and forecasts.
Scoring stays deterministic. Every number a manager sees is traceable to its inputs. AI takes on the harder parts: drafting coaching plans, generating performance insights for every associate, letting auditors chat directly with the data, and giving managers forward-looking team predictions.
coaching-plan · marcus.r · jun 2026
Coaching planTrigger
2 consecutive T5
Trend
T2 → T3 → T4 → T5
Driver
Quality (-9.4%)
Summary
Productivity is steady at 11.2/hr, but counting defects rose sharply in May once critical-counting started flowing into effective defects. The drop is concentrated in addressable defect types, not throughput.
Suggested actions
- Pair with a T1 peer for two audit-deep-dives this cycle.
- Re-take the critical-counting refresher; verify in week 2.
- Weekly 1:1 with focus on the top 3 defect categories.
Draft · awaiting manager edit2.1s · 1,284 tokens
AI chat · Q2 2026
Chat with dataWhy did Marcus's quality slip in May?
Critical-counting started flowing into effective defects on May 4. His counting defects rose +3.2/mo. Productivity unchanged.
How does he compare to T1 peers?
T1 peers average 1.1 effective defects/mo. Marcus is at 4.3. Gap is 100% counting-driven.
Ask anything about a person, a tier, a cycle…↵
Team insights · L4 sandeep.k
ForecastTier distribution
T1
T2
T3
T4
T5
Headcount
316
At-risk next cycle
12
Predicted T5 next cycle
- marcus.r87% conf
- priya.k74% conf
- devon.t61% conf
Coaching plans for at-risk associates
When an associate hits a real signal, AI drafts a humane plan grounded in six months of their own productivity, quality, and utilization data.
Performance insights for every associate
Plain-language summaries of trends, drivers, and outliers so managers walk into 1:1s already knowing what to talk about.
Chat with performance data for auditors
Auditors can ask questions in natural language and get traceable answers backed by the same locked metrics shown in the portal.
Team insights and forward predictions for managers
Roll-ups, risk forecasts, and next-cycle tier projections so managers can act before a slip becomes a pattern.
Manager stays in the driver's seat
Every plan and insight is a draft. Managers edit, approve, or reject before anything is finalized.
Fully auditable
Every prompt, response, model, latency, and token count is persisted, so any output can be debugged, replayed, or explained.
One portal. Every number explainable.
Sign in with your StackSense AI account to view your team's rankings, submit boost & LP scores, or review pending approvals.