Practical tools for operations leaders making decisions at scale.

Decide which manual review queues to scale, which to optimize, and which need root-cause investigation. V2 adds false positive cost (LTV impact of blocking legit orders) and portfolio mode (allocate the next marginal agent across multiple queues).

Detect BIN-level attack patterns before they hit chargeback. Combines volume velocity, new-user clustering, and CBK rate trending into a single dashboard with adjustable thresholds and a 7-day drill-down per BIN. Includes a Python reference implementation.

Surge pricing using PID control. V2 adds polygon-level optimization — a 3×3 city map (Airport, Downtown, Stadium, Residential, Industrial, Suburbs, etc.) showing per-zone surge state with a time scrubber across 24h. Nine independent PIDs (one per zone) vs a single city-level PID, because city-average surge hides geographic disparity.

Walk through a chargeback like a senior fraud analyst. Interactive decision tree wizard guides you from reason code to action recommendation in 4-6 questions, with the evidence package ready. Plus the 8-section SOP structure any team can adapt.

How many agents do you actually need? Multi-case-type forecast based on volume × complexity × productivity × shrinkage. Compares 100% in-house vs vendor-split scenarios with cost comparison. Four pre-loaded operational presets.

Translate a marketplace forecast into required couriers, vehicles, sorters, and dock capacity, with day-over-day backlog cascade, storage occupation tracking, origin mix decomposition, and operational levers (boost capacity, defer origin). Two personas (3PL last-mile + fulfillment warehouse), two granularities (daily + intraday hourly+zone), side-by-side operational + financial output, sensitivity at +/- 15% and +/- 30%.

Two questions every ops leader asks at standup, fused into one tool: are we behind on SLA right now (backlog health monitor with hours-to-breach, agents-needed-to-recover, recovery cost) and what should the agent pick next (queue prioritization with risk × age × value scoring, three reprioritize modes, top-5 highlight). Four operator presets for fraud, content moderation, disputes, CS escalations.

Most auto-escalate AI tools skip the taxonomy. They surface alerts but cannot say which deserves a human, which should wait, and which would self-resolve in four more hours. This tool maps exception type, alert age, SLA budget, customer impact, volume, and dollar exposure into a single recommendation: auto-resolve, wait, route to human review, or page on-call. Includes a recommendation-across-alert-age curve, cost-of-escalating vs cost-of-waiting comparison, and a per-tier action checklist. Five operator presets covering stuck shipments, payment failures, partner non-response, quality drift, and fraud-flagged transactions.

Quantify the dollar leak from preventable customer feedback. The number you bring to the executive conversation when the question is fix-it-now vs absorb-it-forever. Monthly leak, annualized leak, payback period, 12-month NPV, 3-year cumulative savings, Pareto category breakdown, sensitivity at +/- 15% and +/- 30%. Two personas (Operations Manager + CFO / Finance), 5 operator presets for marketplace, fintech, last-mile, DTC, and SaaS onboarding shapes.

The structured one-pager every new operations leader needs. Generate the 30-60-90 day plan mapped to JD priorities: executive summary, per-priority deliverables across discover / decide / deploy phases, north star metrics, risk register, asks to the leader. Output as Markdown for direct paste into your follow-up email or executive memo. Five operator presets covering Series A marketplace, scale-up CPG, mature marketplace supply, fintech fraud, and last-mile logistics.

Rank vendors against a weighted multi-dimensional scorecard (on-time delivery, defect rate, cost variance, lead time, response time). A/B/C tier classification with sensitivity analysis on weights, action recommendations per vendor, and money-on-the-table math. Four operator presets for kit sourcing (overseas manufacturers), 3PL last-mile carriers, BPO offshore support, and balanced multi-vendor portfolios.

Rank drivers and couriers on weighted multi-dimensional performance with tenure-based bias correction so new drivers aren't unfairly penalized by thin data. Four-tier classification (Top / Reliable / Watch / At-Risk), best-gets-first-call simulation showing which tier absorbs the next 100 orders, turnover risk surfacing, and per-driver action recommendations. Four presets for instant delivery, gig couriers, ride-hail, and 3PL last-mile.