🧾
Persona · Billing
Billing Specialist
Oversees autonomous billing agents and ensures billing logic and policies are correctly applied across every invoice generated. Focuses on billing accuracy, policy compliance, and anomaly resolution — not invoice keying.
◉ 92% autonomous
Dashboard · Billing Agent Control Room
Autonomous Billing Performance
Updated 4 min ago
Invoices Generated
41,820
MTD · 93% straight-through
Billing Accuracy
99.6%
▲ +0.8pp vs LQ
Policy Violations
42
Of 41,820 · 0.1%
Avg Gen Time
2.1s
Per invoice · agent-driven
Leakage Prevented
$1.4M
Caught at billing, not close
Billing Agent Exception Queue42 open
InvoiceCustomerPolicyAmountStatus
INV-284419Acme FoodsTiered volume > threshold$84,200Review
INV-284420North Bev CoContract minimum shortfall$14,700Hold
INV-284421Pacific RetailRebate accrual mismatch$22,180Review
INV-284422GlobalCo EMEAFX rate tolerance > 2%€9,420Hold
INV-284423Summit SpiritsPromo price window expired$6,300Agent fixed
Policy Application Mix
Contract pricing tiers98.4%
Rebate & discount logic96.1%
Tax determination99.2%
Multi-currency / FX94.6%
Freight & surcharge91.3%
Agent Skills · Low-Code Foundation
Billing Specialist — Skills Library
Each skill is a reusable, low-code building block for the Billing Agent. Skills are composable, governed, and owned by the Billing Specialist.
◉ 6 skills
invoice.audit
Audit
Inspects every generated invoice against the contract pricing table, promo calendar, and rebate accruals. Flags deviations > configurable tolerance.
Input: invoice JSONOutput: issue list
policy.enforce
Logic
Applies billing policies from a versioned rule library: contract tiers, minimums, surcharges, tax codes. Returns pass / hold / review decision.
Trigger: pre-postSLA: < 3s
anomaly.detect
ML
Scores invoices for unusual combinations (price × volume × customer × period) using trained pattern models. Catches issues rule-based audits miss.
Model: XGBoostThreshold: tunable
exception.resolve
Workflow
Routes held invoices to the right reviewer with recommended fix, backing data, and evidence trail. Auto-releases on approval.
Human-in-loopAudit trail
leakage.recover
Finance
Identifies missed charges (freight, surcharges, minimum commitments) and generates corrective invoices or credit notes automatically.
Runs: dailyValue: tracked
billing.narrate
GenAI
Generates plain-language explanation of every invoice for customer CX and internal reviewers — line by line, with source references.
Model: LLMMultilingual
📈
Persona · Revenue Assurance
Revenue Accountant / Revenue Assurance Analyst
Monitors AI-driven revenue integrity dashboards and resolves high-judgment revenue exceptions. Closes the books faster by catching recognition errors at source, not at month-end.
◉ ASC 606 / IFRS 15
Dashboard · AI-Driven Revenue Integrity
Revenue Integrity Monitor
Period: FY26 Q1
Revenue Booked YTD
$742M
Auto-recognized 96.8%
First-Pass Accuracy
98.4%
▲ +6.2pp YoY
Open Exceptions
58
23 high judgment
Close Cycle
2.6d
From 6.2d · target 2d
Revenue Integrity Trend · 6-month▲ Improving
Nov94.1% · $118M
Dec95.3% · $131M
Jan96.7% · $122M
Feb97.5% · $127M
Mar98.1% · $130M
Apr MTD98.4% · $114M
High-Judgment Exceptions23 open
TypeCountValue
Variable consideration9$2.8M
Contract modification6$1.4M
Multi-element allocation4$840K
Principal vs agent2$520K
Cut-off & timing2$310K
Revenue Leakage — Detected & Recovered$4.8M YTD recovered
SourceCountExposureRecoveredStatus
Unbilled fulfilled orders127$1.6M$1.6MClosed
Missed contract escalators42$1.2M$1.1MIn progress
Underbilled usage overages89$980K$950KClosed
Missed rebate true-ups21$640K$580KIn progress
Expired promo continued18$420K$420KClosed
Agent Skills · Low-Code Foundation
Revenue Accountant — Skills Library
Skills that power the Revenue Assurance Agent. Each skill is a reusable low-code building block with defined inputs, outputs, and audit trails.
◉ 6 skills
revenue.recognize
ASC 606
Applies the 5-step model: identify contract, performance obligations, price, allocate, recognize. Returns recognition schedule per contract line.
Input: contractOutput: schedule
variance.triage
ML
Ranks revenue variances by materiality × judgment × risk. Routes low-judgment items to auto-resolve, high-judgment to the analyst.
Priority queueExplainable
contract.interpret
GenAI
Extracts performance obligations, payment terms, consideration clauses from contract PDFs. Structured output ready for rev rec engine.
Model: LLMHuman review
leakage.scan
Finance
Cross-joins fulfillment, billing, contract, and usage data to find revenue that was earned but not billed. Produces recovery queue.
Runs: nightlyValue: tracked
close.accelerate
Workflow
Pre-populates month-end journals, flux explanations, and disclosures. Reduces analyst prep from days to hours.
Trigger: BD-1Audit ready
disclosure.draft
GenAI
Generates disaggregated revenue disclosures, remaining performance obligations, and contract balance tables with source citations.
Audit trailSOX ready
🛡️
Persona · Credit & Collections
Credit & Collections Analyst
Guides AI-generated credit and collection strategies and intervenes only in high-risk customer situations. Moves from dunning caller to risk strategist and negotiator.
◉ 1.1K accounts
Dashboard · Customer Risk & Collection Strategy
Credit Portfolio Overview
1,124 customers · $496.5M AR
DSO
49.4d
Target 42d · ▲ 2.1d MoM
Past Due
$213M
43% of total AR
High-Risk Accts
38
Requires analyst review
Agent Auto-Actions
2,147
MTD · $18.6M recovered
Bad Debt Reserve
$8.4M
1.7% of AR · stable
AI-Generated Collection Strategy
Customer Segments & Recommended Action
Low Risk · Strategic847 accts
Autonomous agent handles end-to-end. Gentle reminders, auto-reconcile, no analyst touch.
Auto-remindersSelf-service portalRelationship preserve
$214M
AR balance · 93% current
Medium Risk · Engaged239 accts
Agent proposes strategy; analyst approves cadence and payment plans for borderline accounts.
Tailored plansMulti-channelAgent + analyst
$189M
AR balance · 21% past due
High Risk · Analyst-Led38 accts
Human analyst leads. Negotiation, credit holds, litigation referral. Agent prepares the brief.
Direct negotiationLegal reviewExecutive escalation
$93M
AR balance · 71% past due
High-Risk Customer Queue · Analyst Action RequiredTop 6
CustomerPast DueRisk ScoreAgent RecommendationAnalyst Note
Atlantic Distributors$4.2M92Hold future orders · legal demand letterCall scheduled
Meridian Foods Intl$2.8M88Payment plan · 90 days · partial releaseApproved
Coastal Beverages$1.6M76Escalate to customer CFOIn progress
Northland Retail Grp$1.2M71Factor receivable · reduce exposureReview
Vista Spirits LLC$940K85Bankruptcy watch · reserve 60%Legal notified
Rainier Supply Co$780K68Restructure terms to net 30Pending
Agent Skills · Low-Code Foundation
Credit & Collections Analyst — Skills Library
Skills that power the Credit & Collections Agent. The analyst configures, trains, and supervises — the agent runs the volume work.
◉ 6 skills
risk.score
ML
Produces 0–100 credit risk score per customer using payment history, industry signals, bureau feeds, and macro indicators. Refreshed daily.
Sources: 12SHAP explain
dunning.orchestrate
Workflow
Runs the right dunning cadence per customer segment (email, SMS, call, portal nudge) with AI-chosen channel and timing.
Multi-channelA/B tested
payment-plan.propose
Finance
Designs payment plans customers will actually keep — balances amount, cash flow, and commercial relationship. Analyst approves.
Human-in-loopOptimizer
escalation.brief
GenAI
Prepares executive briefing on any high-risk customer — timeline, touchpoints, exposure, recommended action — in under 60 seconds.
Model: LLMOne click
credit-limit.recalibrate
Policy
Reviews credit limits quarterly using current risk score + usage. Proposes increase, hold, or decrease with rationale.
Runs: quarterlyAuto-exec < 5%
dispute.prevent
Predictive
Predicts which invoices will be disputed before they are — based on line, customer, amount pattern. Flags for proactive outreach.
Precision: 0.84Preempt
💰
Persona · Cash Application
Cash Application Specialist
Supervises autonomous cash-matching processes and manages only the rare exceptions requiring judgment. Remittance interpretation, multi-invoice matching, and deduction research — all driven by agents.
◉ 93% touchless
Dashboard · Cash Matching Control
Incoming Payment → Application Flow
Today: $14.2M · 2,417 payments
◉ Incoming Payment
2,417
$14.2M · ACH, wire, check, card
→
◉ Auto-Matched
2,248
93.0% touchless · avg 0.8s
→
◉ Matching Exceptions
169
7.0% · 28 analyst-escalated
Touchless Match Rate
93.0%
▲ +21pp vs baseline
Unapplied Cash
$1.8M
From $8.4M · −79%
Exception Avg Age
6.4h
From 3.2 days
FTE Redeployed
8 of 12
Moved to analysis work
Exception Types · Today169 open
ExceptionCountValueTypical Cause
Short-paid · deduction58$412KChargeback · OS&D claim
Over-paid12$48KDuplicate payment
No remittance / unclear47$684KWire w/o reference
Multi-invoice lump sum31$1.1MConsolidated check
Customer mismatch14$186KParent vs subsidiary
FX / bank fee variance7$52KIntermediary fees
Channel Match Performance
Lockbox · check88.1%
ACH · CTX with remit98.4%
Wire · w/ reference91.6%
Wire · w/o reference62.4%
Card · gateway99.7%
Customer portal99.2%
Agent Skills · Low-Code Foundation
Cash Application Specialist — Skills Library
Skills that power the Cash Application Agent. The specialist reviews confidence-scored exceptions and teaches the agent new match patterns.
◉ 6 skills
remittance.interpret
GenAI
Reads remittance advice in any format (PDF, email, EDI 820, portal screenshot) and extracts invoice numbers, amounts, deductions.
Model: Vision LLMMultilingual
match.multi-invoice
Algo
Solves the subset-sum problem to split a lump payment across multiple invoices with high confidence. Explains every choice.
ILP solverExplainable
deduction.classify
ML
Classifies short-pays into OS&D, promo, pricing, unauthorized, tax — and routes to the right owner (AR, sales, logistics, trade).
12 classesRoute auto
customer.reconcile
Graph
Resolves parent/subsidiary/branch relationships using a customer graph. Matches payments from the right legal entity.
Neo4jFuzzy name
unapplied.clear
Workflow
Drives unapplied balances to zero. Correlates with customer inquiries, reaches out automatically, proposes reversals or corrections.
Runs: dailyTarget: < 1%
pattern.learn
Feedback
Captures specialist fixes as new match rules. The agent gets smarter every day — without code changes.
Active learningLow-code
📜
Persona · Contract & Pricing
Contract & Pricing Analyst
Designs and governs the pricing and contract rules that AI agents use to configure offers and interpret terms. The rule-setter — not the rule-executor.
◉ 247 active rules
Dashboard · Rule Library & Offer Configuration
A-Spec Configuration Pipeline
247 rules · 84 offers
◉ Step 01
Categorical Parameters
Industry, region, tier, channel, volume band — the inputs that drive pricing logic.
→
◉ Step 02
Logic
Rules, decision tables, and waterfall pricing — authored in low-code, versioned, and tested.
→
◉ Step 03
A-Spec Configuration
Agent-ready offer specification — deployed to the Billing, Order, and CPQ agents.
Active Rules
247
Across 84 offers
Rule Coverage
98.6%
Of deal volume
Offers This Quarter
18
New · 12 revised
Time-to-Deploy
2.1d
From 21d · −90%
Rule Library Activity · 30dGoverned
RuleTypeScopeUsedStatus
Tier 1 Enterprise — GlobalWaterfallVolume > 10k4,218Active
Rebate Accrual — BeveragesAccrualIndustry1,864Active
FX Protection — EMEA EURSurchargeRegion1,412Active
Retail Promo Q2-26DiscountChannel984Expires 30d
New Customer Intro KitBundleAll642Active
Sustainability SurchargeAdd-onAll318Draft
Offer Configuration Status
Production · live71
Staged · testing9
Draft · authoring4
Archived last 30d6
Pending review3
Agent Skills · Low-Code Foundation
Contract & Pricing Analyst — Skills Library
Skills used by the analyst to author, govern, and deploy rules. The agent consumes the rules; the analyst curates them.
◉ 6 skills
rule.author
Low-code
Visual rule editor — categorical parameters → logic → outcome — with natural language preview and dry-run simulation on historical deals.
No-code UIVersioned
contract.extract
GenAI
Extracts pricing terms, rebate schedules, and SLAs from contract PDFs into structured rule candidates for analyst review.
Model: LLMPDF / DOCX
offer.configure
A-Spec
Assembles rules into agent-ready offer specs (A-Spec). Deploys to Billing, Order, CPQ agents with single publish action.
A-Spec JSONOne-click
impact.simulate
Analytics
Replays a proposed rule change against 12 months of history — shows margin, volume, and customer impact before go-live.
What-ifHistorical replay
governance.review
Workflow
Enforces approval chains by rule type, value, and risk. Keeps audit trail of every rule change and approver for SOX.
SOX readyMulti-approver
rule.lint
Quality
Detects rule conflicts, dead branches, and unreachable conditions before publish. Prevents pricing incidents in production.
Static analysisBlock publish
🤝
Persona · Customer Finance
Customer Finance Manager
Uses predictive insights from AI to prevent disputes and proactively improve customer financial experiences. Moves finance from back-office function to customer relationship partner.
◉ Predictive-first
Dashboard · Predictive Customer Financial Insights
Prevention & Experience Monitor
Top 100 customers · 72% of AR
◉ Prevent Disputes127 predicted
−64%
Dispute rate reduction YoY
AI flags invoices likely to be disputed — wrong ship-to, unclear line item, customer payment pattern — before they are sent.
◉ On-Time Payment94% predicted paid
+12pp
Improvement in on-time rate
Customer-specific payment behavior forecasting. Proactively intervenes with invoices at risk — before they go late.
◉ Fraud & Risk Reduction$2.3M avoided
−38%
Fraud exposure reduction
Anomaly detection across payments, remittances, and bank changes. Stops social engineering attacks before funds leave.
Customer NPS (finance)
+62
▲ +18 YoY
Disputes Prevented
127
MTD · $1.9M value
Proactive Interventions
84
Calls / emails auto-triggered
Customer Effort Score
4.6/5
Finance touchpoints
Proactive Intervention Queue · This WeekPredictive
CustomerPredicted EventConfidenceRecommended ActionStatus
Harbor DistributorsDispute on INV-284419 (price)0.91Pre-call customer finance lead · explain price changeScheduled
Summit BeveragesLate payment risk · 14 days0.78Offer early-pay discount 1%/10 net 30Proposed
Pacific Foods IntlChargeback on promo claim0.86Send settlement proposal ahead of claimSent
Lakeside RetailCredit limit strain in 30d0.72Offer reviewed terms + quarterly true-upPending
North Star SpiritsBank detail change · verify0.94Callback on file number · block until verifiedVerify
Agent Skills · Low-Code Foundation
Customer Finance Manager — Skills Library
Skills that power the Customer Finance Agent. The manager shapes the strategy; the agent runs the predictive signals at scale.
◉ 6 skills
dispute.forecast
Predictive
Predicts invoices that will be disputed using customer, line, amount, channel, and timing features. Routes for proactive outreach.
Precision: 0.84Pre-send
payment.forecast
ML
Predicts per-invoice pay date per customer. Feeds DSO forecast and cash flow planning. Flags at-risk invoices for intervention.
MAE: ±2.4 daysPer invoice
experience.nudge
CX
Sends the right communication at the right moment: early reminder, payment option, statement clarification, or thank-you — no spam.
Multi-channelPersonalized
fraud.detect
ML
Spots social engineering (bank change asks, vendor impersonation) and payment anomalies using graph + sequence models.
Graph MLReal-time
journey.map
Analytics
Maps every finance touchpoint with each customer — billing, payment, dispute, dunning. Highlights pain points for journey redesign.
Per customerTime series
relationship.brief
GenAI
Generates 1-page customer finance briefing before every key call: exposure, trends, issues, wins, talking points.
Model: LLMOn-demand
🎯
Persona · O2C Leadership
AR Manager / Director of Order-to-Cash
Orchestrates end-to-end autonomous O2C operations, focusing on policy, risk management, and performance optimization. Owner of the agentic O2C platform and outcomes.
◉ End-to-end O2C
Dashboard · O2C Control Tower
End-to-End Performance & Orchestration
All 7 agent domains live
DSO
29.1d
▼ −20.3d YoY
Revenue Integrity
98.4%
First-pass · ASC 606
Touchless O2C
85%
End-to-end autonomous
Bad Debt
0.31%
Of revenue · −41% YoY
Annual O2C Value
$48M
FY target · 312% ROI
Agent Domain Performance
Each Persona, Each Agent — One View
DomainOwner (Persona)AutonomyExceptionsValue YTDHealth
Contract & PricingPricing Analyst98.6%3$4.1M● On track
BillingBilling Specialist93.0%42$6.8M● On track
Revenue AssuranceRevenue Accountant96.8%58$9.2M● On track
Credit & CollectionsC&C Analyst81.2%38$11.4M◐ Monitor
Cash ApplicationCash App Specialist93.0%169$7.6M● On track
Customer FinanceCustomer Fin Manager89.4%84$8.9M● On track
End-to-End O2CDirector O2C85.1%394$48M● Green
Policy & Risk PostureGoverned
Credit policy coverage99.2%
SOX controls · pass100%
Agent action audit trail100%
Concentration risk < 15%11.8%
Reserve adequacy1.7%
Performance Optimization PlaysNext 90 days
PlayDomainExpected Value
Segment dunning by industryCollections+$2.1M
Early-pay discount 1%/10Customer Finance+$1.4M
Remit learner auto-trainCash App+$0.9M
Rev leakage sweep · EMEARevenue+$1.6M
Rebate true-up agent v2Billing+$1.2M
Agent Skills · Low-Code Foundation
AR Manager / Director O2C — Skills Library
Skills used by the O2C leader. Orchestration, governance, and performance skills — consumed across every persona domain.
◉ 6 skills
orchestrator.route
Control
Routes O2C events between agent domains with SLAs, retries, and escalation paths. Single backbone for the agentic operation.
Event-drivenSLA tracked
policy.govern
Governance
Enforces policy across all persona domains — credit, pricing, revenue, collections. Versioned, approved, auditable.
CentralSOX ready
kpi.forecast
Analytics
Projects DSO, bad debt, revenue integrity 13 weeks forward. Simulates the impact of policy or process changes.
13-weekMonte Carlo
agent.observe
Telemetry
Monitors every agent action — autonomy rate, exception rate, latency, drift. Alerts leader when a domain degrades.
Real-timeAlerting
value.track
Finance
Tracks the realised value of every agent skill and play. Ties dashboards back to P&L, DSO, and working capital impact.
Audit-linkedReal dollars
exec.brief
GenAI
Generates the monthly O2C executive briefing — performance, risks, plays, asks — with source citations and a one-page summary.
Model: LLMBoard-ready