Energy AI agent · Billing & tariff
Billing & Tariff Support AI Agent
Retrieve the bill and the filed tariff sheet that priced it, explain the charge against that sheet, surface the escalation route, and leave the arrangement itself to a billing adviser.
- Typical delivery
- 4–6 weeks
- Deployment
- Your stack
- Adviser review
- Tariff-cited
- After launch
- Agent Care
What this agent does
Explains the bill, without repricing it
In
- 01
Ingesting the account, the disputed bill and the tariff schedule in effect from supported billing, CIS or meter-data sources.
- 02
Normalising rate codes, meter reads and charge lines, and carrying each one forward with its source.
Reason
- 03
Explaining each charge against the filed sheet that priced it, line by line.
- 04
Applying the jurisdiction, rate class and tariff version configured for that account.
- 05
Citing the tariff sheet and its effective date rather than paraphrasing a rate.
Decide
- 06
Flagging a meter doubt, a high-bill enquiry or a hardship signal as a dispute, not an enquiry.
- 07
Routing arrangements, adjustments and disconnection questions to a billing adviser.
Out
- 08
Retaining the enquiry, the sheet cited, the adviser's answer and the escalation offered against the account.
- 09
Executing write actions only inside the approval boundaries agreed during implementation.
Product statement
The agent explains and cites; a billing adviser decides what is charged, adjusted or arranged, and the utility answers to the regulator for the tariff.
Example workflow
One enquiry, bill to answer
- 1Bill enquiry receivedPhone, web chat, email or a callback from the billing queue
- 2Bill and tariff retrievedThe billed period, the meter reads, the rate class and the sheet in effect for it
- 3Explanation draftedCharge lines, the sheet cited and confidence
- 4Controls appliedTariff-citation checks, jurisdiction rules, dispute-signal flags and confidence threshold
No human action required
Stages 1 to 4 run unaided, and nothing is charged, adjusted or arranged at any of them — the adviser's lane opens at the confidence gate.
Goes to the billing adviser to release.
Adds a tariff-desk read first.
Adviser review
The answer is held with the sheet it cites, its dispute flags and the confidence.
Approve · Correct · Send to tariff desk
Approved — released to the customer
- 6Billing record updatedOnly where write access and approval policy allow it
- 7Outcome evaluatedCitation accuracy, adviser corrections, dispute classification and repeat contacts
Corrections
Each adviser correction is counted in the evaluation.
What should not run autonomously
Human approval stays in control
Outside the boundary — human approval required
8 items
- Quoting a rate that is not on the filed sheet.
- Agreeing a payment arrangement outside the tariff.
- Disconnecting service, or restoring it afterwards.
- Deciding a medical-necessity or hardship claim.
Automation boundary
Agent acts unaided
- ✓Retrieve the bill and the filed tariff sheet that priced it.
- ✓Explain each charge against that sheet and its effective date.
- ✓Record a doubt about the meter or the bill as a dispute.
- ✓Surface the escalation route and hold the answer for an adviser.
Any write happens inside the boundaries agreed at implementation, never ahead of the adviser.
- Adjusting a bill, waiving a charge or issuing a credit.
- Issuing an estimated bill, or the true-up after it.
- Determining eligibility for an assistance programme.
- Changes to tariff mapping or escalation thresholds.
Example output
One bill enquiry, annotated
Everything the agent says is attached to the sheet it was read from.
Billing answer · single enquiry
Illustrative example- Account
- Residential electric
- Charge explained
- Summer usage crossing two tier thresholds
- Billed amount
- $318.42
- Tariff sheet cited
- Filed Schedule E-1
- Confidence
- 91%
- Escalation
- PUC complaint route stated
As received
Taken from the billing record and the tariff sheet in effect — nothing on this side is written by the agent.
Tariff basis used
- Filed rate schedule
- Meter reads of record
- Tier thresholds on the sheet
Why this answer
It cites the sheet that priced the bill rather than restating a rate.
What the score decides
Below the configured threshold the answer picks up a tariff-desk read before it.
Action
ApproveCorrectSend to tariff deskValue
Where AI adds value
The same four claims, placed at the point in the workflow where each one applies.
Where the value lands
Value 01 – 04
Every bill enquiry
From the billing record
Read the sheet that priced it
Draw on the filed tariff, the meter reads and the back-billing and notice rules configured for that jurisdiction.
Read the tariff, not around it
Routine charge explanations arrive already cited to a sheet.
Put advisers where the dispute is
Dispute signals and low-confidence answers are marked, so the adviser's read starts where the exposure sits.
Build an evidence trail
The answer, the tariff sheet behind it and the adviser who released it stay on the account.
Integrations
Typical integrations
Five system groups connect to the same agent. Which of them are in scope is decided in discovery.
Billing and CIS
- Oracle CC&B
- SAP IS-U
- Hansen
- Gentrack
Metering and usage
- MDMS
- AMI head-end
- Green Button exports
Tariff and rates
- Filed tariff library
- Rate engine
- rate schedules
Agent
Billing & tariff support
- Reads the bill
- Cites the sheet
- Holds for an adviser
Contact and workflow
- Salesforce
- Genesys
- Case and complaint systems
Observability & evaluation
- OpenTelemetry
- Langfuse
- Supported monitoring/evaluation sources
Integration availability depends on the client's existing systems and API access.
Agent controls
Six layers between the model and the customer
Every layer wraps the next. What none of them catches is in the map below.
L6 · Outermost — last line of defence
Inward → L1 · closest to the model
- L1 – L2
- Test whether an answer may stand
- L3
- Puts the release in an adviser's hands
- L4 – L5
- Keep the answer and the tariff sheet behind it
- L6
- Falls back to bill retrieval when signals degrade
How Nestack evaluates it
Evaluate the whole enquiry — not only the sentence the customer hears.
Coverage runs the whole depth of the workflow, and every layer is cut by slice.
Surface — the answer the customer hears
- E1Final-output evaluationDid every rate in the answer appear on the sheet cited?
- E2Step-level evaluationDid the agent use the right account, tariff version and jurisdiction rules?
- E3Tool evaluationDid it read the correct bill and the correct tariff schedule?
- E4Confidence calibrationDo low-confidence answers actually attract more adviser corrections?
- E5Slice evaluationHow does performance change across specific rate classes?
- E6Business outcomeHow many answers needed a correction after the customer had relied on one?
Floor — the outcome the utility answers for
Failure modes
Where each failure originates in the agent
Seven ways an answer goes wrong, by the stage it starts.
Agent lifecycle
Direction of processing →
01 · Retrieval
1 mode
- JK-03
Superseded tariff sheet
A rate read from a schedule no longer in effect.
Stage gathers
Bill, meter reads, rate class and the tariff of record
02 · Reasoning
2 modes
- JK-04
Rate stated from memory
A rate is paraphrased instead of read off the filed sheet.
- JK-06
Escalation discouraged
The customer is steered away from a meter test or a complaint.
Stage proposes
The charge explanation, the sheet cited and confidence
03 · Tool / write
2 modes
- JK-02
Arrangement implied
An answer commits the utility to terms outside the tariff.
- JK-05
Dispute logged as an enquiry
A meter doubt is closed as resolved rather than raised.
Stage writes
Only where write access and approval policy allow it
04 · Output
1 mode
- JK-01
Rate not on the filed sheet
A quoted figure does not appear in the filed schedule.
Stage returns
The answer the adviser releases and
05 · Change / Version
1 mode
- JK-07
Silent tariff regression
A model or mapping change widens what the agent will assert.
Stage tracks
Model, prompt, tariff mapping and jurisdiction config
- Sev-1 · acts outside the boundary
- Sev-2 · a wrong rate reaches the customer
- Sev-3 · source degrades, enquiry routes to review
Affected slices
Overall accuracy can hide one bad cohort
Break the correction rate down by rate class before you trust it. A small number of billing cohorts absorb most of the rework, so Nestack reports the adviser-correction rate by slice, not only in total.
Slice performance — reported separately, not only in aggregate
Illustrative example| Slice | Failure rate | Lift | Lift vs. threshold | Status |
|---|---|---|---|---|
| Back-billed and make-up bills | 5.3% | 3.4× | Review | |
| Estimated-read accounts | 4.5% | 2.9× | Review | |
| Time-of-use and demand rates | 2.8% | 1.8× | Watch | |
| Standard single-rate bills | 1.1% | 0.7× | Normal |
Bar: adviser-correction-rate lift vs. single-rate baseline · scale 0–4.0× · tick marks the 2.0× review threshold
2 of 4 slices over threshold
Evidence-linked improvement
A cycle ends in a test, not an apology
The loop shuts when the miss is a case in the suite, not when it has been explained. That suite is what the next bill enquiry answered is measured against.
Improvement cycle · five stages
Switchback — the path turns at Improve and returns at Learn
- 01
Detect
Adviser-correction rate rises in a rate class.
- 02
Diagnose
Pull the enquiries and the sheets cited behind them until the cause narrows to one.
- 03
Improve
Whatever changes ships against a version, with the enquiries that prompted it attached.
- 04
Verify
Nothing ships until the affected cases pass a second time.
- 05
Learn
The suite grows by one case; so does the tariff mapping.
Learn → Detect
The return edge. The next detection is measured against a longer suite than this one.
Typical build scope
Twelve workstreams across six weeks
The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, enquiry workflow, evaluation, integration, then production validation and handover.
Workstream
Week 1Week 2Week 3Week 4Week 5Week 6
- 01Billing workflow discovery and automation-boundary definition.Week 1
- 02Billing, CIS and meter source assessment.Week 1, Week 2
- 03Tariff mapping, rate classes and jurisdiction rules.Week 1, Week 2
- 04Bill and tariff-sheet retrieval.Week 2
- 05Explanation logic and citation binding.Week 2, Week 3
- 06Confidence scoring and dispute routing.Week 3
- 07Adviser review workflow.Week 3, Week 4
- 08Billing-system and CIS integration.Week 5
- 09Rate-citation regression cases.Week 4, Week 5
- 10Guardrails and escalation controls.Week 4
- 11Enquiry-trail instrumentation.Week 5, Week 6
- 12Deployment, documentation and Agent Care handover.Week 6
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel
Final scope and sequence confirmed in discovery
Engagement tiers
What each tier includes
Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.
| Capability✓ in scope · — not at this tier | PilotOne rate class, one jurisdiction | ProductionProduction billing systems | AdvancedMultiple jurisdictions / tariffs |
|---|---|---|---|
| Introduced at Pilot | |||
| Explanation cited to the filed sheet | ✓ | ✓ | ✓ |
| Adviser release | ✓ | ✓ | ✓ |
| Citation-accuracy baseline | ✓ | ✓ | ✓ |
| Introduced at Production | |||
| Reporting by rate class | — | ✓ | ✓ |
| Adviser review in your systems | — | ✓ | ✓ |
| Approved write-back | — | ✓ | ✓ |
| Billing-system integration | — | ✓ | ✓ |
| Introduced at Advanced | |||
| Multi-state billing rules | — | — | ✓ |
| Multi-stage billing approvals | — | — | ✓ |
| High enquiry volume | — | — | ✓ |
| Multi-jurisdiction tariff controls | — | — | ✓ |
| Build price | From $5,000 | From $8,000 | Custom quote |
- Final build price
- Confirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
- Separate from build
- Build pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.
What we need from you
What you bring, and what we build with it
Each input maps to a piece of build scope and a week in the delivery timeline.
You bring
We build with it
01Your bill formats and the accounts behind them
→Bill and tariff-sheet retrieval and field mapping
Week 102Representative bill enquiries and the answers given
→Explanation baseline and tariff-sheet citation binding
Week 203Your filed tariff sheets and their effective dates
→Tariff mapping, rate classes and jurisdiction rules
Week 104Access to relevant APIs, feeds or exports
→Billing, CIS and metering assessment, then integration setup
Week 205Answers you would not want relied on
→Citation cases and failure-mode testing
Week 406Where an answer must stop and wait for an adviser
→Confidence scoring, dispute routing, guardrails and escalation controls
Week 307Named billing advisers to release answers
→Adviser review workflow, then pilot and production validation
Weeks 5–6
Nothing else is required
Deployment, documentation and Agent Care handover are ours.
Delivery timeline
Four phases across six weeks
The bands follow real work rather than a plan, so evaluation and pilot genuinely share the fifth week.
Phase
W1
W2
W3
W4
W5
W6
- DiscoveryW1
- BuildW2 – W3
- EvaluateW4 – W5
- Pilot & LaunchW5 – W6
Week focus
W1Billing workflow discovery, tariff mapping and the automation boundary
W2Billing and metering integration, and the explanation baseline
W3Enquiry workflow, confidence logic and escalation controls
W4Evaluation suite, citation checks and failure-mode testing
W5Billing-system integration, pilot enquiries and targeted corrections
W6One billing cycle answered under supervision, then Agent Care handover
- Reading the band
- A band covers the weeks its work is named in, and nothing else. The week 5 overlap is real, not padding.
- At the end of W6
- Once the cycle validates, Agent Care owns the running agent.
- Duration
- Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.
Next step · Energy AI agent
Build a billing agent around your filed tariff, not around your call scripts.
Show us your bill enquiries, the tariff sheets behind them and who releases an answer. If you bring those, we'll map the workflow, set the automation boundary and name what stays with the adviser.
- Typical delivery
- 4–6 weeks
- Deployment
- Your billing and tariff stack
- Adviser review
- Tariff-cited
- After launch
- Agent Care available
Nestack Agents · Billing & tariff support
AGT-EN-03 · Agent Care available after launch