AI Voice for City Water Utility Billing: Balance, Payment Plans, and Leak Adjustments

City Utility Billing Software with AI Voice: What to Look For

The decision to add AI voice to a city utility billing operation is usually triggered by one of three things: the abandoned-call rate has climbed above 20 percent and the finance director cannot keep getting yelled at about it, the city is mid-cycle on a billing platform upgrade and wants to modernize the phone channel at the same time, or a recent rate increase has flooded the line with payment plan and hardship requests that staff cannot absorb. Whichever the trigger, the buyer is comparing platforms against a fairly specific set of requirements. Here is the checklist that comes up in every utility billing AI voice evaluation.

• Native two-way integration with the city's billing platform. Read account balance, consumption history, payment history, current cycle status, and service status in real time. Write payment plan agreements, leak-adjustment requests, address changes, and service-start/stop tickets back to Tyler Munis, Central Square, Cogsdale, NorthStar, Harris ERP, MuniBilling, Edmunds, or Banyon without staff re-keying. Read-only or scraped integrations are a dealbreaker - they create stale data and audit gaps.
• PCI-compliant payment handoff, not in-AI card capture. The AI must transfer the caller to the city's existing PCI-compliant payment IVR or third-party processor (Tyler Cashiering, Point and Pay, MuniciPay, GovPay, InvoiceCloud, PaymentVision) for the card capture step. The AI never hears, stores, or processes the card data. This keeps the city's PCI DSS scope unchanged and avoids a year of audit work.
• Strong authentication before releasing account data. Account number plus a second factor (service address, phone of record, last four of meter ID, or last payment amount). Configurable to match the city's existing customer-service authentication policy.
• City-specific leak-adjustment policy walkthrough. The AI must walk the resident through the city's actual leak-adjustment policy - eligibility window, documentation required, how the credit is calculated, whether the city requires a plumber's invoice - and capture the structured request directly into the billing platform for finance staff review.
• Hardship and assistance program intake. If the city offers LIHWAP, LIHEAP, senior/disabled discounts, or local hardship programs, the AI must screen for eligibility, explain the program, and either schedule an appointment or intake the application. Compassionate, accurate, multilingual.
• Bilingual or multilingual by default. Spanish is table stakes in most U.S. cities. Mandarin, Vietnamese, Korean, Russian, Tagalog, Arabic, and Haitian Creole come up in specific markets. Multilingual coverage at the safety screen, not just the conversation.
• Warm transfer to a billing rep with full context. When the AI can't resolve, the human shouldn't start at zero. Transfer must include the account number, the question asked, the system lookups already done, and any actions the AI has already taken.
• After-hours and weekend coverage. Most residents call about utility bills evenings and weekends. AI runs 24/7 at the same per-call cost as business hours - no overtime, no third-party answering service.
• ADA accessibility. The phone channel must be accessible. TTY and relay services supported, configurable playback speed, single-word transfer to a human at any point.
• Audit trail of every call. Full recording, full transcript, structured intent classification, every system read and write logged with timestamps. Required for FOIA compliance, billing disputes, and the occasional code enforcement appeal.
• Public-sector data residency and security. Underlying AI and telephony on FedRAMP-authorized platforms. For cities sharing customer data across departments, role-based access controls and field-level redaction matter.
• Procurement path that does not require a year-long RFP. Cooperative purchasing (Sourcewell, NASPO ValuePoint, OMNIA Partners, BuyBoard) or piggyback on a partner-held state master contract is usually the fastest path. Vendor should bring the documentation - capability statement, references, insurance certificates, sample contract language - not make the city's procurement office build it.

The rest of this guide explains how each of those requirements gets met in practice, what the operational picture looks like once the AI is live, and the numbers cities are reporting after the first quarter of deployment.

The Utility Billing Phone Problem

City utility billing is structurally set up to be overwhelmed by phone volume. The math is straightforward. A mid-size city with 30,000 utility accounts generates roughly 5,000 to 12,000 inbound calls per month into the billing line. That number doubles around bill mailing, late-fee posting, rate increases, summer leak season, and any month a major weather event affects service. The team answering those calls is usually four to eight billing reps, supplemented by a couple of cashiers who pick up the overflow when they can.

The result is predictable. During the first week after bills go out, abandonment runs 20 to 40 percent. Hold times stretch past 15 minutes. Voicemails pile up and get returned a day late, usually catching residents while they are at work. Walk-in traffic increases because residents who could not get through on the phone show up at the counter, which makes the lobby worse, which means counter staff fall further behind on the rest of their work.

Cities have tried to push the volume online. Self-service portals (Tyler's Citizen Self Service, Central Square's CustomerCare, InvoiceCloud, Paymentus, MuniBilling's resident portal) absorb the bill-pay traffic well enough, but they do not handle the conversation. Residents who need a payment plan, who think they have a leak, who want to know why the bill is double last month, or who are calling about a shut-off notice are not going to fill out a form. They are going to call.

AI voice rebalances the equation. It absorbs the volume that does not need human judgment, leaves the human work for the billing reps, and shrinks the phone queue down to the calls that actually need expertise.

How AI Handles a Utility Billing Call

The single most common call into a city utility billing line is a balance check, often paired with a "when is this due" or "can I set up a payment plan" follow-up. Here is what that call looks like end-to-end with AI on the line.

1. The call is answered on the first ring. Morgan identifies itself clearly: "You've reached the City of Example utility billing line. I can help with your account, payment plans, leak adjustments, or service start and stop. What can I help you with?"
2. The caller states the request. "I'm trying to find out what I owe and set up a payment plan." Morgan parses the intent and prepares to authenticate.
3. Morgan authenticates the account holder. Account number plus a second factor - service address, phone of record, last four of meter ID, or last payment amount. The authentication policy matches the city's existing customer-service standard.
4. Morgan reads the account in real time. Behind the scenes, it queries Tyler Munis or Central Square or Cogsdale or whichever platform the city runs. It reads back the current balance, the due date, recent payment history, and current service status: "Your balance is $342.18, due on the 28th. Your last payment of $185 posted on the 12th."
5. Morgan offers next-action options. "Would you like to set up a payment plan, make a payment today, talk to me about a leak adjustment, or transfer to a billing rep?" The caller picks; Morgan executes.
6. For a payment plan, Morgan walks the city's policy. Minimum down payment, maximum number of installments, eligibility rules. Morgan creates the payment plan agreement in the billing platform, captures the resident's preferred payment dates, sends the agreement summary by SMS or email, and confirms the next charge date.
7. For a payment, Morgan warm-transfers to the PCI processor. "I'll transfer you to our secure payment line in a moment. You'll be charged $342.18. Sound good?" The transfer hands off to Tyler Cashiering or Point and Pay or whichever processor the city uses. Morgan never hears the card data.
8. Morgan confirms the outcome and ends cleanly. "Your payment plan is set up. First payment of $114.06 is scheduled for the 28th. Confirmation just went to your phone. Anything else?" Call resolved end-to-end in 90 to 150 seconds, no billing rep involved.

For leak-adjustment calls, the workflow walks the resident through the city's eligibility rules and documentation requirements, captures the structured request, and files it directly into the billing platform for finance staff review. For service start/stop calls, the AI schedules the work order and confirms the date. For shut-off avoidance, the AI runs the hardship program screen and either schedules an in-person appointment or intakes the application directly.

Call Types AI Handles for Utility Billing

Not every utility billing call belongs on the AI. The split between AI-handled and human-handled is something the city controls. Here is the typical split for a utility billing operation that has been live with AI for a quarter.

Balance and Due Date Lookups

The highest-volume category and the easiest to fully automate. The AI authenticates the caller, reads the balance and due date, reads the most recent payment, and offers payment-plan setup or live payment transfer.

Payment Plan Setup

Walked through the city's specific payment plan policy: minimum down payment, maximum number of installments, eligibility rules, fees. The AI creates the payment plan agreement directly in the billing platform and confirms the schedule.

Leak Adjustment Requests

The AI walks the resident through the city's leak-adjustment policy, confirms the consumption spike against history, captures the structured request (leak location, date discovered, repair documentation), and files it into the billing platform for finance staff review.

Service Start, Stop, and Transfer

New service start (move-in), service stop (move-out), and address transfer requests. The AI captures the date, address, and forwarding information, schedules the field work order if needed, and writes the request into the billing system.

Shut-Off Avoidance and Hardship Programs

For accounts in shut-off status or approaching it, the AI explains the city's reconnection requirements, walks through the available hardship and assistance programs (LIHWAP, LIHEAP, senior/disabled discounts, local relief funds), and either schedules an in-person appointment or intakes the application directly. Sensitive workflow - the AI is configured to be patient and to err toward warm transfer for residents in distress.

Bill Explanation and Consumption History

"Why is my bill double last month?" The AI reads the consumption history, identifies the variance, and explains the most common drivers - seasonal usage, a meter re-read, a rate change, a possible leak. If the variance suggests a leak, the AI offers to start a leak-adjustment request.

Rate Schedule and Tier Questions

"What tier am I in?" "How much does the second tier cost per gallon?" The AI reads the city's published rate schedule and explains the resident's specific tier based on consumption.

Auto-Pay Enrollment and Changes

The AI walks the resident through auto-pay enrollment, warm-transfers to the PCI processor for the bank or card capture, and confirms the auto-pay start date in the billing platform.

Payment History and Receipts

The AI reads the last 12 months of payment history and offers to email or text individual receipts directly from the billing platform.

Routing to a Specific Billing Rep or Cashier

"I need to talk to the rep handling my dispute." The AI looks up the assigned rep on the account note and warm-transfers with full context.

Calls That Should Always Transfer to a Human

Formal disputes. Anything involving litigation or a chapter 7/13 filing. Escalated complaints. Anything where the caller is audibly distressed or in crisis. Any caller who asks for a human at any point. The AI is configured to err toward transfer - the cost of an over-transfer is a couple of minutes; the cost of an under-transfer is a legitimately upset resident showing up at council.

Integration with Municipal Billing Software

The value of AI voice for utility billing depends entirely on whether it can read from and write to the billing platform the city already runs. Morgan integrates with the major municipal billing systems.

• Tyler Munis (Enterprise ERP). The most widely deployed municipal ERP in U.S. cities, with a strong utility billing module. Morgan reads account balance, consumption history, payment history, and service status; writes payment plan agreements, leak-adjustment requests, and service start/stop tickets.
• Central Square (Naviline, Finance Enterprise). Native two-way integration. Account reads, payment plan writes, and customer portal lookups happen during the call.
• Cogsdale. Now part of Tyler. Morgan reads and writes via the Cogsdale API for cities still running the standalone Cogsdale environment.
• NorthStar Utilities. Common in smaller and mid-size municipal utilities. Native account lookup and payment plan integration.
• Harris ERP. Read and write integration via the Harris API.
• MuniBilling. Cloud-native utility billing platform common with growing mid-size cities. Native API integration.
• Edmunds GovTech. Common in the Northeast and Mid-Atlantic. Account reads, payment history, and service-status writes via the Edmunds API.
• Banyon Data. Common with smaller cities and special-purpose utility districts. Read and write integration.
• InvoiceCloud, Paymentus, Tyler Cashiering, Point and Pay, MuniciPay, GovPay, PaymentVision. Payment processor integrations - the AI warm-transfers to the processor's PCI-compliant flow rather than capturing card data itself.
• Custom and in-house systems. Cities running custom-built billing systems integrate via REST API, webhook, or structured file exchange. We have not encountered a billing platform we could not integrate with given a willing vendor and a published API.

Beyond the billing platform, Morgan integrates with the adjacent systems that show up in a typical utility call: the city's GIS for service address validation, the work order system (Cityworks, Cartegraph, Lucity) for service start/stop tickets, the AMI/AMR meter reading system for consumption history, and the document management system for posting leak-adjustment documentation.

PCI-Compliant Payment Handoff

The single most common technical objection from city CIOs and finance directors on this category is the question of payment processing. Cities have spent years scoping their PCI DSS environment narrowly so that the audit is manageable. Adding a voice AI that captures card numbers would expand that scope dramatically and trigger months of re-attestation work. Done right, AI voice in utility billing does not expand PCI scope at all.

The design pattern is straightforward. Morgan handles the conversation up to the point of payment - authenticates the account, confirms the amount, confirms the resident's intent to pay - and then warm-transfers the call to the city's existing PCI-compliant payment endpoint. That endpoint is the city's existing IVR, a third-party processor (Tyler Cashiering, Point and Pay, MuniciPay, GovPay, InvoiceCloud, Paymentus, PaymentVision), or a payment URL sent by SMS that the resident completes in their browser.

Morgan never hears the card data. Morgan does not record the segment of the call where the card data is collected. The processor's existing PCI-compliant environment handles the payment, and Morgan receives a callback confirming success or failure, which it then reads back to the resident before ending the call.

This design has three important properties. First, the city's PCI scope does not change - the same systems are in scope before and after AI deployment. Second, the resident gets a faster, friendlier payment experience because the conversation up to the payment is handled by a natural voice instead of an IVR menu tree. Third, the audit story is simple: card data continues to flow through the city's existing PCI-attested processor, with the AI sitting upstream as a non-PCI conversation layer.

Leak Adjustments and Hardship Programs

Leak adjustments and hardship program intake are the two emotionally heaviest call types in utility billing, and they are exactly the calls billing reps most want help with. A resident calling about a $1,200 water bill they think is caused by a slab leak is not going to be patient on hold for 18 minutes. A senior on a fixed income calling because they got a shut-off notice is not going to navigate a phone tree well. Done thoughtfully, AI voice handles both categories with more consistency and more patience than a tired billing rep at the end of a long week.

For leak adjustments, Morgan walks the resident through the city's specific policy: what counts as a leak (typically a leak behind the meter on the resident's side that the resident has repaired), the eligibility window (often within 60 to 90 days of the bill in question), the documentation required (plumber's invoice, repair receipt, or self-attestation depending on the city), and how the credit is calculated (typically averaging the prior 12 months or applying a fixed percentage). Morgan captures the structured request and files it directly into the billing platform for finance staff review, with a clear note about what documentation has been collected and what is still outstanding.

For hardship and assistance programs, Morgan screens the resident against the program's eligibility rules - income threshold, senior/disabled status, household size, current arrearage - and either schedules an in-person appointment at city hall or intakes the application directly. For state-administered programs like LIHWAP and LIHEAP, Morgan provides the program contact information, confirms eligibility windows, and offers to text the resident the program intake link. The AI is configured to be patient and to err toward warm transfer when a resident is audibly distressed.

The outcome these workflows produce, repeatedly, is residents who hang up knowing what happens next instead of hanging up frustrated. That outcome is the single biggest reason finance directors greenlight utility-billing AI pilots after a balance-and-payment-plan proof of concept lands well.

ROI for City Utility Billing

The financial case is built on five numbers: cost per phone-handled call before AI, cost per AI-handled call after, walk-in spillover the city no longer absorbs, shut-offs avoided through better hardship program reach, and time billing reps reclaim for dispute resolution and audit work.

For a city utility with 30,000 accounts handling 100,000 calls a year at a loaded $7 per call, current phone-handling cost is roughly $700,000. AI deployment that absorbs 70 percent of calls at an average $0.90 per AI-handled call (and leaves 30,000 calls for billing reps at the same $7) drops total cost to approximately $273,000 - a 61 percent reduction. The savings funds the AI subscription with significant margin and gives the billing rep team back roughly an FTE of capacity to put into dispute resolution and audit work.

The number that usually matters most to the finance director is not the cost line - it is shut-offs avoided. Better hardship program reach typically reduces shut-offs by 15 to 25 percent in the first year, which translates to lower reconnection costs, fewer write-offs, lower bad-debt expense, and a quieter inbox for the city manager.

Procurement Paths That Skip the RFP

The biggest objection from city procurement officers is that AI procurement will require a full competitive solicitation that takes a year and burns through political momentum. It does not have to. Cities have multiple procurement paths that get a pilot live in 30 to 90 days.

• Cooperative purchasing. Sourcewell, NASPO ValuePoint, OMNIA Partners, BuyBoard, and TIPS-USA let cities piggyback on competitively bid contracts that other governments have already awarded. Most cities' procurement codes explicitly authorize cooperative purchasing as a substitute for an independent solicitation.
• State master contracts. Many states maintain master contracts cities can use directly. Texas cities and political subdivisions can procure BetaQuick through partner contract Texas DIR DIR-CPO-6057, which is held by BetaQuick's partner Compass Solutions, LLC. The partner-held vehicle is active through October 2030.
• Direct purchase order. Pilots under the city's competitive threshold (typically $50,000 to $100,000, varies by jurisdiction) can be procured by direct PO. A first-year utility billing pilot often fits cleanly inside that ceiling.
• Sole-source or piggyback on another city's contract. Some procurement codes allow piggybacking on another city's competitively awarded contract. Sole-source determinations work for narrow use cases where no equivalent vendor exists.
• Full RFP. Available if a competitive procurement is preferred or required. We routinely respond to RFPs and bring complete documentation packages.

How to Deploy in 60 to 90 Days

City utility billing deployments follow a structured rollout designed to land safely and prove value before expansion. The standard path is six to twelve weeks from kickoff to live, then department-by-department expansion across the next two to four quarters.

Weeks 1 to 2: Discovery and Policy Mapping

We sit with the finance director, utility billing manager, and a senior billing rep to map call volume by type, identify the top 10 to 12 call categories (which usually represent 75 percent of total volume), and load the city's specific policies into the AI: payment plan rules, leak-adjustment policy, hardship programs, authentication standard, shut-off rules. We confirm integration scope with the billing platform and the payment processor.

Weeks 3 to 5: Configuration and Integration

Morgan is configured with the city's specific account-data reads, payment plan write-back, leak-adjustment intake structure, hardship program scripts, and warm-transfer rules. Connections to Tyler Munis, Central Square, Cogsdale, or whichever platform the city runs are tested in the city's sandbox or staging environment. The PCI payment handoff is tested end-to-end.

Weeks 6 to 8: Internal Testing and Billing Rep Training

Billing reps test Morgan with realistic call scenarios across every configured call type, including hardship-program edge cases. The supervisor is trained on the monitoring dashboard, call review, and escalation queue. The finance director approves the final policy walkthrough language.

Weeks 9 to 10: Soft Launch

Morgan goes live on a defined slice of call volume - typically after-hours and weekends first, then overflow during business hours. Call quality, accuracy on policy decisions, and resident feedback are monitored daily for the first two weeks and weekly thereafter. The city retains the ability to disable any specific call type at any time.

Weeks 11 to 12: Full Utility Billing Coverage

Morgan handles the full utility billing call volume. The billing rep team continues to monitor and field warm-transferred calls. The supervisor reviews a sample of AI-handled calls weekly. Quarterly reviews with BetaQuick refine policy scripts as rates change, hardship programs evolve, and call patterns shift.

Quarter 2 and Beyond: Adjacent Departments

Once utility billing is stable, the same AI infrastructure can extend to adjacent city finance and resident-service functions. Property tax inquiries, business license renewals, parking citation lookup, court fines - all share infrastructure and reduce per-department cost as the deployment grows.

Frequently Asked Questions