Skip to content

AI FOR UTILITIES

AI support trusted to answer every storm call

Every outage call, high-bill question, and payment answered instantly (storm or blue sky) in 90+ languages. No busy signals.

DoorDashCapital.comToastRemedy MedsFlexPostmanAfriexRio

KEY OPPORTUNITIES

In a storm, your contact center is all customers have, and the first thing to buckle.

  • Outages don't queue politely

    A major storm drives call volume 10–50× in hours, up to 100× in the worst events. The legacy IVR hits its ceiling, and you can't staff your way out of it.

  • Billing confusion, not billing errors

    Your biggest steady volume: customers confused by tiered rates, time-of-use plans, and estimated reads. One in six call back.

  • Some calls can never wait

    A gas leak or downed line must reach a human instantly, and life-support accounts need medical-baseline handling. Payments carry disconnection rules and PCI scope, zero margin for a wrong answer.

Tickets#8263Text

Conversation

My power's been out 3 hours. When's it coming back on? I have someone on oxygen here.

Customer · 6:42 pm

I'm sorry. Your address is in a confirmed outage; restoration is estimated for 9:45 PM, and I'll text you the moment that changes. Since someone's on oxygen, I'm flagging your account for medical-baseline priority and connecting you to a specialist. If you ever smell gas, leave the house and call 911 from outside. Can I text ETR updates to this number?

Giga · 6:42 pm

Yes, text this.

Customer · 6:43 pm

What happened

  1. The customer's power had been out three hours, with someone at home on oxygen.
  2. Giga confirmed the outage with check_outage_status: ETR 9:45 PM.
  3. It flagged the account for medical-baseline priority with flag_medical_baseline.
  4. It set up ETR texts with subscribe_etr_texts, then connected a specialist with transfer_to_specialist.
To a specialistOutage confirmed, ETR texts on, flagged for medical-baseline priority.

VOICE EXPERIENCE

Answer the storm surge, no busy signal, no robotic menu

An ice storm hits and the legacy IVR maxes out, busy signals that make the evening news. Giga answers instantly at 10–50× storm volume, reads the OMS for a real ETR, and hands the judgment calls to a human.

ReviewTickets
Match allChannel is Voice, text, chatCreated is Last hour
DescriptionCreatedOutcome

OMNICHANNEL SUPPORT

Reach customers before they call

Most calls start with a surprise: the bill jumps with no explanation, the power dies with no word. One Giga agent gets there first with high-bill heads-ups, mid-outage ETR texts, and assistance nudges, across voice, text, chat, and email.

Tickets#8317Text

Heads-up before your next bill: summer cooling has moved your usage into a higher tier, so it will run higher than usual. Want to switch to budget billing?

Giga · 10:00 am

Yes, switch me to budget billing.

Customer · 10:14 am

RAPID DEPLOYMENTS

Live before storm season, even on legacy CIS and OMS

A real utility call touches the CIS, the OMS, and meter data, a stack that's aging, on-prem, or browser-only with no API. Giga goes live in under two weeks and works even the browser-only screens your agents click through today.

Data sourcesConnections

Connected by API

No API: through the browser

Default actions

PROVEN RESULTS

Numbers that hold when the lights go out.

  • RESOLUTION AT SCALE

    Over 90% resolved

    The hard, real-time cases, handled without a human.

  • SELF-IMPROVING RESOLUTION

    Resolution climbs to 98%

    Improves on its own as policies refine, no re-platforming.

  • ACCURACY YOU CAN REPORT

    Under 1% error rate

    Proven across 1.2M voice turns, every answer defensible to the PUC.

KPIsAutonomous resolution

Autonomous resolution

98%38 pts since launch

FAQs

Questions utility teams ask us

Have a compliance, OMS, or integration question we didn't cover? Talk to the team

Giga answers instantly through the 10–50× surge that overwhelms a legacy IVR: no busy signals, no DTMF fallback. Storm night and blue-sky day run identically, so abandonment stays under 20% in the first two hours and you're not carrying peak headcount for a few weeks of weather.

Those bypass everything: the agent detects a safety hazard and routes it straight to a live human, flagging medical-baseline and life-support accounts for priority. It never resolves a safety emergency itself. You define the must-escalate lanes, and escalation is reliable, not best-effort.

Yes. It reads the account in your CIS and explains why the bill actually rose (a tiered or time-of-use rate, an estimated read, summer cooling), then offers the next step: budget billing, a payment plan, or assistance. That's how the repeat "why is it so high" call finally stops.

No. The Browser Agent works the same legacy, browser-only screens your agents click through today (policy-governed, fully logged) and completes the whole workflow: a payment, a deferred payment agreement, a service transfer. No multi-quarter integration build.

In production in under two weeks, before the next storm season, not quarters after. Rollout goes by channel: outage status and bill explanation first, then payments and arrangements as your audit record builds.

Both. It reaches out before the phone rings (high-bill heads-ups, mid-outage ETR texts, assistance nudges on the moratorium calendar), the kind of outreach regulators and satisfaction studies reward. And context carries across voice, text, chat, and email, so a text followed by a call never starts over.

Guardrails are built in: it never asserts an ETR the OMS hasn't confirmed, never promises a payment stops a disconnection unless the account and rules confirm it, and keeps card data out of PCI scope. Every action lands in a replayable audit trail behind your reported speed-of-answer, abandonment, and billing-accuracy standards.

Every voice response is checked against your policies and the account before it's spoken. That cuts hallucinations from 4–5% to under 1% with zero added latency. Giga also turns root causes into policy updates ranked by the KPI you answer to, which is how resolution climbs from roughly 60% toward 98%.