GridCORTEX Live · Customer Experience  ·  ← All Demos
On this page What you are watching The business case Run this at your utility Where you see it and how you say yes

The Energy Advisor Synthetic Data · Simulation

Every other demo on this site is for the people who run the grid. This one is for the people who pay for it. Follow one billing month at the Rivera household (fictional), EV, electric water heater, pool pump, two working parents on a time-of-use rate they never chose to understand. The advisor reads their actual smart-meter shape, moves the flexible loads to cheap hours without touching their comfort, warns them mid-month before the bill becomes a surprise, and turns a grid emergency into a $24 credit instead of a blackout story. The bill drops $92. The trust goes up more. And the utility gets a customer who leans in, 40,000 of whom just became a 34 MW virtual peaker.

DAY 1 · 06:00
NEW BILLING CYCLE
⏳ DECISION POINT: TIME SLOWED
SYNTHETIC DATA
Off-peak $0.09 Mid-peak $0.16 On-peak $0.34 (4–8 PM) Appliance running Event window (credit $2/kWh)

Same house. Same month. $92 apart.

What a grid-aware advisor is worth to a family, and what a million of them are worth to the grid
,
This month's bill
,
Saved vs doing nothing
,
Comfort changes felt
The Rivera Household
WithoutWith AdvisorΔ
The Utility (× 40,000 enrolled homes)
WithoutWith AdvisorΔ
Illustrative simulation on synthetic data; rates, loads, and savings are placeholders; households and programs are fictional. In a GridCORTEX pilot, the advisor runs on YOUR rate book, YOUR AMI data, and YOUR DR programs, with opt-in consent and your brand on the app. See UC 7.2 "Demo and Proof Plan."
$0.00
Month-to-date bill
$244
Projected month-end
$0.00
Saved so far
$0.00
Event credits earned
The Riveras' Phone, advisor app · opt-in · their choice, always
Day 1
Day 10
Heat wave
Event day
Bill day
The Validated Use Cases Behind This Scenario
UC 7.2
Personalized Energy Advisor
The whole demo: the home's real AMI shape + the real rate book + the family's comfort rules, optimized nightly.
UC 7.1
Grid-Aware Customer Chatbot
"Why is my bill trending high?" answered with the customer's own data, not a generic FAQ.
UC 7.4
Billing Exception Resolution
The mid-month forecast alert that replaces the bill-shock call to the contact center.
UC 6.4
VPP Orchestration
The utility side: 40,000 advisor households on event day = a 34 MW peaker that runs on trust.
Inside the Demo
What you are watching, and what it proves

This demo follows one 30-day billing month at the fictional Rivera household: two working parents with an electric-vehicle (EV) charger, an electric water heater, a pool pump, a dishwasher, and air conditioning. They are on a "time-of-use" rate, meaning the price of electricity changes by the hour: 9 cents per kilowatt-hour overnight, 16 cents midday, and 34 cents during the 4-to-8 PM evening peak. They never chose to understand any of that, which is normal, and last month's bill was $231. The screen shows their house, a 24-hour schedule laid over the cheap and expensive hours, and a running "bill race": what the month costs if they do nothing (red) versus if they follow the advisor (green). The feed is their phone. Every suggestion is opt-in: one tap to accept, one tap to say no.

Three suggestions arrive during the month, and each waits for the family. Day 2, the smart schedule: the advisor reads the home's smart-meter data, notices the car charging at 6 PM, the single most expensive hour of the day, and the pool pump running all afternoon. It proposes charging the car after 11 PM, running the pump overnight, heating water before the 6:30 AM showers, and delaying the dishwasher an hour: worth about $55 this month, with nothing changing but the clock. Day 6, it trims the pool pump from 6 hours to the 4 the water actually needs, another $4. A day-11 check-in shows the bill heading to $170, which is $61 below last month, so there will be no surprise at month end. Day 15, the heat-wave plan: six days of 101-degree weather are coming, on track to add $19 to the bill. The plan cools the house to 71 degrees between 1 and 4 PM, while electricity is still cheap, then lets it drift up to 75 through the expensive evening hours. The stored coolness carries them through, saving about $14 of that $19, and their usage in the expensive window drops 48% with no one feeling a difference. Day 20, the conservation event: the grid is short on power for tomorrow evening, and the utility offers a $2 credit for every kilowatt-hour a home shifts out of the 5-to-8 PM window. The advisor proposes pausing the car charger, delaying the water heater, and pre-cooling at 2 PM. The Riveras tap "I'm in," feel nothing, and earn $24.20. Their neighborhood's peak demand drops 11%, and the 40,000 enrolled homes together free up 34 megawatts, the output of a small power plant, without building one.

The month closes at $153: $92 less than doing nothing, including the $24 event credit. Along the way the advisor also flags an odd usage pattern in the garage as a possibly failing freezer, the kind of catch that saves a customer hundreds. If the family ignores every suggestion, the demo plays the usual version instead: the car keeps charging at the most expensive hour, the heat wave runs straight through peak prices, the conservation-event email joins the 91% that go unread, and the bill lands as a $245 surprise, higher than last month on the supposedly money-saving rate. Then come the 40 minutes on hold, the 1-star app review, and the search for rooftop solar. The closing tiles: this month's bill $153 versus $245, saved versus doing nothing $92, comfort changes felt none. The dollar figures are recomputed live from the bill race on each run; the values here are the month the feed narrates.

Without GridCORTEX

The failure is translation. The utility rolled out an hourly-priced rate and handed the family a brochure. Nobody notices the car charging in the most expensive hour, because noticing is nobody's job. The mid-month drift toward $229, before the heat wave even hits, is invisible until the bill arrives. The conservation event is a mass email that gets 9% participation and the Riveras never see it, so the utility fires up its expensive backup power plant instead. The month ends at $245, which is $13 higher than last month on the "savings" rate, and the bill itself damages the relationship: the hold queue, the 1-star review, the solar quote from a competitor.

With GridCORTEX

The Personalized Energy Advisor reads the home's smart-meter data, works out which appliance is using what, and prices every scheduling choice against the actual rate. It proposes; it never imposes. The family approves all three suggestions and can override any of them on any day. The car charges at 9 cents instead of 34, the house banks cheap cooling ahead of each hot evening, and the grid emergency becomes a $24 credit instead of an outage story. The bill closes at $153 with zero comfort changes felt. On the utility side, 40,000 households like this one add up to 34 megawatts of relief on the worst evenings, delivered by participation instead of by building a new plant.

The KPIs, side by side
KPIWithout AdvisorWith AdvisorDelta
The billwhat the Riveras owe for the month$245$153$92 kept
Annualized savingsthe monthly saving projected across a full yearnone~$938real money
EV charging pricethe electricity price the car charges at34¢/kWh at 6 PM9¢ overnight74% cheaper per charge
Heat-wave surchargewhat six days of 101-degree weather add to the bill~$19~$6cheap cooling banked ahead of the peak
Event daywhat the grid-emergency day meant for this familynever saw the emailearned $24trust built
Comfort changes feltwhether anyone in the house noticed anything differentn/azerosame life, cheaper
Bill surprisewhether the month-end number was a shock or a known quantity$13 HIGHER on the "savings" rateknown by Day 11no shock
Peak-event capacity (× 40,000 homes)the power freed up when every enrolled home shifts together during the emergency9% email response34 MW dispatcheda power plant made of goodwill
Bill-shock contact callsangry calls to the call center after surprise bills, each costing the utility money to handle1 per surprise × thousandsdeflected by forecastdollars saved per call avoided
TOU rate perceptionTOU means time-of-use, the rate where prices change by hour; this row is what customers come to believe about it"a trap""a savings machine"the rate gets accepted
App rating trajectoryhow customers rate the utility's app, a proxy for the relationship1-star reviewsdaily-use utility appcustomers engage
Churn/defection signalearly signs a customer is looking for a way around the utilitygoogling rooftop solarleaning into programscustomers kept
Feeder peak (advisor homes)peak demand on the neighborhood power line; lower peaks delay expensive upgradesbaseline−11% on event daywire upgrades deferred
Consumer-protection auditwhether the utility can prove every nudge it sent was fair and in the customer's interestad-hocevery suggestion logged with its reasoningdefensible to regulators
Live KPIs on the dashboard
Month-to-date billThe Riveras' charges so far this month on whichever path the run follows. It feeds the red and green curves of the bill race; the slower it climbs, the better the month.
Projected month-endThe advisor's forecast of where the bill will land, known by day 11 instead of day 31. A number the family already expects is a bill-shock phone call that never happens.
Saved so farThe running gap between doing nothing and following the advisor. It starts growing the first night the car charges at 9 cents instead of 34, and every day it grows is a good day.
Event credits earnedDollars earned during the conservation event, at $2 for each kilowatt-hour shifted out of the emergency window. It ends the month at $24.20 if the family opts in, and $0 if the email goes unread.

The Business Case: Safety, Hours, and Cost

A utility does not buy a demo. It buys a safety exposure that goes away and a cost that goes down. Below is that case for every use case behind The Energy Advisor, written the way a plant manager, a safety lead, and a CFO each need to read it. Every hour and every dollar is a formula you run with your own rates and volumes. There are no vendor benchmarks in here and no invented percentages. If a number is not yours, it is not a number.
UC 6.4 Virtual Power Plant Orchestration and Optimization

What happens today, without this

The VPP, or virtual power plant, program director runs an event by logging into each vendor portal in turn: one for thermostats, one for residential batteries, one for the commercial storage pilot. Each schedule is set separately, the combined delivered megawatts is a guess made by adding the vendors' own optimistic numbers, and opt outs are tracked by refreshing several dashboards during the event. After the event an analyst pulls a comma separated file out of each platform and stitches them together in a spreadsheet to report performance against the demand response commitment, which takes most of a week each month.

What it replaces or shrinks

  • Logging into each vendor platform separately to configure and launch the same event
  • The manual estimate of combined delivered megawatts assembled from each vendor's own numbers
  • Per platform pre event health checks done by clicking through dashboards
  • Spreadsheet stitching of vendor exports into a monthly program performance report
  • Shrinks the staggering decision, when to lead with batteries and when to call thermostats, to a reviewable recommendation rather than an intuition

Why it is safer

The safety effect is indirect. A VPP that actually delivers what it promised on a peak day reduces the emergency operations that follow when it does not: manual load transfers, callouts, and field staff working a hot afternoon into the night.

Counted in units you already track:

  • Switching operations performed for emergency load transfer on peak days
  • Road miles driven for peak day field callouts
  • Night driving hours for staff recalled during and after evening peak events

Man-hours it gives back

Event day hours come back to the program director and the monthly reporting week comes back to the program analyst.

HOURS AVOIDED PER YEAR = events per year x vendor platforms x minutes per platform to configure, monitor, and stand down, plus programs x monthly reporting hours per program x twelve, minus the review time the director spends approving each combined dispatch plan.

The numbers we need from you to run that formula:

  • Events called per year and the number of vendor platforms in the portfolio
  • Minutes spent per platform per event on configuration and monitoring today
  • Monthly hours spent building the performance report per program
  • Enrolled capacity by program and your commitment obligation in megawatts
  • Loaded hourly rate for the program director and for the program analyst

Where the dollars come from

Cost driverHow it is calculated, from a rate you supply
Capacity value deliveredincremental megawatts delivered against commitment x your capacity price or your own avoided capacity cost
Program laborevent and reporting hours avoided x your loaded program staff rate
Underdelivery exposureyour penalty or shortfall charge per megawatt x the shortfall megawatts you currently incur in a typical season
Market revenuemegawatt hours bid into energy or ancillary products x the settled price in your market, for the hours the portfolio was previously idle
Incentive efficiencyyour incentive payment per enrolled device x the devices you no longer need to call because the dispatch is better ordered

Reliability and maintenance

Reliability
This touches peak day reserve margin and the probability of emergency operations rather than SAIDI or SAIFI directly. A portfolio that reliably delivers its commitment is capacity you do not have to buy or build.
Maintenance
The pre event check across all platforms finds dead telemetry, offline devices, and stale enrollments before an event rather than during one. That converts platform housekeeping from a post mortem into scheduled work.

What else it moves

CustomerOrdering the call so that batteries lead and thermostats join later means fewer customers asked to be uncomfortable, which is the single biggest driver of program attrition.
ComplianceOne performance record across all programs, in the form the market operator and your regulator ask for, instead of four vendor formats reconciled by hand.
WorkforceProgram staff spend event day on judgment rather than on operating four user interfaces at once.

What it costs you, stated honestly

You pay for the orchestration platform, for an integration to each vendor DERMS dispatch interface, and for the contract work to get the vendors to expose those interfaces at all, which is often the slow part. Your program staff also need time to validate the combined forecast against a season of real events before they will offer against it.

How to build the payback case

Payback is usually dominated by the capacity value of delivering the commitment plus the penalty exposure you stop carrying. Program labor is real but it is the smaller line.

This is a planning model driven by your event counts, your capacity price, and your penalty terms. It is not a vendor claim. Re run it with the actuals from one full program season.
UC 7.2 Personalized Energy Advisor

What happens today, without this

A customer asks whether a different rate would save them money. Today an energy advisor or a program analyst pulls twelve months of interval data into a spreadsheet, rebuilds the bill under each eligible tariff by hand, and answers that one customer. It takes most of an hour, so it only happens for customers who call and complain loudly enough. Everyone else gets a generic rate brochure and a web calculator that asks them to estimate their own usage. Program and rebate target lists are built the same way, by an analyst filtering on annual kilowatt hours because that is the only field that is easy to filter on.

What it replaces or shrinks

  • The manual spreadsheet rebuild of one customer's bill under each eligible tariff
  • The annual kilowatt hour filter used to build program, rebate, and demand response target lists
  • Shrinks the one at a time rate consultation an energy advisor does over the phone
  • The generic rate brochure and the self entry web calculator for any customer who has interval data
  • Shrinks the post campaign analysis of who took the offer, because the recommendation and the outcome sit on the same record

Why it is safer

There is no direct safety benefit here and we will not manufacture one. This is desk analytics. The honest mechanism runs through arrears: a customer on a rate that fits their usage, with a plan they understand, is less likely to fall behind, and what falling behind produces is field collection and disconnect visits, which are among the highest confrontation and assault risk activities a utility asks of its people.

Counted in units you already track:

  • Road miles driven on field collection and disconnect visits, reached only through the arrears path and only if the plan genuinely lowers bills
  • Night driving hours and after hours premise visits by collectors working evening routes
  • Energized area entries at the meter for disconnect and reconnect for nonpayment

Man-hours it gives back

Advisor and analyst hours come back, and the advisor stops rebuilding bills and starts having the conversation.

HOURS AVOIDED PER YEAR = customers analyzed per year x hours an advisor spends today rebuilding one rate comparison x your loaded advisor rate, plus campaign lists built per year x analyst hours per list x your loaded analyst rate, minus the review hours your programs team spends approving each outreach batch and answering the questions a personalized plan generates.

The numbers we need from you to run that formula:

  • Customers with twelve months of usable interval data, and how many tariffs each is eligible for
  • Hours an advisor spends today on one manual rate comparison, and the loaded advisor rate
  • Campaign and target lists built per year, analyst hours per list, and the loaded analyst rate
  • Inbound contact volume that is rate related or high bill related, and average handle time on those contacts
  • Your cost per outreach by channel: email, portal, print, and outbound call

Where the dollars come from

Cost driverHow it is calculated, from a rate you supply
Advisory labormanual rate comparisons avoided x hours each x your loaded advisor rate
High bill contact volumerate and high bill contacts avoided x your average handle time x your loaded representative rate
Program marketing efficiencyyour current cost per enrollment x the enrollments you would otherwise have bought with untargeted mail, using your own take rates on targeted versus untargeted lists
Arrears and bad debtyour own write off rate x the balance that no longer ages, using the share of high bill arrears you believe a rate switch would have prevented, which you set
Program administrationanalyst hours avoided on list building and post campaign reconciliation x your loaded analyst rate

Reliability and maintenance

Reliability
This does not touch SAIDI, SAIFI, or any generation metric. Where it touches system risk is on the demand side: customers moved onto a time of use rate that actually benefits them, who then shift load, are load you do not have to serve at the peak hour. That quantity belongs to your load forecasting group to size, not to us.
Maintenance
No plant maintenance effect. The operational maintenance effect is on your tariff library. The rate engine has to be kept current with every filed tariff change, and when it is, your advisors, your web calculator, and your campaign lists all read from the same source instead of three.

What else it moves

CustomerA customer who receives an analysis of their own usage rather than a brochure is the customer who does not show up as a detractor on billing and price in the residential satisfaction survey.
ComplianceEvery recommendation is logged with the usage data and the tariff version behind it, which is the record you want when a commission or a consumer advocate asks how customers were steered between rates.
WorkforceEnergy advisors spend their time on judgment and conversation instead of spreadsheet mechanics, which is the part of the job that keeps them in it.

What it costs you, stated honestly

You pay for the GridCORTEX analysis service, for integration into your meter data management system, your customer information system, and your existing email and portal delivery channels, and for your own tariff analyst time to validate the rate engine against a sample of real issued bills before anything is sent. That validation is not optional and it is the largest internal cost in year one.

How to build the payback case

Payback is usually carried by advisor hours and by high bill contact volume, not by program enrollment lift, because enrollment lift depends on your offers and your channels rather than on the analysis. Build the case on the two you can count and treat program lift as upside.

These are planning models you drive with your own rates, volumes, and take rates, not vendor claims. Re-run them against the actual results of your first outreach batches before you scale the program.
UC 7.1 Grid-Aware Customer Chatbot

What happens today, without this

During a storm the contact center takes the same three questions thousands of times: am I out, when will I be back, and is the wire in my yard live. Today a customer service representative answers each one by hand, tabbing between the outage management system and the customer information system, while a supervisor retypes the restoration estimate into the hold message, the web banner, and the chat script every time the estimate moves. After hours those contacts either sit in queue or roll to an overflow vendor who cannot see your outage map at all. The chatbot you have now answers billing questions and hands anything grid related straight to a person.

What it replaces or shrinks

  • The representative lookup of outage status for a service address on a routine am I out contact
  • Manual retyping of the restoration estimate into hold messages, web banners, and chat scripts every time the outage management system changes it
  • Shrinks duplicate outage reports from premises already inside a known outage, because the customer is told the outage is already logged
  • The after hours overflow vendor script for outage status, which never had grid context to begin with
  • Shrinks the chat to representative handoff on downed wire reports, because location and hazard detail are captured before the escalation

Why it is safer

The safety effect is indirect and we will say so plainly. A chatbot does not keep anyone off a pole. The real mechanism is two steps down: duplicate and already restored outage reports generate trouble dispatches that put a truck on the road for nothing, and a downed wire report captured with a clean location and hazard description reaches dispatch faster, which shortens the time a live conductor sits on the ground with the public near it.

Counted in units you already track:

  • Road miles driven on trouble dispatches raised by duplicate or already restored outage reports
  • Night driving hours during storm restoration, when redundant status checks are most likely
  • Energized area entries by trouble crews responding to tickets that current outage status would have closed

Man-hours it gives back

Average handle time and after call work come back to the contact center, and the supervisor stops being the human broadcast system for restoration estimates.

HOURS AVOIDED PER YEAR = contained sessions per year x your average handle time for an outage status contact x your loaded representative rate, plus estimate changes per storm x storms per year x supervisor minutes to republish the estimate across channels, minus the time your escalation team spends reviewing transcripts and working the sessions the service hands off.

The numbers we need from you to run that formula:

  • Annual contact volume split into outage status, billing, and everything else, and your current containment rate on the existing chatbot
  • Average handle time and after call work minutes for an outage status contact, by channel
  • Loaded hourly rate for a representative and a supervisor, and your per minute or per contact rate for the overflow vendor
  • Storm and major event days per year, and how many restoration estimate changes each one produces
  • Duplicate outage report volume as a share of total outage reports, from your outage management system ticket data

Where the dollars come from

Cost driverHow it is calculated, from a rate you supply
Contact center laborcontained contacts x your average handle time x your loaded representative rate
Overflow and after hours vendorcontacts contained outside business hours x your contracted per minute or per contact rate
Supervisor broadcast workestimate changes per year x minutes to republish across channels x your loaded supervisor rate
Avoided trouble dispatchduplicate or already restored tickets avoided x your fully loaded cost per truck roll, a figure your dispatch group already carries
Complaint handlingescalations and regulatory complaints avoided x your loaded hours per complaint case, using your own share of complaints that trace to a wrong restoration estimate

Reliability and maintenance

Reliability
This does not change SAIDI or SAIFI by one minute and we will not pretend it does. It changes what the customer experiences during the same outage, and it keeps trouble crews from being sent to premises that are already logged or already restored, which is capacity your storm room can put on real work.
Maintenance
The maintenance effect is on content, not plant. Answers are traced back to a source system rather than to a hand edited script, so when a tariff or an outage process changes there is one place to change it instead of a script library, a hold message, and a web banner drifting apart.

What else it moves

CustomerOutage status is answered in seconds at two in the morning, which is exactly where residential customer satisfaction studies such as JD Power penalize utilities, on communication during an event.
ComplianceEvery session is logged with the estimate that was live at the moment it was given, which is what you need when a commission asks what customers were told and when.
WorkforceRepresentatives stop repeating the same three sentences all night and spend the shift on contacts that need a person, which is the difference between a storm shift you can staff and one you cannot.

What it costs you, stated honestly

You pay for the GridCORTEX conversational service, for read integrations into the outage management system, the customer information system, and your contact center platform, and for your own staff time to write and approve the answer set and review transcripts during the pilot. Transcript review is the part utilities underestimate, and it is heaviest through the first two storms.

How to build the payback case

Payback is contained contacts times average handle time times your representative rate, and it concentrates in storm months. Build the case on contained volume you can count in your own contact center reporting and treat avoided trouble dispatch as upside your dispatch group has to confirm.

This is a planning model driven by your contact volumes, handle times, and rates, not a vendor claim. Re-run it with your actual containment and handle time data after one full storm season before you size the program.
UC 7.4 Billing Exception and Anomaly Resolution Engine

What happens today, without this

Every billing cycle drops a block of accounts into the exception queue: a read that never arrived, a rate code that does not match the service agreement, an interval file that loaded twice, a multiplier that was never updated after a meter swap. A billing analyst works these one at a time, with the customer information system, the meter data management system, and the work order history open in three windows, trying to establish which one is wrong. The oldest cases age past a full cycle and start generating their own calls, so the customer hears that it is still being worked on twice before anyone explains anything. Estimated bills go out in the meantime and get rebilled later, which produces a second call.

What it replaces or shrinks

  • The three window manual trace of an exception back to its source read, rate code, or interval file
  • The analyst's written diagnosis note, because each case arrives with root cause, evidence, and a proposed correction
  • Shrinks the check read truck roll on exceptions the interval data can already explain
  • The manual grouping of exceptions into common root cause batches, such as every account touched by one tariff update
  • Shrinks the rebill cycle, because fewer accounts go out estimated and come back for correction

Why it is safer

This is back office work, so the safety mechanism is one step removed and specific. Unresolved exceptions generate check read and meter verification truck rolls, and unresolved exceptions turn into arrears, and arrears turn into field collection and disconnect visits. Field collection and disconnect visits are among the highest assault and confrontation risk activities a utility asks of its people, and every one avoided is also road miles not driven.

Counted in units you already track:

  • Road miles driven on check read and meter verification truck rolls raised only to explain a billing exception
  • Energized area entries at the meter socket when a meter is pulled or tested to chase a billing discrepancy
  • Driveway and premise entries by field collectors on accounts that fell into arrears behind a bill nobody could explain

Man-hours it gives back

Billing analyst hours come back, the queue stops aging, and the analyst moves from tracing to deciding on the cases that genuinely need a person.

HOURS AVOIDED PER YEAR = exceptions per year x analyst minutes per exception today x the share the engine diagnoses to a confident root cause, plus rebills avoided per year x analyst minutes per rebill, plus exception driven contacts per year x average handle time x your loaded representative rate, minus the review minutes an analyst still spends approving each batch and working the held high balance cases.

The numbers we need from you to run that formula:

  • Exceptions per billing cycle by root cause category, and cycles per year
  • Analyst minutes to work an average exception today, and the current age of the backlog
  • Rebills issued per year and the analyst minutes each consumes
  • Contact center volume that traces to a billing exception or an estimated bill, and average handle time on those contacts
  • Loaded hourly rate for a billing analyst and a representative, and your fully loaded cost per truck roll

Where the dollars come from

Cost driverHow it is calculated, from a rate you supply
Billing operations laborexception and rebill minutes avoided x your loaded billing analyst rate
Contact centerexception driven contacts avoided x your average handle time x your loaded representative rate
Field verificationcheck read and meter verification truck rolls avoided x your fully loaded cost per truck roll
Unbilled revenue carryaverage unbilled balance sitting in the exception queue x the days the queue is shortened x your weighted average cost of capital expressed daily
Bad debtyour own write off rate x the balance that no longer ages behind an unresolved bill, using the share you attribute to billing disputes

Reliability and maintenance

Reliability
This does not touch the grid. It touches revenue reliability. Unbilled revenue sitting in an exception queue is money you have earned and not billed, and days sales outstanding moves when that queue drains. That is the reliability number your chief financial officer is actually tracking.
Maintenance
The pattern view is the real maintenance value. When a block of accounts all trace to the same tariff update, you fix the configuration once instead of correcting every bill individually, and the same signal tells your meter shop which meter models and which install crews are producing the reads that fail.

What else it moves

CustomerBills come out right the first time, which removes the second and third call and the complaint that follows a rebill nobody explained in advance.
ComplianceEvery correction carries its diagnosis and its evidence, so a billing accuracy audit or a commission complaint response is a query rather than a reconstruction.
WorkforceBilling analysts stop doing forensic tab switching all day, which is the work that makes the role hard to keep staffed.

What it costs you, stated honestly

You pay for the GridCORTEX resolution service, for integration to the customer information system adjustment interface and your meter data management system, and for your own billing controls staff to define which corrections may post automatically, which hold for a human, and where the high balance threshold sits. Expect a genuine internal effort to agree those thresholds before go live.

How to build the payback case

Payback is dominated by analyst minutes per exception times your exception volume, with contact center savings second. Unbilled revenue carry and bad debt are real but they are the lines your finance group will challenge, so carry them as upside.

This is a planning model built from your exception volumes, cycle times, and loaded rates, not a vendor claim. Re-run it against your actual queue age and auto resolution rate after two full billing cycles.
Each of these opens in full on the use case page, alongside the integration plan, the data ask, the path to production, and the operator console. Open the use case library.
For Your Architects and Data Owners
Run this at your utility

What is this, exactly? It is AI software: intelligent agents and models built and delivered by SoftServe, running on NVIDIA accelerated computing. It is not a hardware appliance and it does not replace the systems you run today. It deploys in your own cloud or on your premises, connects read-only to your existing systems, and recommends; your people approve every action, starting in shadow mode until it earns trust.

An analysis and recommendation service for customer programs teams that turns each customer's interval data into a personal energy plan: rate comparison, savings actions, and matched program offers, delivered through your existing channels. The demo above uses synthetic data; everything below describes what the real deployment needs from your organization.

Systems it connects to

Your systemTypical productsHow we connect
Metering (AMI head-end and meter data management)Itron, Landis+Gyr; Oracle or Itron meter data systemsdatabase replica refreshed nightly
Customer Information System (CIS) / billingOracle CC&B, SAP IS-Udatabase replica refreshed nightly
DER management (DERMS) and DER program platformsUplight, EnergyHub, Virtual Peakerread-only API
Document and knowledge storesrate tariff library, program catalogsdocument upload
Outage Management System (OMS)GE PowerOn, Oracle NMS, ADMS outage moduleread-only API
Asset / work management (EAM/CMMS)IBM Maximo, SAP PMdatabase replica refreshed nightly
Market and grid operator interfacesPJM, MISO, CAISO portals; settlementsread-only API

Data it needs from you

How it runs on your systems

Runs in your cloud account on GPU instances; customer interval and billing data stay inside your environment under your privacy rules with need-to-know access, and all connections are read-only. Recommendations flow through your existing channels, and the pilot runs against a control group.

Path to production

Data access and setup (Weeks 1-4)
Meter data, rates, and program catalogs are connected; privacy approval is the usual gate.
Pilot with 1,000 customers (Weeks 5-16)
Personalized plans go to 1,000 AMI customers while engagement, enrollment lift, and satisfaction are measured against a control group.
Evaluation (Weeks 17-18)
Pilot metrics versus control are reviewed and the go or no-go decision is made.
Production hardening (Months 5-6)
Security review, content approval workflow, monitoring, and marketing platform integration.
Production and scaling (Months 6-8)
The programs team sends refreshed personal plans to the full AMI population on a regular cycle.

What we need from your team

Full integration, data, and timeline detail for each use case in this scenario: UC 7.2 · UC 7.1 · UC 7.4 · UC 6.4
For Your Operators and Dispatchers
Where you will see it and how you say yes

The Approve button you just clicked in the demo above is the real workflow. This is what it looks like on the screen of the customer programs manager in the GridCORTEX console:

GridCORTEX ConsoleSigned in: the customer programs manager
Notifications
Advisor run complete: 640 of 1,000 pilot customers save on time-of-use; median savings $214 a year
Daily model refresh complete; all connected feeds healthy
Recommendation
Send personalized energy plans to 1,000 pilot customers
  • 640 customers save on a time-of-use rate, median $214 a year
  • 212 are strong fits for the heat pump rebate
  • Plans use each customer's latest 12 months of interval data
✓ Approve outreach batchModifyDecline
After you approve: The plan batch queues for delivery through the email and portal channels tied to the CIS, and an audit entry records who approved it and why.
Computed from data as of 17:42:10 local; every card shows the timestamp of the data behind it.

What happens when you hit approve

Approve queues the plan batch in your existing email and portal delivery channels tied to the CIS, in pending status until your marketing rules and suppression lists run; GridCORTEX never emails customers directly. Your team can pull any customer from the batch before send.

How you tell it what it cannot see

Interval data, rates, and program catalogs flow in automatically; the manager only sets the target segment and send window when creating a batch.

Live data, not stale data

Uses each customer's latest 12 months of AMI interval data, refreshed daily, and current tariffs; each batch card shows the data as-of date.

Where it lives day to day

Customers see plans in the portal; the team works in the GridCORTEX console; email push when a batch is ready. The console runs in a browser beside your existing screens on day one; embedding into your own systems is a roadmap step once the read-only phase has earned trust. Approve, Modify, and Decline are all captured in an audit trail your compliance team can pull, and GridCORTEX never blocks or overrides anything in the systems you run today.

The Gap: Why Your Existing Systems Don't Already Do This

The fair question from any Chief Customer Officer: "We have a mobile app, energy-use tips on the portal, and a demand response program, what's new here?" Here's the honest answer.

What you own keeps doing its job

  • CIS / billing, remains the system of record for rates, bills, and payments.
  • The mobile app, stays your channel and your brand; the advisor lives inside it.
  • DR programs, enrollment, incentives, and M&V continue; the advisor makes them usable.
  • The customer, approves every schedule and can override anything, any day. Their house, their rules.

The gap GridCORTEX fills, above them, not instead of them

  • Tips are generic. This home isn't. "Run appliances at night" helps nobody. The advisor reads THIS home's meter shape, finds THIS EV and pool pump, and moves THESE kilowatt-hours, $92 of them.
  • Nobody understands their TOU rate. Utilities rolled out time-of-use pricing and handed customers a brochure. The advisor is the missing translation layer: the rate becomes a savings machine instead of a trap.
  • Bill shock is a trust event. The angriest call in the contact center is the $80-higher bill nobody warned them about. A mid-month forecast with the cause and the fix, sent to their phone, deletes that call.
  • DR events feel like the utility taking something. Framed by the advisor ("we'll handle it, you'll earn $24, tap to confirm") the same event becomes the utility giving something. That's how you get 78% event participation instead of 9%.
  • The aggregate is a grid asset. 40,000 homes shifting on the advisor's schedule is a 34 MW virtual peaker with zero steel in the ground; the customer program and the grid program become the same program.
Accent, don't replace: GridCORTEX reads your rate book, your AMI data, and your DR catalog · computes each home's cheapest comfortable day · and delivers it through YOUR app with the customer approving every change. The bill drops. The calls drop. The peak drops. Same kilowatt-hours, better month for everyone.
Under the Hood: What GridCORTEX Took Into Account in This Scenario

When someone asks "what did it actually calculate?", this is the list. In the simulation these factors drive the storyline; in a pilot they are computed from your rate book, AMI data, and program catalog.

🏠 The Home Model

  • Appliance disaggregation from 15-minute AMI intervals, the EV, water heater, pool pump, and HVAC identified from the meter shape alone, no extra hardware
  • Comfort constraints learned, not assumed: when the family actually showers, cooks, and charges, flexibility found around life, not instead of it
  • Thermal model of the house for pre-cooling: how many degree-hours the envelope can bank at 2 PM prices

💰 The Money Model

  • The full tariff engine: TOU windows, tiers, riders, event credits, every scheduling choice priced against the actual rate book
  • Bill forecasting from weather + usage trend: the month-end number known by day 11, not day 31
  • Savings attribution the customer can check: "your EV charged at 9¢ instead of 34¢; that was $48 of your $92"

🤝 The Trust Model

  • Opt-in everything: each schedule proposed, never imposed; one-tap override; no dark patterns
  • Event framing: what we'll do, what you'll feel (nothing), what you'll earn, before we do it
  • Proactive beats reactive: the bill-shock warning, the outage heads-up, the "your usage looks unusual, possible failing freezer?" nudge

⚡ The Grid Payoff

  • Population-scale aggregation: 40,000 households' flexibility scheduled as one 34 MW resource on event day (UC 6.4)
  • Peak shaved where the feeder needs it; the advisor knows which homes are on which transformer
  • Runs on the NVIDIA Agent Toolkit, NeMo-powered conversation in your app, RAPIDS-scale disaggregation across millions of meters, every recommendation Relay-traced for consumer-protection review

Presenter's one-liner: "The advisor read one family's meter, moved their EV and pool pump to nine-cent hours, warned them before the heat wave hit their bill, and turned a grid emergency into a twenty-four-dollar credit. They saved $92 and felt nothing. Multiply by forty thousand homes and the utility got a 34-megawatt peaker made entirely of goodwill."

GridCORTEX Live Scenario Demo · Synthetic data throughout; the Rivera household is fictional; no utility, rate, or program is real · Customer approves every action; opt-in always · SoftServe + NVIDIA · Created by Ronnie Mauldin, NVIDIA Solutions Director, Power & Utilities, SoftServe · JUL 2026