Data scattered across systems
SCADA, meter data, work orders, billing and spreadsheets all speak different languages, so every report starts with a manual stitch.
Industries / Utilities
Agentic BI for UtilitiesYour network data has a goal.It finds why service drops.
The agentic BI platform for utilities. Your AI analyst investigates outages, losses and demand across your SCADA, metering and billing data. Every number cited, never guessed.
No credit card · 14-day team trial
Utilities analysisnetwork_events.csv
LiveQuestion
Why is reliability slipping in some districts?
Reliability target met, by district%
Root cause
Hillside lines fail in high wind, and 64% of its lost customer minutes trace to ageing lines on just three circuits.
60s
to connect your database and start asking
100%
of numbers cited to a query that ran
24/7
monitoring on the networks and meters you put on watch
The utilities data problem
Every meter reports. Almost nothing is explained.
SCADA, smart meters, work orders and billing each hold part of the picture. Seeing that outages or losses rose is easy. Knowing why takes engineers days, while customers wait for answers.
Outages with unclear causes
Customers lose supply and the cause hides in weather, ageing assets and switching logs.
Losses you can’t locate
Leaks, theft and metering errors drain revenue, and nobody can say which zone they come from.
Demand surprises
Peak demand arrives earlier or harder than forecast, and the network finds out at the worst moment.
Ageing assets, reactive repairs
Transformers and pipes fail before they’re due for replacement, and maintenance runs on fixed cycles.
Dashboards that go stale
Static exports are out of date the moment they’re emailed, and two teams rarely agree on the same KPI.
How it works
From connected to monitoring in two days
No SQL, no ticket to the data team. SchemaChat does the heavy lifting.
See it in action
Ask the questions network and operations teams actually ask
Four real-world questions, answered the way SchemaChat answers them: the number, the why, and a chart of what it found.
Ask
Why did outage minutes rise 18% this quarter?
- Compared outage minutes by district, circuit and cause
- Isolated Hillside and its storm events
- Matched each outage to asset age and repair time
SchemaChat
Customer minutes lost rose 18%, driven by storms in the Hillside district. Three circuits account for most of the increase.
Root cause
Ageing overhead lines on three circuits fail in high wind, and crews wait an average of 4.2 hours for access after each fault.
Ask
Where are we losing the most water?
- Compared water supplied with water billed by zone
- Looked for night-time flow that doesn’t match customer use
- Cross-checked against pipe age and pressure readings
SchemaChat
Non-revenue water is 27% of supply, up from 22%. Zone 14 and Zone 9 account for 58% of the increase.
Root cause
Night-time flow in Zone 14 has tripled since a pressure change, which points to a burst main on 1960s cast-iron pipe.
Ask
Which substation will run out of capacity first?
- Compared peak demand by substation with rated capacity
- Extended each trend using the last three winters
- Checked growth from new connections and electric heating
SchemaChat
On current trends, peak demand will pass Eastside substation’s rated capacity in the second week of January, about three weeks earlier than planned.
Root cause
New heat-pump connections in the Eastside area add evening load that the network wasn’t sized for.
Ask
Which meters are under-reading or not reporting?
- Compared each meter’s readings with its own history and its neighbours
- Flagged meters with flat, zero or sharply dropping readings
- Grouped them by meter model, install date and area
SchemaChat
1,240 meters report unusual consumption or none at all, an estimated $310k of unbilled usage a year.
Root cause
Most flagged meters are one older model installed in 2015–2016, whose communications modules are failing.
What teams use it for
One analyst for every utilities question
From the control room to the billing office, the same governed numbers answer every team.
Outage minutes · Frequency · Repair
Outages & reliability
See where, when and why supply is interrupted, and which circuits, causes and weather events drive it.
Consumption · Read rate · Gaps
Smart metering & consumption
Understand consumption by customer, tariff and area, and spot meters that stop reporting.
Loss rate · Night flow · Pressure
Leaks & network losses
Find where water or energy is lost by zone, and prioritise the repairs that recover the most.
Peak load · Load factor · Headroom
Demand & load forecasting
Understand daily and seasonal load by substation, and see where capacity will run out first.
Asset age · Fault rate · Repair time
Asset health & maintenance
Spot ageing transformers, lines and pipes before they fail, and plan work by risk, not by calendar.
Unbilled usage · Estimates · Disputes
Billing & revenue protection
Catch meter and billing gaps early, and make sure every unit supplied is a unit billed.
Automated dashboards
Operations dashboards that build themselves
- Ask “build me a network performance dashboard”. It picks the KPIs and charts.
- Live, not snapshots. Every widget re-queries on each view.
- Click any district on the map to cross-filter the whole board.
Try it: click a district on the map.
Build me a network performance dashboard
Outage minutes74 min+18%
Network losses8.9%+0.8pt
Peak load86%+3pt
Meter read rate98.2%−0.4pt
Districts & feedersColour = outage minutes
Outage minutes per customer
Outage minutes, last 12 weeksNow 74.0 min · was 52.0 min
Integrations
Works with the data you already collect
SchemaChat reads the database your network systems already feed. There’s no new tracking to install and no data to re-enter.
- SCADA & network telemetry
- Smart meter data
- Outage & switching logs
- Work orders & crews
- Asset register
- Billing & customer data
- Weather & GIS feeds
- Spreadsheets & CSV
Your database read-only
SchemaChat reads from
- PostgreSQL
- MySQL
- BigQuery
- Snowflake
- MongoDB
- CSV files
schemachatAI data analyst
- Root-cause analysisInvestigating outage minutes
- Live dashboardsNetwork performance · re-querying
- Alerts & monitoringPeak load > 90% · watching
- Cited answersEvery number traced to its source
Read-only, credentials encrypted.See all integrations →
FAQ
Questions utilities teams ask
Which utility systems does SchemaChat work with?
SchemaChat connects to the database your systems write into: PostgreSQL, MySQL, BigQuery, Snowflake or MongoDB. If your SCADA, meter or billing data lands in one of those, or in a CSV export, you can start asking questions straight away.
Do engineers or analysts need SQL knowledge?
No. You ask in plain English. SchemaChat writes and runs the queries, and shows the exact SQL behind every answer, so your data team can check it any time.
Can it combine data from different systems, like meters and billing?
Yes, if the tables are in the same connected database, SchemaChat can join them. For example it can link meter readings to billing records and the asset register by meter ID, and define the join once as a governed metric.
How do I know the numbers are right?
Every figure comes from a query that actually ran, with the step behind it cited. If the data isn’t there to answer a question, SchemaChat says so rather than guessing.
Can it control or affect our network?
No. SchemaChat only reads the database you connect. Queries are SELECT-only, enforced at the SQL parser layer, so it cannot modify your data and it has no connection to the equipment that operates your network. Credentials are encrypted at rest, and you choose which tables and views it can see.
Can it handle smart meter volumes?
SchemaChat runs its queries against your database, so aggregated tables, for example hourly or daily consumption per meter, are the best fit. Very high-volume raw interval data is better rolled up first.
Can it alert us when a network metric goes off track?
Yes. Put any finding on watch, for example a substation’s peak load or a zone’s loss rate, and you’ll be alerted when it crosses a threshold, along with a digest of what changed.
Can we pull answers into our own tools?
Yes. The REST API lets you ask questions, run read-only queries and pull charts from your own backends, Slack bots and automations using a Bearer token.
How much does it cost to try?
The 14-day team trial lets you connect your own database (read-only) or upload CSVs, run deep analyses and build dashboards, with no credit card and up to 3 seats. After the trial, Premium and custom plans for larger teams are on the pricing page.
Put an AI analyst on your utilities data today
Connect a database or upload a CSV of recent readings and ask your first question in about a minute.
- 14-day team trial, no credit card
- No credit card
- Cancel anytime