All articles

Product

Meet Ray, your portfolio's newest coworker

Ray is an AI coworker for solar teams. He works in Microsoft Teams with your live AMMP data, flags problems early and answers questions on demand.

7 October 2026 · 3 min read

It's 8am. You're an asset manager or O&M lead, and before your first coffee there are alerts waiting in a dozen vendor portals, each with its own login and its own idea of what "fault" means. Somewhere in the pile is the one issue that's actually costing money. And the monthly report is due on Friday.

Ray is the AI coworker we built for that morning. He lives in the tools your team already uses, starting with Microsoft Teams and then Slack, and he's connected to your live AMMP data. He flags problems before anyone logs in, and he answers questions when you ask.

On the job before the team logs on

While your team sleeps, Ray reads the portfolio. He checks production against expectations, looks for inverters and strings drifting away from their neighbours, notices sites that have gone quiet, and compares what he finds with open work orders and contract terms.

By the time you open Teams, there's a short morning message waiting. It doesn't list every alert. It ranks the few issues worth your attention, says why each one matters, and links straight to the evidence. You start the day with a plan instead of a backlog.

Across a portfolio, that can mean [TODO: number] alerts reduced to a handful of ranked priorities.

Proactive when you don't ask. On demand when you do.

Ray works in two modes. Some mornings he comes to you. Other times you go to him with a question, the way you'd ping a colleague who knows the fleet.

#portfolio-ops · Microsoft Teams

Ray

Good morning. Three things worth a look today. Top of the list: at Brightwater Solar, INV-03 shows underperformance across multiple strings, producing 38% less than its neighbours. No open work order yet. Want me to draft one?

Sources: Telemetry · Work orders

A

Asset manager

Yes, draft it and assign it to the field team.

Ray

Done. Draft is ready for your review in AMMP OS.

Proactive: Ray's morning message

Direct message · Microsoft Teams

O

O&M lead

Did Valle Soleado meet its availability guarantee last month?

Ray

No. Availability was [TODO: number]% against a [TODO: number]% guarantee, mostly from two unresolved inverter trips. The O&M contract's availability clause applies. I can build the chart for your monthly report.

Sources: Availability data · Work orders · O&M contract

On demand: a question from the team

Every answer shows its sources, so you can check the working before you act on it. Ray investigates, prioritises and prepares the work. Your team stays in charge of the decisions.

Wherever your team already works

Nobody needs another tab. Ray starts in Microsoft Teams, with Slack next, so your team can ask questions in the channels where they already talk about the portfolio.

The same Ray will also power the chat inside the AMMP web app. Ask from Teams in the morning and from AMMP OS in the afternoon, and you're talking to one coworker with one view of your data.

Some teams want to bring their own AI tools. For them, there's the AMMP MCP server. It connects any MCP-compatible assistant, such as Claude or Cursor, to your assets, KPIs, alerts and raw time-series, using the same access controls you already have in AMMP.

Built on 10,000+ sites of data

An AI coworker is only as good as the data under him. Ray sits on top of AMMP's vendor-agnostic data layer, which already brings together telemetry from more than 10,000 sites in 35+ countries.

With 30+ integrations across inverters, meters, batteries, weather data and business tools, data from different vendors lands in one consistent model. That's why Ray can compare an inverter from one brand with its neighbour from another, and why a question about the whole fleet gets one answer instead of twelve.

Ray also follows the permissions you've already set in AMMP OS. He only sees what the person asking is allowed to see, so nothing new needs to be configured to keep data in the right hands.

What's next

Ray is rolling out in Microsoft Teams first, with Slack and the chat in the AMMP web app to follow. Teams that want to connect their own assistants can use the AMMP MCP server and Data API as add-ons today. If you'd like to see Ray on your own portfolio, we'll walk you through it.