Adam Sroka

Free Gift! The Same Page Tool is the most read and shared article from my newsletter. I've put together an accompanying guide and template form for people wanting to use it. Just head to this link to claim yours: lnkd.in/e9kjrcBQ #Data #DataScience #Leadership

The energy transition will be won by companies that are brilliant at software and happen to own physical assets. Not by companies that are brilliant at building things and hope the software sorts itself out. Which one are you building?

If you spent 160x more on the building than on the security system, you'd call that negligent. In energy, we spend 160x more on the asset than on the intelligence that protects and operates it. And we call it normal.

1/ Decision velocity in energy isn't about technology. It's about organisational design. I've seen companies with world-class AI that still take 3 weeks to act on its recommendations because of sign-off chains. 2/ The technology is ready. The bottleneck is almost always a human

Who has root access to your battery management system right now? Not who should have access. Who actually does. If you don't know the answer immediately, that's the problem.

6.8 GW of battery storage connected to the UK grid. Each one generates millions of data points daily. The companies treating that data as exhaust fumes will lose to the ones treating it as a nervous system.

The best AI project I've seen in energy cost about £15k to build. It reads maintenance logs, flags patterns, and sends a Slack message when something looks off. No machine learning platform. No data lake. No 18-month roadmap. Just one narrow agent solving one problem well.

Every time someone says "fully autonomous" in energy, ask them: who's liable when it goes wrong? Humans decide. Machines scale. Liability doesn't automate.

The infrastructure landlord trap: you build the asset, outsource the intelligence, lose the margin, then wonder why your returns look average. The asset is the easy part. The operating intelligence is the moat.

$66B in global storage investment last year. How much went into the intelligence that operates those assets? A rounding error. The 160:1 Asset-to-Algorithm Ratio is the single biggest misallocation of capital in energy right now.

Speed kills in energy trading. But not the way people think. It's not about millisecond execution. It's about going from "I see a pattern" to "we changed our position" in minutes instead of meetings.

Every energy asset generates data. Very few energy companies treat that data as a product. The ones that do build a nervous system across their portfolio. The ones that don't build a spreadsheet per site.

1/ Data sovereignty in energy isn't abstract. It's very concrete. Can you switch your SCADA vendor without losing 3 years of operational history? Can you give a third-party tool read access without a procurement cycle? Can you query your own telemetry without calling someone? 2

"What's your AI strategy?" Better question: what are the 50 decisions your team makes every day that follow a pattern? Automate those. Call it whatever you want. That's the strategy.

The future of energy isn't "AI replaces humans." It's one operator managing 20 sites because narrow agents handle the routine decisions, and humans handle the ones that matter. Humans decide. Machines scale. That's the model.

Two identical batteries in the same market. Same capacity. Same grid connection. One earns 2x the other. The difference? Decision velocity. How fast you go from signal to judgment to action. 20-35% revenue improvement from advanced optimisation (McKinsey). That's the gap betwe

Energy sector IT spend: 1.5-3% of revenue. Banking: 7-8%. Pharma: 4-6%. Same data complexity. Same regulatory pressure. Wildly different investment. The 160:1 ratio isn't unique to batteries. It's an industry-wide blind spot.

The EU Data Act now requires connected device manufacturers to provide data in standardised formats. Translation: the regulatory world is telling you that vendor lock-in on your operational data is ending. Get ahead of it or get caught by it.

Two types of energy company in 2035: 1. Software companies with physical assets 2. Infrastructure landlords paying rent to Group 1 The difference isn't capital. It's decision velocity. Same battery. Same market. 2x the returns.

What's the most boring, repetitive task your operations team does every single day? That's your first AI agent. Not a chatbot. Not a dashboard. Just that one task, automated properly.

700+ GW sitting in the UK grid connection queue. Most of those projects are planning the asset. Almost none are planning the operating intelligence. They're building future landlords by default.