The next frontier of data isn’t another dashboard
Most energy organisations have invested heavily in data platforms, analytics, machine learning, and, now, AI. Yet many critical business processes still rely on spreadsheets, emails, manual approvals, and disconnected workflows.
The gap now isn’t insight. It’s action – the distance between what the data tells you and what actually gets done. Closing that distance, not generating another report, is where the real gains are hiding.
The last mile nobody talks about
The past decade of data investment has been remarkable. Platforms, pipelines, warehouses, lakehouses, machine learning, and now agents – most energy businesses have the lot, or are well on the way. And yet walk into almost any operations team and the real decisions are still being made with a spreadsheet.
A curtailment figure gets questioned. Someone exports it, adjusts it by hand, saves a new version, emails it to a manager for sign-off, waits for a reply, then re-keys the corrected number somewhere else. While the analytics were world-class, the last mile was a mess.
This is the quiet problem with insight. Businesses have spent years getting very good at telling people what happened and what might happen next, while leaving the action to whatever ad hoc process grew up around it.
What actually closes the gap
Closing that gap takes more than another report. It takes applications where data, models, and workflow come together in one place – where the same trusted data and analytics that powers a dashboard also drives the decision, captures the action, and writes the results back to the systems everyone relies on.
It helps to think of it as the top rung of a ladder the industry has been climbing for years:
- Dashboards tell you what’s happening.
- Models predict what might happen.
- AI agents recommend what to do.
- Applications let you actually do it — on the same trusted data.
The first three generate insight. The last one closes the loop.
And doing covers far more than fixing a number. The same principle, integrated data and analytics forming the basis for the workflow, underpins a whole class of operational applications:
- Business planning and scenario analysis: Planners test assumptions against live data and models, and the results are captured and signed off rather than lost in someone’s spreadsheet.
- Pricing and bid/offer engines: Analysts tune the inputs, the models recompute, and the output is reviewed before it goes to market.
- Fleet and schedule management: Operational data and constraints drive a schedule that people adjust, approve, and act on – without leaving the platform where the data resides.
- Operational decision support: A model or agent surfaces a recommendation, and the operator acts on it in the same place, with the action recorded.
- Data capture and adjustment: The example we’ll walk through below.
Different jobs, same shape. Trusted data and analytics as the foundation, a workflow on top, and a governed path from decision to action.
What it looks like in practice
Rather than describe it, we built one.
The Ignite Curtailment Data Management Portal takes one of those use cases – data capture and adjustment – and makes it real. It does an unglamorous but important job: keeping curtailment records correct. Anyone who has reconciled generation data knows the drill. A settlement figure looks off, an adjustment is needed, and the reason has to be recorded and signed off before it touches anything downstream.
We built it on Databricks Apps. That choice mattered less for the interface than for what sat behind it: the app runs next to the data, under the same permission model, with no second copy, no shadow database, and no separate hosting stack for security to sign off on. Whatever Unity Catalog already says a person can and can’t see holds true inside the app, and every action is logged the way everything else is. The governance isn’t bolted on afterwards – it’s inherited.
Here’s the workflow, end to end.
None of that required anyone to open a notebook, touch a table, or go near the backend. And because it all lives on the same governed platform, the governance came for free.
That last point is worth dwelling on. Governed write-back – letting business users change trusted data safely, with approvals and a full audit trail of who did what – is exactly the thing spreadsheets can’t give you and most operational tools make painful. Here it’s simply the default behaviour.
What we learned building it
Building it taught us more than the demo shows. A few things worth passing on:
- The unlock is the governed workflow, not the interface
Anyone can build a form. The hard, valuable part is letting business users act on trusted data safely – with approvals and a complete audit trail – which is exactly what a spreadsheet can never give you. If you take one thing from this, take that.
- The approval queue is the point
The moment changes stop hitting the table directly and start queuing for review, an ungoverned edit becomes a reviewable, attributable, reversible one. That single design decision is what makes business-user write-back safe enough to allow.
- Decide what can’t be changed first
Locking the keys – settlement date, unit identifier – is a data-integrity decision, not a UI one. Get it wrong and a well-meaning app becomes a fast way to corrupt a trusted dataset.
- The hard work is the process, not the code
Because hosting, authentication, and governance come with the platform, almost all the effort goes into modelling the real-world workflow. That’s a feature, not a cost: it’s the part that deserves your attention anyway.
- The review that stalls most internal tools mostly disappears
Internal apps usually die waiting on infrastructure and security sign-off. When the app runs on the same governed data under the same permission model, there’s no new stack to approve – so the thing that normally stops these projects reaching production largely goes away.
Why this matters for energy and utilities
For energy, utilities, and asset-intensive operators, this is the difference between analytics that inform and analytics that run the business. The payoffs are practical rather than theoretical: spreadsheet-driven processes move onto governed datasets, decisions and approvals happen where the data lives with an audit trail by default, and business users get to work with the platform without being exposed to its plumbing – closer to the data, further from the complexity.
And in a sector where a wrong number can flow into settlements, market submissions, or a regulator’s inbox, that governance is a process you can defend.
Is your process a candidate?
Not every spreadsheet needs to become an app. Here’s a quick test – the more of these that are true, the stronger the case:
- The work combines data, a calculation or model, and a human decision.
- A trusted dataset gets exported, adjusted, and loaded back in — or would be, if the tooling existed.
- More than one person needs to sign off before a change or decision is official.
- You need a defensible record of who changed what, and when — for audit, compliance, or a regulator.
- The output has downstream consequences: settlements, pricing, schedules, market submissions.
- It’s a recurring process on a cadence, not a one-off clean-up.
- The people doing the work are business users, not engineers.
If most of those hold, you’re looking at a candidate. If it’s a one-off, nobody needs to sign off, or the output doesn’t feed anything downstream — a spreadsheet is probably still fine, and that’s okay.
It’s not a Band-Aid; your foundations matter the most
An app on top of the wrong foundations isn’t a fix – it just gives you a faster way to make mistakes. If your data and governance are a mess, you must sort that first. It’s for the operational last mile, not a wholesale replacement for good dashboards, well-built pipelines, or sound modelling. Plenty of questions are still better answered by a chart than a form.
But for the growing list of processes where the real work is doing something with the data, this is the piece that’s been missing. The challenge in most energy businesses is not generating insight. It’s turning insight into action without leaving the platform that produced it.
That’s the shift worth paying attention to. And it doesn’t look like another dashboard.
We built the curtailment portal to see what the pattern could do — and it’s only one of many. If you’ve got an operational process that still runs on exported spreadsheets and email approvals, we’d be glad to show you what it could look like instead.