The fastest path to the intelligent factory may start with Excel

A spreadsheet used to control production is rarely just a spreadsheet. It may contain the real sequencing rules, recurring exceptions, material substitutions, shift knowledge and customer promises that never made it into the official system. Colours carry meaning. Comments explain edge cases. A macro written years ago quietly coordinates work across departments.

This is fragile, difficult to scale and dangerous when knowledge holders leave. But it is also compressed domain knowledge. Treating it only as technical debt throws away the very context a modern operational system needs.

The traditional catch-up plan is familiar: hire a transformation team, map every process with consultants, procure separate systems for planning, quality, maintenance and reporting, integrate them, migrate the data, and hope the organisation still recognises the original problem at the end.

Sometimes specialist expertise and dedicated systems are necessary. But their existence in a larger company does not make them mandatory steps for everyone else. Ennea's principle is to integrate before replacing: keep the ERP, machines, databases and spreadsheets that still contribute value, then build shared context and better operational loops across them.

The starting point can be an Excel workbook, a PDF, an email thread, a folder of work instructions, machine exports or simply a conversation. Konnie's role is to inspect these sources, identify the apparent objects and relationships, and propose a structured understanding that people can correct.

The company does not need to translate its operation into perfect requirements first. Its people provide what only they know: why a column matters, which exception is intentional, who may approve a change, what must never happen automatically and what a good result looks like. Context becomes the scarce human input; much of the translation becomes AI work.

The operating model we are building toward is deliberately ambitious: Konnie should handle roughly 80 to 90 percent of the repetitive groundwork—reading source material, proposing mappings, structuring factory context, drafting an app or workflow, preparing documentation, testing assumptions and revising the result after feedback.

That percentage is a design target, not a universal measured outcome or a promise that every factory is already receiving today. The important idea is the division of labour. People should spend their time on meaning, exceptions, priorities, permissions and validation—not on months of configuration and specification work that AI can help compress.

Speed disappears when every operational problem creates another disconnected application. Planning, a quality checklist, a maintenance workflow and a management report may look like separate projects, but they often depend on the same orders, machines, materials, people and time.

Ennea brings integrations, shared factory context, planning, apps and automations into one platform, with Konnie as the common intelligent interface. A manufacturer can begin with one urgent loop and let the operational model grow from there instead of buying a different island for every department.

AI can accelerate the wrong interpretation just as easily as the right one. A colour in a workbook may be decoration or a safety-critical signal. An informal workaround may be outdated or may protect the factory from a failure nobody documented. Speed without traceability simply turns hidden risk into automated risk.

Every proposed mapping should remain reviewable. Unknowns should stay visible. Permissions and approvals belong in the design from the start. A workflow should first assist, then prove itself, and only automate more when the evidence and operational boundary support it. Leapfrogging means compressing unnecessary implementation effort, not skipping responsibility.

Choose one file that people use to make a real decision every day. Identify its owner, the decision it supports, the sources it depends on and the exceptions that matter. Let Konnie propose a model and a small working application or workflow. Correct it with the people who know the operation, then run it beside the existing method until it earns trust.

The future of manufacturing does not have to wait for a perfect data estate, a large transformation office or a five-year replacement programme. It can begin with the knowledge already keeping the factory alive—and an AI employee capable of helping turn that knowledge into the next operational system.