How Enlight went from zero AI to a live AI operation
Mexico's #1 distributed-solar installer — 10+ years across 47 cities for clients like Bimbo, Liverpool, Honeywell, and ZF. We started with a blank slate: no AI anywhere in the company.
A list of opportunities, no order, no execution
Coordinating distributed solar across 47 cities is an operational sport. Every installation, sales handoff, and post-sale visit lives in someone's calendar, Slack thread, or notebook. As Enlight scaled past 139 people, the same pattern that built the company started limiting it: the team knew what AI could do, but opportunities were scattered across departments — with no shared baseline of where time actually went, and no way to compare the impact of one bet against another.
Their commercial, operations, and post-sale teams were running fully manual processes that nobody had measured because nobody had time to look from the outside. The cost of those processes never appeared on any report — until somebody listened to the systems themselves.
How we did it
Listening infrastructure
diezX connected as a silent observer to Slack, Drive, Calendar, and Google Meet. We mapped the organization, ran voice interviews across the entire org — not just leadership — and let the system accumulate weeks of operational context before proposing anything.
Pattern detection
From observed behavior, the platform surfaced 357 prioritized automation opportunities, 42 patterns, and 25 motifs. Each scored by impact and effort, with the real cost of every manual process attributed to a process owner — not to a generic budget line.
Production baselines
We built the top-priority workflows directly into Enlight's operation. 24 baselines went live with Diego — our AI copilot — embedded in each, answering technical questions in real time and handing off to the internal AI Champions that emerged from the discovery.
The systems we integrated
No data warehouse, no rip-and-replace. diezX connected to the systems Enlight already used — across ERP, CRM, construction ops, payments, comms, and automation — and made them legible as a single operation.
From a blank slate to a live AI operation
Outcome-based contracts only work when the outcomes are measurable. These are the numbers from the engagement to date — the operational reality, not a deck.
That last number is value nobody saw before. Now it sits prioritized in the backlog — capturing it is only a matter of execution.
As an industrial engineer, I had always struggled to reconcile two things: how much value a company loses in its manual processes, and what to do about it without halting the operation. diezX does the second part without asking me to stop anything. They listen to the systems, prioritize, and embed someone who executes. It's the implementation model that was missing.
Ready to operate AI, not just plan it?
diezX listens to your systems for weeks, prioritizes the 20% of automations that move 80% of the impact, and embeds the experts who execute them. Book a call to see whether we're the right fit.
Book a call