L.E.B. Electric: the first working system was live before the meeting ended

L.E.B. Electric has spent more than three decades building a reputation on Long Island and into New York City. LEBS Electric, the Florida company, is run by Brian Chierichella as president and partner alongside William Parker. Two operating companies, two labor markets, two margin profiles, one leadership team, and commitments that cross both constantly.

Those commitments get made where the work happens: on a job site, in a truck, on a call. They were landing in four places at once. Email, text, phone, and a handwritten notebook, none of them connected. One executive's inbox held tens of thousands of unread messages and was still the only place anything was tracked. Monday's action items faded by Wednesday, and follow-through depended on someone remembering to go looking for it.

This is the pattern Strong Tide sees in most well-run middle market companies, and a two-entity structure sharpens it: the people holding cross-company context are the same three who cannot hold any more of it. It is not a discipline problem. It is a design problem.

What made it hard was a stack that was already working. L.E.B. runs on Microsoft 365 and Exchange, supported for two decades by the same outside IT provider. Estimating and project systems were in place and functioning, and Microsoft Copilot licenses were already deployed. Any AI layer had to sit on top of that environment and respect both the technology and the relationship. And it had to be usable the same day, because a tool that requires a training program does not get adopted by people who spend their day in the field.

Strong Tide started with the operator, not the org chart. Discovery ran per executive, mapping how commitments actually moved through each company, then coordinated with the group's outside IT provider on licensing and mobile mail before anything was configured. The first deliverable was deliberately small: a capture layer that turns anything spoken, written, or recorded into a tracked commitment with an owner and a date. Everything else in the roadmap sits downstream of that one habit.

The conventional move at this size is a company-wide program: choose a platform, license everyone, run training, then measure adoption. Strong Tide recommended against it. In an owner-led company, the operating system is not software. It is three or four people. Every consequential decision across L.E.B. and LEBS passes through the CEO, the CFO, or the LEBS president, so the engagement started at the bottleneck on purpose. Nothing was migrated, purchased, or taken away from anyone, which is what made a same-week start possible. If it had not worked, there was nothing to unwind.

The sequence ran the same way for each executive: discovery with the operator, alignment with existing IT on licensing and permissions, then a single working session where the system was configured and demonstrated live against that week's real commitments and left running before anyone stood up. Short weekly sessions since then fix what the last week exposed and add the next workflow.

Three workflows, built around how the work already happens. Each was demonstrated live against real commitments and left running in the executive's hands at the end of the session.

Capture: say it once, and it is tracked. A commitment made in the field is spoken into a phone in plain language, the way it would be said to a person. It is parsed, assigned an owner and a date, and routed to the right list automatically. No form, no app switching, no typing on a job site. Spoken to tracked in under a minute, without stopping the work.

Accountability: two lists replaced a dozen open loops. Rather than a generic task manager, the workflow was built around the two questions an operations lead actually asks: who owes me something, and what do I need to verify myself. Every captured item routes to one or the other, so nothing sits in an ambiguous middle state, and every commitment carries an owner and a check date.

The paper problem: the notebook connects to something. Handwritten notes were the hardest habit to replace, so Strong Tide did not try to replace them. A photo of a notebook page is read, the action items are extracted, and they land in the same tracked queue as everything else. The same path works for a recorded call: transcript in, structured list out. Nobody retypes anything.

The build followed four rules. Nothing new to log into. Do not change a working habit; the grid-paper notebook stayed, and the system was built to read it. An owner and a date, or it does not enter. One session, one working outcome.

Nothing here is custom software. Each piece was chosen for one job, and the systems the group already depended on were left alone.

• Claude (Anthropic) handles the judgment: turning plain speech into an owner and a date, reading a photographed notebook page, structuring a transcript. It runs in L.E.B.'s own team workspace, with an L.E.B. employee as administrator and per-executive instructions written from each person's actual priorities.
• Microsoft 365 Copilot was licensed and deployed before Strong Tide arrived, and it stayed in place for document and mail work inside Office. Treating this as a two-tool question rather than a bake-off kept the work moving instead of stalling in a platform debate.
• Granola records and summarizes the calls that used to produce nothing durable, including a standing Monday block. Action items route into the same tracked queue.
• Exchange, Outlook, and Microsoft To Do are the system of record. A commitment captured by voice in a truck appears in Outlook on the desk and in Microsoft To Do on the PC. Tasks live in the Microsoft account, not a consumer cloud.
• The phone already in the pocket is the capture surface, with four lists structured for accountability rather than by project: My To-Do, Team Follow-Ups, Verify, Personal.
• Estimating and ERP stayed untouched, by design. No integration was attempted in this phase, so no part of the work depended on one.

Every component is licensed product L.E.B. could have bought directly, and both the workspace and its administrator belong to L.E.B. No code was written, so there is nothing to patch and nothing that stops working if the engagement does.

Three executives are live across two companies, one session each. Configured, demonstrated against real commitments, and running before each session ended. No pilot phase, no rollout plan, no training curriculum.

Adoption did not have to be driven. Each executive began entering real commitments in his own words during his own session. In the LEBS session, five live commitments were captured by voice inside the first hour, unprompted, and a photographed notebook page became tracked action items in front of the person who wrote it. Nobody was asked to give up the notebook, which is why the change held past the first week.

Nothing was replaced and nothing was bought. Copilot stayed. Strong Tide worked directly with the group's IT provider of two decades and cleared the mail configuration issue blocking production use, in person, in a single visit.

New York and Florida now run the same capture layer, the same list structure, and the same tracked queue, so commitments that cross both companies stop living in two separate heads. And the next scope came from the client: leadership is assembling its own AI use-case list across both companies, and the LEBS president wants the system in the hands of a 20-plus person Florida team. The expansion conversation is being led by the business, not by the consultant.