In the words of Danny, who runs the operational overview across VACA's locations.
VACA runs several restaurant locations. Danny handles the operational overview across all of them: suppliers, staffing, contracts, the day-to-day follow-ups that keep a multi-location business from drifting.
Before working with Cimalys, that overview lived across email, calendars, documents, task lists and separate operational systems, held together mostly by memory and messages to other people.
"It was not one big task, but a lot of small follow-ups every day. I do not have a reliable hours-per-week number, but it was enough fragmented work that important follow-ups could easily be delayed."
Cimalys built Danny a personal AI assistant on top of his own email, calendar, documents and systems. Here is what changed.
Finding the source instead of guessing
The benefit Danny points to first is not speed, it is retrieval. He can now ask "What did we agree last time?" or "What does the contract actually say?" and get back to the original document instead of relying on memory.
Into a negotiation with the paper trail
Danny says he used to know the broad picture but did not have time to pull up the exact email, contract clause or invoice pattern before a meeting. Now he asks for a short brief with the source material attached, which lets him challenge a supplier on specific points instead of arguing from memory.
Fewer people to ask before answering
Questions about shifts, availability and leave used to mean checking with several managers, especially across locations. Danny now gets a structured overview directly. He is clear about the limit:
"It does not replace local management judgment. A schedule does not always show what is happening in the restaurant right now."
But it gives him a faster, more precise starting point.
What still needs a human
Danny draws a hard line here, unprompted: the assistant prepares work and flags risk, but for supplier agreements, staffing, finance and external communication,
"a person still needs to make the decision and approve any action."
Data quality and system access matter too. When the underlying data is incomplete or inconsistent, the answer still needs checking.
No number, on purpose
Danny declined to publish a figure, and asked for more time before putting one on it.
"The value is mainly in fewer missed follow-ups, faster preparation and better control of information across locations and suppliers. We should measure time saved and response times over a longer period before putting a number on it."
Would he do it again
Yes. He would start where the value showed up first for him: email, calendar, tasks, contracts and staffing, the places where information already exists but is hard to use in practice day to day.
His recommendation, conditional rather than a pitch
"If you run several locations and spend too much time chasing information across email, calendars, suppliers and staff, I would say it is worth testing. Do not expect it to run the business for you. The value is that it gives you better preparation, fewer loose ends and a faster way to get from a question to a decision."