What CrewFlow's AI actually looks at when it builds your schedule

"AI-generated schedule" can sound like a black box. In CrewFlow it isn't — every assignment comes with a written reason, and the process behind it is straightforward enough to explain in full.

What goes in

When you click Generate Schedule, CrewFlow sends its scheduling engine your employee availability, skills, and equipment; your client preferences, required certifications, and estimated job hours; and the geographic locations of every employee and job site.

What it does with it

• Filters eligible employees for every job, by availability, required skills, and required equipment
• Calculates drive times between all job sites, so it knows the real cost of any given route
• Builds a full week of assignments that minimizes total drive time while honoring client day, time, and preferred-employee preferences
• Writes a plain-English reason for every single assignment

The whole pass takes 15 to 45 seconds for a typical week. That's the same optimization problem that used to take a manager several hours, run consistently, every week, without getting tired of it by Thursday.

Reading the reasoning

Every assignment card has an expandable reasoning note. In practice, these tend to read like:

"Maria is the client's preferred employee and is available Tuesday morning."

"James has the required floor scrubber equipment and lives closest to this site."

"No preferred employee was available — assigned Carlos as next best match based on skills."

That last example matters: preferred employees are a strong preference, not a hard rule. If the preferred person genuinely isn't available, CrewFlow assigns the next best match and says so explicitly, rather than silently picking someone and leaving you to wonder why.

The AI Summary

After generation, an AI Summary calls out the things worth a second look before you publish — employees near capacity, clients whose preferred employee wasn't available that week, and any job that couldn't be fully resolved due to a skills or availability gap. It's meant to be read before the schedule, not after something goes wrong.

Every schedule is a draft until you publish it. You can reassign any single job by hand without regenerating the whole week — the AI does the heavy lifting, you keep the final call.