Skip to main content

FMI Consulting

Resources › AI analysis of scheduling data
Interactive demo · AI and data quality

AI analysis of scheduling data: the quality of the data comes first

What an AI tool can do with PlanningPME and PlanningConnect data, and why it all depends on the data

An AI tool can build charts, comparisons and answers from your scheduling data and recorded hours in minutes, even if you are not an analyst. But it can only tell you something useful if the data is precise and specific. Try the demo to see both sides.

See what AI can do with precise data

An interactive dashboard built on example data extracted through the PlanningPME and PlanningConnect reports: planned hours compared with recorded hours, analysed from different perspectives.

The demo works on a computer or a phone. People and clients are shown as codes, and the category labels inside the demo are in Italian.

Open the demo full screen in a new tab ↗

The AI is the easy part

Today almost anyone can ask an AI tool to turn a spreadsheet into charts, compare two periods or find where something went wrong, with no formulas and no programming. The analysis that once needed an expert is now within reach of the person who runs the scheduling.

The hard part is no longer the analysis. It is what the AI has to analyse. A tool that does not know who did what, for which client and on which day can only give you general advice or a confident guess. The more specific the data, the more useful the answer.

For more on what AI tools change in a company, and what they do not, read AI tools in your company.

Six analyses from the same data

In the demo, one set of precise data answers six different questions.

Overview

How did the period go?

Planned and recorded hours for the whole period, month by month, with the main totals at a glance.

Resources

Who worked how much?

Planned against recorded hours for each person, to see who ran over the plan and who stayed under it.

Clients and activities

Where did the time go?

The same comparison for each client and for each type of activity, such as billable work, warranty or maintenance contracts.

Saturation

How loaded was the team?

Planned work against the weekly capacity of each person, with absences already taken out.

Absences

What was not available?

Holidays, sick leave, permits and other absences, shown separately from the working hours.

Variances

Where did plan and reality differ?

The gaps between planned and recorded hours, flagged as in line when the difference is small.

Why the data makes the difference

Every analysis in the demo is possible because each hour is tied to a person, a client, an activity and a day.

Specific

Each event says who, for which client, which activity and which day. Without that detail, a question such as which client takes more time than planned has no answer.

Complete

The plan alone shows intentions, and the recorded hours alone show results. Only the two together, with absences, show where they differ and why.

Consistent

The same client, the same category and the same person must always appear in the same way. If a name is written three ways, the AI counts three different clients.

Recorded at the source

Hours noted by the person doing the work, when it happens, are more reliable than hours typed in at the end of the month from memory.

Where the data comes from

Two sources, collected as the work is done and not rebuilt afterwards.

Plan

PlanningPME

Every task assigned to a person on the shared calendar is a planned event, with its client, activity and hours. Read more about PlanningPME.

Analysis

The AI tool

An AI tool reads the two sets of data and builds the analyses you ask for. See the monthly timesheet report for an example of the recorded data.

An analysis you can check

Good data comes with clear rules. These are the ones behind the demo, so you can verify any figure.

Planned hours

The hours of the task events in the PlanningPME export.

Recorded hours

The labour hours in the monthly PlanningConnect files.

Months compared

Only months that have recorded hours are compared. In the demo these are June to September 2026.

Weekly capacity

8 hours for each day from Monday to Friday, minus the person’s absences. Context events are left out.

Absences

Holidays, sick leave, permits, medical visits, recovery time and generic absences. Public holidays are shown apart and are not part of the total. Absences come from the plan, not from the recorded hours.

In line

A gap counts as in line when the recorded hours are within 25% of the planned hours.

What AI does not do for you

An AI tool does not know your business and does not decide for you. It can read the data, calculate and present it clearly. It cannot recover an hour that was never recorded, and it cannot tell whether a gap between plan and reality is a mistake in the estimate or a good reason. That judgement stays with the person who plans.

This is why the order matters: first reliable data collected at the source, then analysis. Read how the pieces fit together in job and time management. Where a shift started is one more detail you can record at the source: see GPS clock-in by industry.

Read next

AI tools in your company

What they change, what they do not, and where to start.

Read the guide →
Download

Timesheet and job-costing report in Excel

A monthly report of hours, travel and expenses by technician, client and job order.

See the report →
Guide

Planning shifts with AI and Excel

Where a spreadsheet is enough, and where you need scheduling software.

Read the guide →

Questions about the demo

Short answers about the data and the analysis.

Is the demo built on real client data?
It uses anonymised example data: people and clients are replaced by codes. The demo does not upload or export any file.
Do I need to be a data analyst to do this?
No. With an AI tool you describe what you want to see and it builds the analysis. What you do need is data that is specific and consistent, and the habit of checking the result against the rules behind it.
What happens if my data is incomplete?
The analysis will be incomplete too, and it may look convincing anyway. Hours that were never recorded, or events without a client, leave gaps that an AI tool cannot fill in reliably.
Which months does the dashboard compare?
Only the months that have recorded hours. In the demo they run from June to September 2026.
Can I do the same with my own data?
If you plan with PlanningPME and record hours with PlanningConnect, you already have both sets of data. Contact us to talk about what you could analyse.

Want to see what your own scheduling data can tell you?