Allocate work with AI

On an allocation report, AI can propose assignments for you instead of dragging each work item onto a resource. It looks at all of the work and all of the resources together and proposes a full set of assignments, which you review before anything is saved.

Before you start

You need:

• An allocation report with AI optimization turned on in its settings

• At least one parameter turned on, so there's something to optimize for

• Resources on the report. Without them, there's nothing to assign work to

Learn more about setting up an allocation report.

Allocate a set of work

1. Select Allocate work

2. Set your date range

3. Select the working days and working hours to allocate within

4. Choose which work to include, such as all unassigned work

5. Choose an optimization goal

6. Select Allocate

Review the proposed assignments on the timeline, then:

• Select Save to keep them

• Select Undo to step back through changes

• Select Redo to reapply changes you've undone

• Select Cancel to back out

Nothing is assigned until you save.

Choose an optimization goal

An optimization goal tells the AI what a good result looks like. The same work and the same resources can produce different assignments depending on which goal you choose.

It's worth knowing how the AI works before choosing. It doesn't fill one opening at a time and move on. It evaluates complete sets of assignments and returns the one that scores best against your goal. So a result that looks odd in one place is often the cost of a better result overall.

Fair distribution

Spreads work evenly across qualified resources, so no one ends up with much more than anyone else. Choose this when balance matters more than cost or headcount.

Fewer resources

Covers the work using as few resources as it can, filling each one before adding another. Choose this when you'd rather have four people fully booked than eight people half booked.

The AI works in terms of available working time rather than headcount, so it may give one resource more work than seems necessary if that avoids drawing on another resource's availability.

Minimize cost

Assigns work to the lowest-cost resources, based on the hourly rates you've mapped.

The AI compares complete sets of assignments and returns the one with the lowest total cost. It isn't simply working down a list from cheapest to most expensive, so a result that uses a more expensive resource can still be the cheapest overall.

Suggest a placement for one work item

When you only need to place one thing, you don't have to run a full allocation. Right-click a work item and choose to have AI suggest a placement for it. The suggestion respects the same parameters your report is set up with.

When work isn't assigned

An allocation won't always place everything, and the message you get when a run finishes doesn't tell you what was left out. Compare the timeline to what you expected before you save.

Work is usually left unassigned for one of three reasons:

• No resource has the skills the work calls for. This applies when the Skills parameter is on

• Resources who could do the work don't have a free block long enough for it in your date range

• A slot was available, but the AI left the work out because including it would have produced a worse result against your goal

If you're using the Availability parameter, a resource with no availability in your date range isn't considered at all. If someone you expected to see gets no work, check their assignments and exceptions on the availability report first.

Learn more about setting up an availability report.

If a run is slow

Runs take longer when there's far more work than the resources have time for, and most of that work can't be placed anyway. If a run is slow or returns few assignments, narrow the date range or the set of work you include, then run it again.

Send feedback

Select Help improve this report to tell us how the results compare to what you expected. Feedback on optimization goals is especially useful while scheduling reports are in beta.