Tag Archives: best practices

Where does TimeControl stand in the use of AI?

TimeControl.ai, TimeControl, Chris Vandersluis, Christopher Vandersluis, Christopher Peter Vandersluis HMS started using AI techniques in TimeControl back in 1999.  Surprised?  Artificial Intelligence has been around awhile and you probably wouldn’t have noticed how we leveraged it back in 99.  At that time we created a communications protocol called “HMI” for TimeControl to transmit enormous volumes of TimeControl data through the Internet.  It had the capability of re-routing traffic based on what paths would be fastest.  With the invention of .Net and other Internet-based technologies we evolved from it.

While we are investigating how generative AI could help TimeControl administrators get the best use of TimeControl now, the questions we more commonly receive on the subject are almost all in one area:

“Can you make TimeControl use AI to automatically fill in my timesheet?”

The short answer is yes but don’t get too excited quite yet.  We have the technical capability to have AI determine what the most likely entry for any particular user’s timesheet.  The algorithm would look at the user’s scheduled work, for example, and perhaps past examples of timesheets submitted by that user and then figure out what the most likely entries would be.

The problem is, should we?

An auditor would say, ”Who entered this data?”  “TimeControl did it on its own,” we’d have to answer.  That would a problem for any audit.

“Ah,” you might answer, “but TimeControl could create the draft timesheet and the end user could just approve it.  Think of the time saved?”

That too is technically possible but imagine this scenario: A user is scheduled to work 30 hours this week on “Task A”. They’re scheduled to work 10 hours on “Task B”.  The end of the week comes and the clever AI generator says “Based on the scheduled work, the timesheet should probably look like the schedule”.  The pressure on the employee to just click Ok would be tremendous.  Perhaps they’d even justify it in their head by saying “I’ll make it up next week by doing the reverse.”

You can see the problem.  Algorithmic calculations of what should have happened don’t mix well with the simple recording of “what actually happened”.  So over the years, despite numerous requests, we’ve resisted putting our AI knowledge and our ability to automatically fill in workloads into the timesheet very deliberately.  This is perhaps why TimeControl is supported by both project administrators and Finance administrators at the same time.

We aren’t however, unsympathetic to the desires to reduce the workload in filling in a timesheet and, as a result, there are many features and functions within TimeControl that can reduce the time required to get one’s timesheet complete.  They include:

Preloading where TimeControl will automatically preload your timesheet with the project name and charge code to which the employee was assigned, and which fits into a particular filter of time.

Filtering which can filter out projects and charges the employee isn’t even working on.

Personal Preloads in which the employee defines projects and charges they always want to appear on their timesheet (think “Internal Meetings”) that they don’t want to go looking for.

Validation Rules and testing those rules on the timesheet.  This allows any errors to be caught before they’re even submitted for approval.

Notifications where TimeControl will remind you by email if your timesheet is late or about to be.

Copy where you can just copy a previous week’s timesheet in its entirety if you know you did the same thing this week as you did last week.

TimeControl can also just get you to the timesheet faster by configuring your personal preferences to have the timesheet entry screen automatically appear.

Keeping the structured financial and auditing rules of Finance, the fast-moving progress of project management and the speed at which end users want to move on with tasks they are not focused on is a balance we’ve had to manage since TimeControl was invented some 30 years ago.

Take a look at the Best Practices area of the TimeControl website for more ideas on how to improve efficiency with TimeControl either as an organization or as an individual.

Just tell us about your week

saving_time_300x200.jpgThere is a feature that you won’t find in TimeControl though it has been asked for more times than any other enhancement request.  The specification usually looks something like this:

“We want TimeControl to automatically populate a timesheet with all the hours that an employee should have done so that they can just click ‘Ok’ if the timesheet is correct and save time.”

“Sorry, we won’t do that,” we reply.

It’s not that we’re being difficult.  In fact client and prospective client feedback is a key source of enhancements in TimeControl.  And it’s not that it would be technically difficult to deliver this feature.  The projected hours could have come from the project plan, for example and populating each cell on a line for that task isn’t a major programming challenge.

No, the difficulty is what would happen if we delivered such a feature.  As desirable as it seems to be for end users, we are certain that the result would be bad data.

Imagine that I ask an employee, “What did you spend your time on this week?”

Instead of answering with hours and tasks, they respond to my question with a question, “What do *you* think I should have spent my time on this week?”

I’d be instantly concerned that if I give my expectations that the feedback I’d receive would be tainted by the employee’s desire to please.

Or, imagine this scenario.  You come to the end of your week and your timesheet happily pops up your expected hours.  “Uh oh,” you think to yourself.  “I can see that I was supposed to spend 35 hours on task #27.  But I didn’t spend a minute on task #27.  I spent 35 hours instead on task #25.”  The pressure you’d feel to put at least *some* hours of your timesheet on task #27 would be enormous.

In the over two decades we’ve been selling TimeControl, we’ve seen these scenarios play out in countless organizations.  It’s not that employees want to lie.  They just don’t want to disappoint their managers.

Here at HMS when we get this request we now know to ask what the client wants this feature for and then we ask what they expect will happen to the quality of data if we were to deliver the feature.  The request always fades away.  As it should.  After all, life happens to projects.  The chance that something changed between the plans for the week on Monday morning when the week started and Friday when the timesheet was filled out will be the exception, not the rule.

Instead of prompting people with what they should have done, we want TimeControl to ask something very simple: “What did you actually do?”  TimeControl can preload tasks from a project system’s schedule or from personal preferences to save time in looking up those tasks but TimeControl won’t prompt with the hours per day that were expected.  The result is high quality timesheet data and in the end, that’s what will make managers happiest.

For those who are interested in how to save a minute here or a minute there on configuring and entering their timesheet, the TimeControl website has a section on Best Practices for timesheets at www.timecontrol.com/resources/best-practices.