Artificial Intelligence for Shift Scheduling

Terry Elliott Products & Technology

Shift scheduling has emerged as a hot new SaaS application for employers of contingent workers in the “Gig Economy”.   The product category is introducing some valuable new products that staffing agencies should consider to save time & money, increase back-office productivity and facilitate more responsive and agile scheduling as changes are inevitably required.   Good shift scheduling systems can also improve contractor and consultant accountability for an agency, reduce no-shows, enhance contractor morale and increase customer satisfaction.

It often helps to “follow the money” to see if new product concepts have the potential to stand on their own or whether they are more likely to ultimately become bolt-on applications for more integrated systems.  Venture capital firms have invested close to $110 million in six shift scheduling companies over the past 2 years.  There are now several hundred “workforce management” software vendors.  Mergers have already begun and shakeouts will occur but the product category seems to be doing well on its own.

Staffing agencies can capitalize on recent rostering product improvements but they need to be careful that the software solutions they consider truly meet the needs of an agency and not a direct employer.   At the front end, the shift management system needs to be integrated with the agency’s customer relationship management system, applicant tracking system and onboarding system to assure that the data on the engagement and the contractor is up-to-date.  Shift scheduling needs to be connected to time & work data collections and time & attendance analytics.  How this data comes into and out of the system are important considerations.  Data integrations are also often required with payroll providers, accounting systems, and HR/ERP systems.  The system selected needs to be able to manage such agency-specific requirements as the customer’s approval of the roster, the invoicing process, and the commission calculation process.

As the Chief Data Scientist for Jobscience, I believe the workforce management product category has a potential for process automation and improvement with artificial intelligence.   A relatively narrow and specific application, like shift management, can benefit from these tools.  I also want to make the point that, while one of the benefits of these systems is reporting that provides deeper insight into operations, greater benefits will come from embedding the logic behind the reports in the underlying process logic of the system and then using artificial intelligence tools to improve the process logic.

Take the problem of “no-shows” in shift scheduling.  The is multi-faceted.  An agency can learn about no-shows after the assignment is scheduled to begin when they no longer have time to find a replacement.  Morale of the contractors that do show up can be undermined by the open assignments left by the no-shows.  Contractors without accurate no-show records can continue to assign no-show contractors to teams when more reliable contractors are available.  Rosters without the no-shows removed can be submitted for customer payment approval, undermining confidence in the agency.  These problems can all arise without accurate no-show reporting and analysis.

Jobscience is evaluating shift scheduling flow-event-logic to determine where process automation can be introduced to manage these problems.  For example, many agencies require advance no-show notices.  These notices should trigger automated tasks to assure that a full team is assigned to the engagement.  If a no-show fails to comply by providing advanced notices, they can be assigned to another automated process, a series of steps which ultimately lead to no longer considering them for assignments.

Once the processes for dealing with no-shows has been automated, statistical and artificial intelligence tools can be used to identify patterns of contractor (and consultant) behavior from the data collected.  Ratings can then be developed to differentiate top quartile through bottom quartile performers with processes for managing top-level through bottom level performers.  Although this can be done with more confidence for staffing agencies with larger numbers of engagements and contractors, blinded data can be compiled across multiple similar agencies in the same region to develop key performance indicators (KPIs) for contractors and consultants.

Improved shift management systems and the application of machine-aided and artificial intelligence processes provides the back-office people that manage rosters with more time to fill each team before no-shows become a problem for the agency.  This, in turn, produces a higher level of contractor satisfaction and customer satisfaction with the engagement.