Artificial Intelligence guides routing and encourages teamwork at Posti

A hands-on approach can give you a lot of insights when you are looking for challenges that could be tackled with data and artificial intelligence. Riku Tapper, Director of Data and Automation at Posti Group, has sorted parcels at a logistic centre and delivered newspapers in the dead of night to gain an impartial view about challenges that postal workers can face.

“I know from experience that even a short delay can feel really long if you are waiting outside at a newspaper drop off point. Especially when its freezing cold and there is heavy sleet pouring down,” says Tapper, Director of Data and Automation at Posti Group.

Soon after Tapper’s field research experience in newspaper delivery, all Posti delivery personnel started receiving real-time information on when newspapers are arriving.

“Delivery work can be arduous. Relatively small things like knowing how long a wait will be can make things more pleasant and easier for our employees. If you know that there will be a fifteen-minute delay, you can go wait somewhere inside instead of shivering outdoors,” he notes.

Posti employs more than 20,000 people. This means that there is never a shortage of development ideas coming from various sources in the organisation. Posti’s artificial intelligence projects are guided by three themes derived from the company’s strategy:

• improving customer experiences relating to package flows
• enhancing workplace well-being for all Posti employees, and
• strengthening environmental and social responsibility.

Forecasts enable improved parcel routing

During just the past year, package volumes have increased by 25%, and simultaneously the number of letters addressed directly to recipients have dropped by 14%. Tapper affirms that all kinds of change always create new opportunities.

“We recently conducted a survey that showed that 87% of Finns have picked up packages from parcel lockers. Research shows that we often even prefer parcel lockers to home deliveries,” Tapper explains.

It’s no wonder that the capacity of popular pickup locations is stretched to its limits during peak seasons such as Christmas. Regardless of how busy the season is, the customers are always happiest when their packages are routed to their preferred pickup locations.

“When we decide where to route a package, we don’t know yet how many parcel lockers will actually be empty in several hours’ time. In the future, data and AI will help us predict this more accurately. We will be able to utilize our capacity much better and deliver as many packages as possible to the parcel lockers that the customers wished for,” Tapper describes.

Posti can forecast what will happen in the next few hours by combining many different data sources:

• Timing: is it a weekday or weekend, a vacation season or a gift giving season?
• Is the parcel locker located in an apartment building or a mall? How often do people usually visit the location?
• Are some of the packages awaiting registered recipients whose previous pickup speed is known?
• What knowledge do we have of the parcel’s contents? Is it consumer electronics or cat food – this too can have an effect on how soon the package is picked up.

AI encourages teamwork between mail carriers

Tapper also talks about a new application that all mail carriers have on their mobile devices. The application utilizes artificial intelligence and it is called Route Master.

“Route Master uses gamification and helps our mail carriers notice times when a colleague might need help. It is in wide use especially in larger cities,” Tapper says.

The Route Master app encourages individual mail carriers to collaborate and teams to work together, as it gathers statistics and awards points for helping out colleagues.

“Delivering mail is known as an individual sport but in the future, it can be carried out more as a team. Teamwork also improves workplace well-being, so using apps like Route Master can also be seen as a way of enhancing corporate social responsibility,” Tapper points out.

Route planning decreased Posti’s carbon footprint

In accordance with its Zero Carbon 2030 commitment, Posti is on its way towards carbon free deliveries. Last year the company decreased its carbon footprint by 14%.

“At the end of the workday, all of our drivers get automatic feedback on their driving. The report shows if their driving was economic, how much idling there was throughout the day and so forth. It helps our drivers change their driving style to cut our carbon footprint,” Tapper says.

Posti also uses data and forecast models to optimize the home delivery routes. This saves both fuel and time and streamlines the drivers’ work.

“The 14% decrease in our carbon footprint comes from a combination of factors. We provide our drivers feedback, we use renewable fuels and electric vehicles – but decreasing the number of kilometres driven has been one of the main actions contributing to our smaller carbon footprint,” Tapper emphasizes.

Posti also has a pilot project ongoing in Vantaa, in which customers can determine a more precise time window for the delivery and follow the driver in real-time.

“Data and AI can help the environment, postal workers, and customers alike when each route is individually optimized utilizing various data sources,” Tapper remarks.

Employees partake in data development

“All of our strategic focus areas have some kind of digital aspect – it plays a very significant role here,” Tapper underlines.

Posti sees many opportunities for utilizing AI to help solve challenges of all sizes.

“When it comes to very simple AI solutions, like parcel sorting machines, our employees have decades of experience of working with them. In recent years we have started investing more resources into technology and much more sophisticated AI development,” Tapper says.

“We have about 50 data and automation experts in our organization – and loads more employees who are also enthusiastic about data development. We have many different training initiatives going on around Posti to make sure that everyone who wants to learn more, can,” Tapper adds.

In years to come, Tapper hopes to see more and more Posti employees partaking in development work and coding themselves. The civic data science trend could also be seen in Tietoevry’s Data x Business -report.

“The closer we get the actual development work to the business, the better. We want to offer our employees the opportunity to learn how to build machine learning algorithms themselves and solve challenges directly,” he envisions.

The article is written based on the podcast in Finnish.

Listen to the podcast (in Finnish).

Find all the newest DataInsiders podcast episodes in English here.