Deutsche Post DHL turns to machine learning to help find the skills of the future

[TrueBlueTerrier]

https://diginomica.com/deutsche-post-dh ... lls-future

Logistics giant Deutsche Post DHL says deployment of an AI-powered internal career marketplace has started to allow its half a million-plus global employee base to take charge of their own career paths. The company also claims that the technology is encouraging team members to build personal profiles that showcases their skills, helping them quickly find relevant training tools to fill skill gaps.

The new system - delivered as part of what the corporation sees not as old-style ‘learning and development’, but more modern ‘learning and growth’ - is also claimed to support retention through internal career progression. It is also seen as boosting productivity, as employees feel more supported and empowered.

The tech - from AI-powered people experience platform supplier Cornerstone - was also able to do in less than five minutes, what an average two years of learning and development (L&D) effort had been unable to: achieve 85% accuracy of skills categorization, even from a non-customized version.

An employee-centric experience
Why an organization would find that useful is because finding the right people for ever-more complex workplace tasks, and at scale, is a huge challenge. The company has a 2025 talent strategy all about identifying the skills currently in the team today and uncovering potential skills for its future. Deutsche Post DHL also has a very wide range of work processes, from desk and knowledge worker to field and parcel handling and delivery.

But one of the biggest challenges in doing that, the company’s VP of Group Learning Talent and Platforms, Meredith Wellard, explained, is how to create a comprehensive database of all the different skills the company needs to have - in HR terms, a skills ‘ontology’.

Deutsche Post DHL is a successor to the former German federal mail handler, growing by acquisition and now active in several global businesses. The DHL arm of the Group Wellard works for offers parcel and international express services, freight transport, and supply chain management services, as well as ecommerce logistics solutions, while Deutsche Post is now the biggest postal and parcel service provider in Europe. Overall, Deutsche Post DHL Group employs approximately 570,000 people in over 220 countries and territories worldwide, generating 2020 revenues of over €66 billion.

Wellard said:

My role is to enable a seamless employee-centric HR experience, so an employee doesn’t have to look at five or six different places to find what they are supposed to be learning or where they need to be updating their CVs. If that’s too hard, they’ll just go to LinkedIn, which we don’t want them to, as we want them to feel that within our organization it's really easy for them to learn, grow and to develop their career.

That’s easier to say than to do, she joked, as the company’s five divisions all have slightly different cultures. However, with her new machine learning and data analytics approach to skills, she feels the company now has a way that simplifies the employment experience employees have. It also ensures, she said, that it can give all Deutsche Post DHL employees an opportunity to grow all the way through their careers and remain with the organization over the long term.

The company tries to deliver on that promise by its in-house certified program, which helps staff become a certified specialist in a specific role, and in parallel grow through learning content available through partners. Recently, the concept of skills had become central to this, she said.

We had started to talk about the new skills required for the future that may not exist today and looking at some of the ‘workplace of the future’ research, the topic of the new skills you would need kept coming up. We started to ask what it would take to really enable targeted learning for employees to develop their skills.

Millions of job and skill data points
A key development here was partnership between her learning team and the Group’s benefits team, which had been working on a new Deutsche Post DHL ‘job architecture’. Wellard decided to combine the two investigations, but also use AI to look at skills at scale her human colleagues wouldn’t have been able to. She said:

Machine learning seemed to be very useful for parsing millions of data points in the market externally on job boards, on career sites and so forth, then collate, curate and sort a set of skills that might be required by us as an organization, both now and for the future.

Wellard adds that the team weren’t AI experts but could see the potential of pattern matching algorithms. She commissioned a small initial proof of concept to explore this, eventually selecting a French supplier called Clustree to develop a pilot. (This company was later acquired by Cornerstone.)

It was this pilot, which ran at the start of 2020, that delivered the impressive ontology results mentioned above. She said:

For many, many years in HR, we’ve been trying to get some control around classifying skills by building taxonomies that describe what skills you would have to have at, say, level 1, 2, 3, and 4 for finance, for HR, operations and so on.

Putting that type of taxonomy together takes a good couple of years, and by the time you've finished it you need to start again, because it's gone out of date. But even the vanilla version of this type of software we trialled, so no trained algorithm, just an off-the-shelf version, but the accuracy of the skills recommendations that came blew our minds.

We couldn't believe that this machine, without even knowing DHL, could recommend something with 85% accurate skills in its first run. That impressed the pants off us.

Wellard and her team quickly moved forward, using the platform to create a new internal skills marketplace. This, she said, gives staff using it very precise learning recommendations for areas the candidate wants to improve in.

This really helps the learning and growth function of the company due to its accuracy but also that the system can recommend not just internal training content to help, but from a wide variety of sources, including free online. That means, she said:

No need for a skills or gap analysis by us, which gives us a huge opportunity for us to respond to a big issue in corporate learning, which is having too much to choose from. Now, the algorithm can do the sorting for you and direct the right learning to the right people.

Making everyone ‘visible’ through technology
Another big advantage is that the system now knows all the career paths that have ever existed in the business, so can say to team members currently in a call center role that with their set of skills, the most common career paths you can follow are A, B and C: would you like to follow one of those or do something different? That support is of particular interest, she said, given today’s shift from an employer-centric market to a more employee-centric one. She said:

Today, employees want to know that they have the freedom to grow and that they have freedom of choice. This enables that, which is super-important for us, and for staff.

Summing up the impact of machine learning skills tools at her company, for Wellard the main return on investment is a focus on the individual. She said,

Everyone was visible before, but because of the hours it used to take to trawl through every individual meant lots of people got missed because it was just too much to do manually. With a machine learning approach, you really can get through everyone in a shorter period.

I was particularly delighted when one user said to us, ‘I've been at the company for 20 years, and this is the first time I felt that anyone's really bothered to think about me and my aspirations.’