Using AI to set up your employees for success at Deutsche Post DHL
As the world’s largest logistics company Deutsche Post DHL (DPDHL) has had its work cut out in the past 12 months with the exponential growth in e-commerce and home delivery. The company created 20,000 jobs last year and now has 570,000 employees in 220 countries, of which some 400,000 are customer-facing out in the field. The company’s operations include post, freight, supply chain and ecommerce services, and revenue increased by 5.5% year-on-year to EUR 66.8 billion in financial year 2020.
With such a large field workforce on whom many businesses and consumers were relying during COVID-19, DPDHL needed to find a way to quickly adapt to an environment where it could communicate swiftly and directly with customer-facing staff who needed to be kept informed and safe. Getting to each of them was a major challenge. No longer could managers hold the regular in-person performance dialogues before people went out on shift. Instead, the company needed to find a way to reach its frontline people quickly to identify what type of information, what new skills and, particularly, what support they needed at that time.
As Meredith Taghi, VP group learning, talent and platforms at DPDHL, says: “We were really committed to ensuring that they were never in a situation where they didn't have the information or the training on how to use the various health items that were distributed.”
The story could end there but the HR team at DPDHL spotted an opportunity. If they could communicate at scale with their field workforce they could not only provide targeted learning to give their people the skills needed today but also start to prepare them for the skills they might need tomorrow and in the future. DPDHL already had two key initiatives on roadmaps prior to the pandemic. An app, called Connect, designed for employees to give them access to corporate information, message boards and the learning management system on their own device was in planning stage. A Career Marketplace initiative, based on artificial intelligence to map skills, had already gone through the phases of receiving expressions of interest to evaluate vendors.
“When COVID came we thought, oh no, we’re going to have these stopped. But the opposite happened, everyone started saying get it out there faster,” Taghi explains.
What DPDHL did
The company was no stranger to digital. Back in 2019 it launched its 2025 strategy with the aim of excellence in the digital world. But artificial intelligence (AI) based tools were a new concept to HR. Says Taghi: “The digital agenda was right there on the forefront, which then kicked off a lot of investigation for HR into what we need to do to change and update and upgrade our digital agenda. We had long-term, traditional talent and career practices that were difficult to scale across our whole organisation. We were talking together about the need to have a skills framework and remembering that last time we did that it took us six years of work and it was used for a year and then everyone forgot about it. Then someone mentioned it was possible to do what we wanted using AI.”
This kicked off a discovery phase and research into AI-based solutions. “We're lucky to be in this period of time where HR is discovering what HR means for the future. But we don't yet have all the skills ourselves,” admits Taghi. “When we started talking about AI we were thinking about this whole ‘robots are taking over’ piece, because you can't imagine that a machine can go out into the market and take millions of data points, analyse them, bring them all together and say, you know what, here are the skills that you need today. So we decided to approach it as both discovery in terms of what product we might want to go with as well as a learning experience for us as an organisation and as professionals in talent, in learning, in recruiting, compensation and benefits and so forth.”
The tool they chose was Clustree (now owned by people development solutions vendor Cornerstone OnDemand) which enables DPDHL to have a sustainable skills ontology. “It’s not fixed. so we don’t have to redo the whole thing every five years but can, month by month, day by day, see what skills are most important for different roles at the moment and what skills might be important over the next six or 12 months,” explains Taghi. “And we link that with another big initiative that is going on in our organisation called Workplace of the Future. And when those two come together, you start to get this complete fit picture of the long term, the immediate need and what transitions in between.”
The company began in mid 2020 with proof of concept where some 200 end-users across eight countries were able to try out the tool and feedback. And, says Taghi, the response was overwhelmingly positive. “They described it as feeling as though someone finally understood what they did. To them it was not a machine but someone interested in their growth, recognising what they did and then going a step further and making them an offer in terms of learning or a next best job. Once we saw their responses we knew we were definitely going forward and, through show and tell, we got the buy in of the HR board and other key stakeholders in our business,” she says.
Since then DPDHL has started training the algorithm. Employees can put their profile in and it recommends a set of skills that match what they do and then they can add additional skills based on what they’ve done before and external activities. “One of the things that we always suspected but never really knew for sure was that employees were doing a lot of self-education outside of regular business time, especially now that there are products on the market where you can just go online and access learning. And what we've noticed is that now people are saying, I did a digital marketing diploma, and now that can be recognised as part of their skills. I think this is probably for me the biggest surprise, how many people spend a lot of time self-educating and we never recognised it,” says Taghi.
The company is rolling out the tool through what it calls its pioneers, who then become change champions encouraging others to try it. The next pioneers will come on board in October to try the learning and job recommendations.
“We roll out and build at the same time. And I like that because it means we're not hitting everyone with everything at the same time. We're gradually helping people develop the confidence and the competence and build the trust. We are starting softly softly by just giving people the chance to say, these are my skills and I'd like to be visible for receiving recommendations on learning and jobs potentially. And then we encourage that and we develop that. Then the next phase is that people are engaged in that process and seeing some benefits. And that will take us 12 months,” explains Taghi.
While it’s still early Taghi says she expects to see the most benefit from increasing internal placements as a result of the targeted skills development. “Talent and learning are long lead time activities. You can't flick a switch or insert something to make that happen. So if you can reduce the time to discovering skills gaps or discovering a potential next best opportunity then you have more time to actually target the development and get someone there. That’s one of the biggest wins.
“The other one is that we have such a large organisation and we have a great opportunity to grow people from the ground up, but we've never had a way to make opportunities visible to them and to really push the opportunity towards them. The AI has this matching capability that lets me say, your interests sit here and your skills are there, here are some suggestions that might suit you for the future. That might be introduced a little later this year and it’s going to just be a game changer for us.”
She also thinks it will help with diversity and engagement. “There's significantly less bias in this type of process than in the manager recommendation model. And we saw that, even in that very short proof of concept, people stayed thoroughly engaged in the process and therefore the business.
“We call it a change project with a bit of IT in it because it really turns everything we know on its head. We no longer talk about roles, because skills are our new currency. We no longer talk about talent management, because now it's a career marketplace. We no longer talk about the manager making the decision, because we want more transparency and decision-making in the hands of the employee.”
Taghi’s lessons on implementing AI
• The biggest lesson is that not all AI is AI. When I first started looking into artificial intelligence, I had made this assumption that there's this one thing called an algorithm, and then it just works this way. In reality AI is everything from text matching to super sophisticated, multifaceted algorithms that do several things all at once, including machine learning. And they're not the same but they're both called AI. So when you look at products, if people are saying that it's based on an AI technology, make sure you really look at what you need and get the right one.
• The second lesson is you don't need a Ferrari to take your kids to school in the morning. So don't buy the big AI if all you need is a text matching solution. But similarly, don't do it the other way around, because it won't be sustainable and it won't give you what you would like to have.
Meredith Taghi, pictured below, is VP group learning, talent and platforms at DPDHL