Conor Doherty: To mark the 10-year anniversary of our partnership with Air France Industries, Lokad traveled to the Paris Air Show to catch up with Air France Director of the AFMAE Training Center Jacques Dauvergne. Today is a special day. It’s the Paris Air Show but it is also the 10th anniversary of Air France and Lokad working together. So, my first question is: How did the relationship start?
Jacques Dauvergne: Well, it was very amazing. I was visiting a friend, the CEO of one of our JV in Germany, and after the meeting, this guy told me, “Hey, maybe you should try Lokad. It’s a very small company but they are very good in optimizing the inventory.” At that time, we had exactly the same concerns. When I was back in Paris, I told my team, “Okay guys, you should try to contact Lokad and let’s see what we can do together.”
Joannes Vermorel: Lokad venturing into aviation started with, in fact, a branch of the Air France group which was Spareliners. That’s a joint venture between Air France Industries and Lufthansa Technik, and Spareliners was like the testing ground. So we acquired some experience for aviation. But 10 years ago when we went with Air France, it was on a completely different magnitude because first, in terms of scale, it seemed much bigger. We are talking like 20 times bigger, but also the diversity of challenges is absolutely enormous.
In the case of Air France Industries, they have everything. They have MRO activity, pool activity, they repair jet engines, there are like dozens of different trades. For us, that was really one of the clients who pioneered the world probabilistic approach at scale.
Conor Doherty: So after 10 years, are there any aspects that Air France is very proud of or pleased with?
Jacques Dauvergne: Yes, of course. A very amazing deliverable we had at the beginning was that the first time we ran the Lokad solution, we saw that we have several millions of surplus of inventory in our stock. It was not visible with traditional tools. Another example could be that I remember one day, one of the Lokad members told me, “You should try to have 120% coverage on some references.” Why 120? Because it’s cheap and at least you can be sure that any time the customer needs the part, you will have the part available.
They have changed the way we were looking at stock optimization, and we moved from a traditional way to the new modern way of working.
Joannes Vermorel: The complexity is, when you say you want to do probabilistic and statistical approach, you need to have all the sources of uncertainty modelled. The traditional vision is the only source of uncertainty is the demand. But if you go into aviation and if you go into a large company that is very complex, no, you have like at least I would say half a dozen of sources of uncertainty.
Your turnaround times are uncertain. Your scrap rates are uncertain. Your purchase prices are uncertain. And yes, the demand for parts is also uncertain. You have unscheduled repairs that need to be taken into account. But it’s only a fraction of the total uncertainty. You have plenty of other things.
For Air France, we really had to develop the tooling to combine all of those uncertainties so that we could have an economic optimization, so that for every dollar invested by Air France Industries, they get really the most out of it, for example reducing the number of AOG per year or reducing the number of days for the turnaround time for the repairs of the aircrafts or the engines.
Conor Doherty: So you just talked a little bit about the change. Are there any other examples of how this project has changed the way Air France controls its supply chain?
Jacques Dauvergne: In fact, 10 years ago the inventory experts were using the EOQ formula, a very traditional way of managing stocks. It was a completely different approach from Lokad. They told us, “You should more look at how much money you have in your pockets. If you have 1 million dollars of investment, maybe you should try to set up a competition between the references.”
And that is, they can provide you a list of what are the priorities for investment or as the opposite, what are the priorities for removing parts from the stock. It was a completely different vision about how to optimize assets.
Conor Doherty: Ten years is a long time. I think it’s a long personal relationship. It’s a long professional relationship. After ten years, how do you keep the relationship strong between the two companies?
Jacques Dauvergne: Of course, we had a good relationship with the CEO Joannes Vermorel, of course, but what was very important was the relationship between our teams. I think the main, the key rules were transparency and reliability. Transparency because, for example, each time we had an initiative and we wanted Lokad to work on it, very quickly they can answer you it’s feasible or it’s not feasible. If after one month or two they say, “We have dug into the subject, we are not going to achieve it,” they tell you it’s not feasible.
Because sometimes you may have some providers who tell you they are going to do the job, they say they are not able to. Also, what was very key is that they achieve what they are saying, so we can be very trustful regarding what they say and when we look at the solution, it works. It’s a good way to set up transparency and trust between the teams.
Joannes Vermorel: I mean, again, at Lokad everything pivots through the supply chain scientist. In ten years, the Lokad technology itself evolved enormously. In fact, we wrote and rewrote the code like three or four times. So it’s not the same optimization that is running now compared to ten years ago. We have literally rewritten the code several times.
But also, we have greatly expanded on the realm of, I would say, type of decisions. We started with something very simple which was investment and divestment mostly for rotables, the expensive parts. Investment is what do you buy; divestment is what do you want to resell for secondary markets. For example, with the composition of the fleet that Air France Industries supports evolving, they may have stock of parts that are not needed anymore.
The granularity of the decision has really really changed. There are plenty of decisions now. I think we have almost something like 20 different types of decisions that we optimize for them. So it has enormously expanded and the supply chain scientists were spearheading this change.
Conor Doherty: Again, we talked about let’s say the business side but personally do you have any favorite moments from this project review here from the beginning?
Jacques Dauvergne: Yes. A key moment for us, it was very critical, was the COVID crisis. As you know, for the airline it was a nightmare, and it was the same for many companies. We had a very specific request to Lokad to stop cash out. We had hundreds of components we were ready to send to the US for repair and we asked Lokad, “Could you please set up an algorithm which tells us you can stop this one and have no risk for service level? You have to send this one even with the COVID crisis, it is very important that you keep the flow going.”
And then in a few weeks, Lokad answered positively and we just had a list of recommendations. We followed the recommendations. At the end of the day, it was a success and it was very key, of course, for our economics.
Joannes Vermorel: There were lockdowns that were complete chaos, and for us what happened is that during the lockdowns, white-collar workforces went home by and large. So for 14 months we were without having those people. Obviously, there were still people supporting us but at a much much lower headcount, and vastly reduced. The supply chain was still running at like 70% capacity because even if passenger traffic was down during this period, the cargo traffic was way up. So in the end for us, it was not like a quiet period. It was slightly quieter than usual, but fundamentally there were still a lot of changes, and that really put the Lokad automation to the test.
Conor Doherty: What are the next steps with the project and importantly how can Lokad help?
Jacques Dauvergne: We have a major milestone in our supply chain improvement which is the SAP implementation. We have discussed a lot with Joannes and his team. As we have developed many solutions between the Lokad tools and our ERP, if we remove the ERP we have to be sure that it is still going to be operational.
So we discuss a lot with Lokad how to make it very pragmatic and how to be sure that we are not going to face any difficulties when moving from our current ERP to SAP. The other topic for the coming years is digitalization. We want to move forward to a digital twin. So far Lokad has helped us a lot to be good with the concepts, but we want to move further in that direction. So we will also need Lokad for solutions for that.
Conor Doherty: I’ll see what I can do about all of those things. Thank you very much for your time. I don’t have any more questions. Have a great air show.
