Unifying Visibility, Planning, and Agentic AI: Kinaxis on the Future of Supply Chain Decision-Making

To have nice agents working, you need nice data put in place, so I would stress that data are important in order to have access to data from different sources. You need to have, at some point, integration between multiple systems that TraceLink can provide, and not gather all those data within one platform to get access to visibility.

In this world of volatility, a partnership between a Visibility Network and a Planning Platform is quite vital for different reasons.

Number one is the way traditional planning was done historically. It was based on steady assumptions and old-fashioned data. Given all those assumptions, the demand swings mean the plan created by the planning team is already out of date and basically needs to be revised because it no longer makes sense.

What brings visibility is this real-time data point on supplier, on carriers, on logistic provider, not to know what's the latest on the supply chain. Now, with planning on top of that, I can quickly react to see what is the best mitigation plan to solve those issues.

Number two is that visibility without planning is just awareness because you only have data points. "I know that my supplier is late and what about that?" Planning enables us to react quickly not only to sense but also to react in order to find a mitigation plan.

At last, when we see a bit of the licensed competitors in the market, the most competitive of them are the ones that have taken resilience on top of mixing sensing and agility in order to have this real-time visibility combined with a planning engine. Why is this partnership between TraceLink and Kinaxis important?

Because if you look at the pharma space, all of the players to life science manufacturers rely on CMOs, 3PL, 4PL and other service providers. On average, life science players would rely on 600 players to orchestrate the supply chain. I was talking to one of the COVID-19 vaccines manufacturers. They rely on 3,000 partners to run their operation.

TraceLink enables us to have this real-time visibility on the partner side. What happened at CMO level, at CRO level, at 3PL level, at distributor level. All those data points are then provided to Kinaxis in order for us to sense those data and also react quickly each time we have a disruption happening on the supply chain.

TraceLink enables us to provide its outside data within our next call solution to orchestrate the answer to those descriptions. A collaboration between TraceLink and ConnectiHub to manage better exceptions is first we better sense the data. We have access to real-time data points on CMO level, supplier level to know what's happening on the supply chain. Then we detect.

We know what's the impact of a supplier being late on the value chain. What Kinaxis is offering on top of that is resolution protocol in order to recommend to the planning team and supply chain team action to mitigate those disruptions.

Then once you're happy with the resolution, we execute and orchestrate processes, workflows, and people. Maestro and TraceLink are also leveraging AI and new technology. We learn from the past.

We learn from the planner's decision in order to automate this decision making and how to focus the planning team on the new exception and the new description that has never been solved in the past. My view on how AI and IdentityAI will help companies is to better, I would say, manage their description in multiple ways.

In pharma life science, of course AI will help a lot, but some processes, quality, regulation, we still need human intervention to validate any of the decisions.

Then on top of that, if you look at the supply chain, I see first prediction, predictive AI - how we can have more data in order to better predict what's going to be the impact on the market, what's going to be the potential disruption on the supplier side.

The predictive AI, also all the generative AI - how both solutions and AI overall will help people to create content, dashboards, exception reports, and reporting.

Then Agentic AI on top of that will also help to automate the way the planning decisions were made in the past, in order to have the team fully focused on what is new and what are the new actions they need to focus on.

Agentic AI, leveraging prediction and a bit of generative AI, will help to focus the team on the new description where the system was not capable to cope with in the past.

My advice to companies that want to launch themselves in the AI area is first, of course, to have nice agents working. You need nice data put in place, so I would stress that data are important in order to have access to data from different sources.

You need to have at some point integration between multiple systems that a TraceLink can provide and then all together all that data within one platform to get access to visibility. Also the process is important. Data without process is also nothing. Once you get the data, once you know that you have a description, you need to know how to react.

For that, we need clear processes and clear organization behind that to manage this amount of data. My advice is that you need to start small. On a big bang approach with agents everywhere along your organization, you need to find the right use cases and then get the value from that and then to be able to expand.

Those use cases are important because AI is not an outcome. AI is just a technology enabler that will support business use cases and focus the business team on what makes sense for them as a description, I would say.