In this tutorial, we'll be working with Maia, our agentic AI, to adapt an existing transformation pipeline in response to new business requirements. Maia can author net new pipelines end to end, but Maia can also amend existing integration logic. By updating the pipeline rather than rebuilding from scratch, Maia will help our teams stay agile, minimising effort while maximising business value. We require a delay probability score table that predicts the likelihood of shipment delays. This table will allow teams to proactively anticipate and mitigate disruption while costing us very little production time to create, thanks to Maia. To achieve this goal, we'll combine three key data sources we have already worked on. This includes historical shipment delays by weather, customer location data, and the three day weather forecast for each customer location. With just this context and the right instructions, Maia will build the probability table for us using the transformation pipeline we have already been working on. We intend to provide Maia with a full set of instructions on how to implement this probability table. The process will begin with creating a lookup table from the historical shipment weather data and creating an average probability based on that historical data of how different weather conditions affect delays. We will ask Maia to link this with customer locations and upcoming forecasts, applying the lookup table to generate probabilities and expected delay times. Finally, the result will be a new delay probability table, which will be a forward looking resource that highlights for each customer and location the chance of delay and relevant forecast details. This will allow teams to proactively mitigate disruption before it occurs to the deliveries. So let's get to it. As you can see, a transformation pipeline has already been created, and the tables we intend to use have been added using table input components. You don't have to create the transformation pipeline already, Maia is perfectly capable of setting up pipelines on its own. In this case however, the pipeline has already been started. So now I will paste into Maia a well drafted plan of action on how I think we should create this delay probability table. The plan of action does not go into too many details, it is just a brief overview of how I think we should go about creating the table. As you can see, Maia starts by summarising the plan of action, and asks if I'm ready for it to start, So I will select yes. Maia is now getting busy implementing our plan of action, which will take a short while to complete. Once it is finished, I'm going to sample the data of each component to check what Maia has done. Now, after sampling the final delay probability table, I can see that Maia has succeeded in filling out our plan. So now I will validate the pipeline, after which I will run the pipeline. As we can see in the task history, the run was successful. And if I check my schemas, I can see our new delay probability table. We started with location data for our clients, and with the help of Maia, we created a complex pipeline to predict delay probabilities in minutes. This shipment delay early warning system will allow us to proactively anticipate and mitigate disruptions in our business. If you want to learn more about Maia and its capabilities, go to Matillion dot comslash Maia.