DHL builds core applications and cuts costs with agentic AI

Challenge

Automating investment approvals and accelerating quality control

Javier Echarte, Head of Digital Transformation, Finance at DHL Supply Chain, is direct about the focus on using technology for standardization and automation.

"That’s our main use case for technologies like OutSystems," Javier shared. "A great example is our investment approval process. It was rooted in a series of business cases written in Word documents and we saw an opportunity to standardize and automate this workflow, while creating a single view of enterprise infrastructure, and supporting pricing."

Because the approval process handles critical contracts and financial data, building this app required speed and performance. "A slow query or failed integration inside the system can stall contracts that drive regional revenue or affect the single source of truth in the risk application," explained Javier.

DHL Supply Chain chose OutSystems after realizing off-the-shelf software couldn’t handle the complexities of the process. As an agentic systems engineering platform, OutSystems was uniquely suited for mission-critical development, offering both flexibility and a foundation for AI.

Since then, DHL has built other solutions with OutSystems, including pricing, risk assessment, and contract management apps.

But keeping these applications healthy meant engineers had to scroll through platform logs by hand and review dozens of pages across dev, QA, and production environments every week. They were losing hours they could have been spending on higher-value work.

Automation with agentic AI offered a solution to the manual process.

Solution

A platform for governed development and automation

AI was on the table to automate the log review process. Together with the OutSystems partner Noesis, DHL launched an AI agent built in OutSystems in just five days.

The agent watches the platform around the clock. It pulls logs from DHL’s OutSystems environment, groups related events, and writes structured reports that identify any affected modules, the probable causes, and the suggested next steps.

Severity drives report classification. When the agent flags a critical issue, it emails the right engineer with a direct link to the full analysis. The workflow runs automatically every night, and engineers can also trigger reports on demand from a simple dashboard.

As for future enhancements, DHL and Noesis plan to adopt a multi-agent orchestration model in which specialized AI agents, each focused on specific log types and error patterns, collaborate in real time to share context and produce a more comprehensive analysis. They also plan to make the system less prompt-specific and transition it into a natural-language agent.

Results

AI saves DHL up to 1,640 engineering hours annually

The investment approval application built with OutSystems captures business cases and routes them into regional or global board meetings for review. This application governs the business case assessment of all new investments.

The agent shifted engineers out of reactive firefighting. DHL and Noesis estimate that time savings can be up to 1,640 hours annually, depending on adoption.

"You really see the value of leveraging AI technology when you implement it at scale," said Javier.

The agent is also helping teams outside engineering learn what OutSystems can do. Non-technical stakeholders can read structured summaries and understand what is happening with the apps that run their processes. That transparency matters to a company where finance, compliance, and commercial teams all rely on those applications.

Looking ahead, DHL Supply Chain plans to expand its partnership with OutSystems by building additional solutions. Future projects include an application to track the full lifecycle of customer orders and a dedicated web shop to manage customer orders and returns.