Joule agents will orchestrate the business process from bottom to bottom. Data Data The data brick solution is a … [+]
SAP (NYS: SAP) recently announced what they called a “historical partnership” with databricks. This is really very significant. There are ten years or so, there is a technology that truly shakes the markets of enterprise software and supply chain. Sellers who embrace new technology get market share. Those who do not fight. We can be at that point of inflation.
SAP is the world’s largest business software provider. Databricks offers a data intelligence platform. The type of databricks solution is increasingly called a “data fabric” or a data platform built into “Data Fabric Principles”.
A data fabric does not store the data itself; It relates and provides access to data from various sources without moving it physically or copying it. This is an impressive feat that sounds almost magical. A data fabric speeds up and simplifies access to data assets throughout the business. It approaches, converts and harmonizes data from numerous sources to make it usable and active for various business use cases. When the data is deducted from the applications, they can be created once and shifted smoothly, in real time, where necessary.
Data fabrics rely on knowledge graphs to contextualize data. Contextualization is the process of identifying and representing relationships between data to reflect the relationships that exist between data elements in the physical world.
A graph of knowledge creates relationships in data resources. SAP recently issued a graph of knowledge. Knowledge graphs “wander” together previously unrelated data, often exist in various applications or lakes of database, and doing so often detect hidden models and relationships, models that no man can discover.
Subtracting the hype curve
Why does this matter? We have begun to obtain clarity about the most advanced form that the enterprise will take, as well as the steps and technologies needed to get there. We are descending the curve of Him Mortgage.
As my colleague Colin Masson – Colin pointed out is an expert of the ARC advisory group for industrial use for him – at the Arc industry leadership forum in 2024, end users were mostly observers who were looking forward to learning about things like Gen It, agents and industrial use cases for the use of it. The change was “visible” to the ARC 2025 leadership forum. Companies as Celanese were able to tell about advanced uses of what they were providing considerable ROI.
The case of advanced use for him is to create a work orchestration throughout the enterprise. This work is supported by a truly advanced co-pilot capable of surfaces exactly the information a worker needs, with the right context, only when needed!
One of the essential principles of supply chain management is the breakdown of silos. But cross-functional cooperation is not only a need for supply chain departments, it is necessary throughout the enterprise.
No big company is based on applications from only one company. Even companies relying on SAPs can have different cases of SAP in different units or regions of business. Even in the same case of an ERP solution, it is not always easy to get information to flow smoothly through the apps where it is necessary. Finally, external and unstable data should often be collected and contextualized to make better decisions.
The journey of he
For this advanced vision of using it for orchestrating the company all over, a company must clean and harmonize the data. Data fabrics support this. Then, that agent is employed to solve distinct problems. That agent is a group of agents working together. These agents should not be based on it. Sometimes, a microservice inside an application can be an agent. In some cases, mathematics is applied to data to provide an answer, and in some cases, agents must rely on forms of artificial intelligence such as learning machinery.
These agents help solve distinct problems. To fully orchestrate all work throughout one enterprise can require the creation of thousands of agents. The SAP platform supports the creation of agents. The agents are then surfaced to the workers in the form of a co-pilot type user interface. SAP is called joule.
SAP is, and has been actively creating agents. For example, ARC was recently informed about the SAP transport management product. SAP co-ordered an agent focused on the intelligent cargo invoice with a large manufacturer of vehicles. This manufacturer has about 1,000 trucks a day to come to their largest plant. Of them, about 20% suddenly appear. These truck shipments did not use an advanced transport notice as they should have been. The guard will then have to go up from an hour passing through the documents to determine if the shipment was necessary and then create a digital record so that the intake process could continue.
Now the agent of he can scan those documents, and then the large language model is trained to look at and find all the relevant information- the origin, destination, product and quantity. This information is used to dynamically create a shipment within the SAP transport management system, which can then be used to continue the intake process. SAP managers of the SAP products told us that taking time for the truck was lowered by one hour for these mismatch shipments in about 15 minutes.
However, the advantage of the data relationship is that there is a great deal of improved ability to create agents in a heterogeneous data environment. There is the ability to create agents in the abundance of gray spaces that exist between applications and processes. In short, this technology can create a truly advanced orchestration layer throughout the enterprise.
This is not an easy trip
What we have heard in our user forum is that creating beginner agents can be a fight. But then companies get better in it, and the rate of agent creation can grow rapidly.
Halucination is rare when a company’s data is used to create a large language model. But they still happen, and they must be detected and fixed.
Creating agents connecting the factory floor with the rest of the enterprise will not be easy. The ARC classifies data fabrics as a decline in two categories, one type is mainly for transactional data found in applications. This is a graph of enterprise data. ARC classifies data data as by providing this type of solution.
A second type of data graph is for deeply messy data found on the factory floor, which ARC calls a graph of industrial data. Copilot agents used on the factory floor are currently created with industrial data fabrics.
Finally, cleaning, harmonization and contextualization data are difficult. Data fabric and knowledge graph helps a lot with this, but building this foundation will still be a heavy hard work for most companies. In short, there will be no quick ROI for these projects.
Companies need to think differently about ROI about this vision of him. They should think on a decade -old horizon and consider all the value that can be created throughout the enterprise if they can really unlock their data and provide advanced decision support. Most will find the possible value to be massive, but it is, in part, a journey of faith.