Can industry process models fix the agentic AI data problem?
In this podcast, MyWave CEO Geraldine McBride says agent-driven data orchestration, compliance guardrails and cross-platform operation can bridge the legacy and cloud ERP worlds.
Technical explanations for the notorious hallucination problem of generative AI typically place much of the blame on data quality. It's almost a truism to say you can't trust AI answers that come from data that hasn't been cleaned up and standardized.
But MyWave, based in Melbourne, Australia, said it has solved AI's data problems with its agentic AI platform. Industry-specific process models, developed over the course of a decade, keep the AI agents from straying from the compliance guardrails and standard business processes employed in 27 different industries. Cross-platform data orchestration, much of it handled by MyWave agents, helps clean up the data accessed by GenAI large language models (LLMs) to improve their accuracy.
In this episode of Enterprise Apps Unpacked, MyWave co-founder and CEO Geraldine McBride explains how the company developed its industry process models, the role they play in the AI agents and where agentic AI technology is headed.
McBride, whose career includes senior executive roles in several regional divisions of SAP, also discusses MyWave's partnership with SAP on its own agentic AI initiatives and how the technology helps with migrating on-premises SAP ERP systems to S/4HANA Cloud.
Data orchestration and cloud containerization are key
MyWave's data orchestration means agents can work across different instances of SAP ERP or other vendors' systems.
"We don't mind if data is messy and we're not dependent on that from an architectural point of view," McBride said. "We can actually take messy data, enrich it, make use of it and turn it into something that's better."
Geraldine McBride
The MyWave agents are unusual because they can work on both on-premises and cloud platforms, which is critical in helping owners of legacy on-premises SAP ERP move their business operations to the newer, cloud-based S/4HANA. McBride said this is accomplished in part by feeding the program code of the extensions and add-ons that typically adorn on-premises ERPs into an agent-creation tool in MyWave. In days or weeks, the result is a new AI agent that does what the add-on did, but in ways that meet the architectural requirements of the cloud.
McBride said MyWave adopted that approach because banks, which were among its earliest customers, were often reluctant to put their data in the cloud.
"They cared about security and wanted things to be on-premise," she said, so MyWave designed its agentic technology to support both kinds of deployment. "MyWave runs behind the firewall of the client, and it can run on-premise or in the cloud -- in fact, it can run on premise and in the cloud simultaneously." The on-premises versions of the agents are containerized so they can also run in the cloud.
McBride predicted model-driven agentic AI will help further the evolution of ERP toward simplified UIs and automated workflows that insulate users from the underlying software code.
"It is much easier to be able to open up a model and change a model than it is to be able to actually write core code," she said. Users can then command agents to interact with the ERP on their behalf. "You've got things that you know they can do, and they're either working autonomously or they're working alongside you."
People are already getting used to working this way, and the menus and data tables of older ERP systems are starting to disappear as more Baby Boomers retire.
"The next generation are just going to be typing into an interface and work is done," McBride said. "Why wouldn't you want to have the software working for you as opposed to you working for the software?"
Other topics discussed in the podcast include the following:
- The evolution of agentic AI technology since MyWave's founding in 2013.
- Progress on getting agents from different platforms to work together.
- How MyWave's agentic AI works in cooperation with the SAP Joule copilot.
David Essex is an industry editor who covers enterprise applications, emerging technology and market trends, and creates in-depth content for several TechTarget websites.
