Covering Disruptive Technology Powering Business in The Digital Age

Written by: Izzat Najmi Abdullah, Tech Journalist, AOPG.

Have you ever considered creating your own custom AI agents to streamline and enhance your business processes? As Artificial Intelligence (AI) technology advances, the opportunity to integrate intelligent agents into your workflow is more feasible than ever. Boomi is currently trying to make this evolution with its AI Agent Framework. David Irecki, CTO for APJ at Boomi, during an interview with Disruptive Tech News, offered his valuable insights into how businesses can leverage this framework to optimise operations and future-proof their systems.

David firstly reflects on Boomi’s historical commitment to integration and automation, noting that AI represents a new frontier in these efforts. “Boomi for many years has been focused on integration and automation,” David explains. Historically, Boomi’s solutions have helped companies connect applications and databases to automate processes and achieve desired outcomes. As AI technology evolves, it offers new ways to enhance these processes.

In the past, companies have used digital transformation initiatives to link various data sources and automate workflows. Today, AI provides the tools to take these efforts further. David envisions AI as a way to build on existing business processes: “AI is going to augment a lot of those processes that are in business today.” This might involve enhancing customer interactions with advanced chatbots or using AI to offer personalised recommendations in retail settings.

David Irecki, CTO for APJ, Boomi

For instance, rather than just having a retail chatbot that answers customer queries, organisations can now leverage AI to analyse past interactions and preferences. By connecting this “AI agent” to the organisation’s data systems, the chatbot can offer a far more personalised experience, which includes the ability to suggest products based on previous purchases or even provide detailed product information.

Boomi’s AI Agent Framework

But what exactly is an AI Agent Framework? Is it akin to the MI6? Let’s clarify: Boomi’s AI Agent Framework is definitely not designed for espionage. It is a framework that is designed to manage and optimise the use of AI agents within an organisation. David outlined the framework’s vision, which consists of three main components:

1. Marketplace: The Marketplace is the starting point for businesses looking for AI agents. “First, there’ll be a marketplace where we will have gold-standard AI agents that Boomi recommends to our customers,” he said. This component offers a curated selection of high-quality AI agents for various business needs. For instance, businesses can find agents to summarise financial data or optimise logistics and supply chains. The Marketplace serves as a hub for discovering and deploying AI agents that have been tested and approved by Boomi.
2. Agent Garden: The Agent Garden is where users can interact with and manage their AI agents. “You’ll be able to pull down one of those agents and interact with it through the GPT,” the CTO described. This component not only allows users to deploy agents but also provides tools for monitoring their performance. Users can access analytics and data to understand how the agents are functioning and what data sources they are using. The Agent Garden ensures that businesses have the insights they need to manage their AI agents effectively.
3. Registry: The Registry serves as a centralised repository for tracking all AI agents used within an organisation. “You will need a registry for the whole company to look up what agents have been used,” David explained. This component helps organisations maintain an overview of their AI agents, ensuring that agents from different departments are accounted for and managed effectively. The Registry helps organisations keep track of which agents are in use, their purposes, and their performance.

In July, Boomi will be introducing another six new internal AI agents as part of this framework. These agents are designed to simplify the use of Boomi’s products, allowing users to interact with the system through conversational interfaces rather than traditional drag-and-drop methods. This development reflects Boomi’s commitment to making AI technology more accessible and user-friendly.

Partner Agents and Real-world Applications

Partner agents are a significant part of Boomi’s AI strategy. A notable example is the collaboration with Vianai Systems, a leader in conversational AI solutions for finance. Together, they have developed Boomi FinTalk, an AI agent that facilitates real-time financial data analysis through natural language interactions. “Using Boomi, the agent connects to sources of financial data in ERP systems, databases, and other data sources,” explained David. “It enables users to ask questions about their data through a natural language interface to discover and analyse financial data and trends in real-time.”

This partnership illustrates how Boomi’s AI agents can be applied to solve complex business problems. With Boomi FinTalk, users can query financial data and gain insights into trends and patterns, helping businesses make informed decisions based on real-time information.

Building Custom AI Engines with Boomi

One of the most intriguing aspects of Boomi’s platform is the ability for businesses to create their own custom AI engines. When asked if this is feasible, David confirms that it is indeed possible and provides a detailed explanation. “If you think of the Large Language Model (LLM), you want to write an essay,” he said. “Depending on what it’s been trained on, it generates an initial output.” The true potential of AI lies not just in generating initial outputs but in refining and improving these outputs over time.

The Boomi CTO for APJ elaborated on this iterative process: “But that’s a single pass signature,” he said. “Data already shows out here in the industry that if you actually feed that insight back and refine the draft, and connect to other agents or data sources, it will provide a much better output.” This iterative refinement process is central to Boomi’s User Agent concept, which involves combining multiple agents and data sources to achieve the best possible outcomes.

By leveraging feedback and integrating additional data sources, businesses can enhance the performance of their AI agents. This approach allows organisations to build sophisticated AI solutions tailored to their specific needs.

Accessibility for Non-Developers

A common concern regarding AI tools is their accessibility to individuals with limited coding experience. Addressing this, David emphasised that “One of the things Boomi has been very proud of is being a low-code development platform.” This platform allows users from various departments, including HR, finance, and sales, to automate processes using intuitive drag-and-drop functions.

With the introduction of Boomi GPT, users can interact with the platform using natural language commands. “With Boomi GPT, AI Agents are taking their instincts for the conversation to physical actions,” David highlighted. While more complex integrations might require IT support, Boomi’s low-code platform ensures that a broader range of users can engage with and benefit from AI technologies.

David is very optimistic about the evolving role of developers and IT professionals in an AI-driven world. He views AI as a tool for augmenting, rather than replacing, human roles. The C-suite executive pointed out that governments, such as Malaysia, are developing AI roadmaps and investing in education to prepare the workforce for the future. “Businesses are trying to work out what’s going to be my company’s framework for responsibly adopting AI,” he added. This emphasis on augmentation rather than replacement reflects a broader trend towards using AI to complement human expertise.

Ensuring Accuracy and Reliability of AI Agents

Ensuring the accuracy and reliability of AI agents is crucial, especially for tasks like data classification and integration design. David detailed Boomi’s approach to maintaining high standards: “Boomi has been around for over 20 years, and we’ve leveraged over 200 million integrations for data and best practices,” he says. By tapping into a vast repository of data from Boomi’s global customer base, the platform avoids common issues like hallucination in LLMs, where models might generate inaccurate information due to limited data.

“Our interface will be able to prompt ‘Hey if you’re looking to build out this integration, this is how our customers have done it before,’” he said. This approach ensures that Boomi’s AI agents are built on a solid foundation of best practices and proven solutions.

Building Smarter Integrations

David Irecki’s insights into Boomi’s AI Agent Framework reveal a strategic vision that combines advanced technology with user accessibility and responsible innovation. Boomi’s platform supports the creation of custom AI engines and democratises access to AI technology through a low-code environment. As David emphasised, the future of AI is about augmentation rather than replacement, offering opportunities for developers to expand their roles and for non-developers to harness AI’s power.

With its Marketplace for high-quality AI agents, the Agent Garden for managing and interacting with these agents, and the Registry for tracking AI usage, Boomi is setting a new standard for AI-driven integration and automation. Tools like Boomi GPT and DesignGen help users build and refine integrations, while Boomi Answers and Boomi Pathfinder provide expert guidance and best practices.

As AI technology continues to evolve, Boomi’s commitment to accuracy, accessibility, and innovation positions the company as a leader in the future of AI-driven business processes.

(0)(0)