Event Archive

Daniel Acton CTO, Accelera Digital Group

Autonomous AI agents can make a massive contribution to the speed and efficiency of an organisation’s processes. However, in order to reap the benefits, data readiness is essential. Daniel Acton, CTO, Accelera Digital Group, outlines why an autonomous AI agent requires a higher standard of data integrity and access than a basic chatbot does, and how to achieve this. 

“AI agents need to be held to higher standards than basic chatbots because the tasks they perform are more complex and they are often authorised to work more closely with your data,” Acton says. “An autonomous agent knows to look for input and create output. It is task-oriented. When the parameters are correct and the data is well organised, its efficiency and performance will add immense value to your business. If the parameters are incorrect or the data it’s prompted to work with is not valid, then you risk confusion and errors creeping in.”

Continuous improvement must be a cornerstone of your thinking when it comes to AI. The better the model that backs the agent, and the better the augmented data, the more its results will improve over time. Poor data cannot lead to good results. Your security posture needs to be strong, and must have guardrails in place that control the agent’s permissions. For example, defining what data a customer service AI agent can access so that it cannot erroneously overstep its boundaries is key to good security.

“Understanding your data governance needs is the first step. The next step is putting in place the controls to make that happen. For example, there are different data retention tiers, including ‘the right to be forgotten’, where any data subject may ask the processor to permanently erase their data. Many organisations must adhere to strict industry regulations in this regard. With the massive value that autonomous AI agents can offer, organisations need to be certain that their data is secured and well managed before they start using them,” Acton says.

The Data Warehouse (or Data Lakehouse) that your AI agents rely on must take care of critical processes, like ETL (extract, transform and load), governance, storage, and more. A strong data foundation includes your Data Warehouse (or Data Lakehouse) as well as how your business (and AI Agents) can use this data (which points to a Data Strategy, Data as a Product, and Data Contracts). ADG can assist with creating a strong data foundation to help CIOs move from legacy data silos into a framework that can provide the data for autonomous agents. This enables you to correctly organise and secure all your data for AI agent readiness.

“At ADG, we work with our customers on every step of their data foundation, including the strategy, governance aspects like access, retention policies, the right to be forgotten, automation, and governance aspects like metadata and data dictionary. The aim is to have simple access, with well-defined semantic meanings that are good for humans and agents alike. This allows AI agents to follow a clearly defined path,” Acton says.

“When it comes to taking the first step, our advice is to find a use case that you are comfortable trying – like approvals in an existing business workflow based on a set of criteria, or automating a certain set of software, with human escalation and overview,” Acton says. “Assign the agent to a specific task, with clear parameters in place. Then assess the convenience and value that it adds.” With the correct foundation in place, you are set up for success. Once organisations see the value in having agents handle the mundane aspects of business processes, they can choose how and where to expand their scope for greater efficiency, cost saving and future growth.

https://www.adg.io/