One Model Blog

AI Agents in Action: Engineering Agentic Workflows for a Smarter HR Workplace

Written by The One Model Team | Oct 11, 2024 8:21:49 PM

In his talk, "Under the Hood: AI-Driven Engineering Workflows for Future of Work," Chris Butler, CEO of One Model, addressed what's coming and how it will impact everything.

 

The key takeaway? AI is about to change the game for productivity by enabling what Chris calls "agentic workflows." Here’s a peek at what that means and why it’s a big deal for your workplace.

The AI Ecosystem is Opening Up

The enterprise AI ecosystem is evolving quickly. Imagine an AI at work that doesn’t just answer questions but can also take action—accessing tools, managing processes, and optimizing workflows. According to Chris, the likes of Microsoft’s Copilot, Apple’s assistant, and other major players like Salesforce are poised to become the AI linchpins of the workplace. Soon, AI won’t just assist; it will be seamlessly integrated into every facet of your digital workspace.

Enter Agentic Workflows

One Model is currently building agentic workflows to turbocharge workplace efficiency. Think of a group of specialized AI agents, each with its own job description, working collaboratively—just like a project team. From gathering data, analyzing it, and critiquing results to creating dashboards, these agents mimic the roles of a traditional data team. The result? Faster, smarter outputs that scale without needing more people.

Chris gave real-world examples of ai agents in action: An AI project manager, data engineer, and analyst worked together to gather compensation data, clean it, and create insightful reports—tasks that normally take days were completed in hours. The AI agents interact with each other in natural language, refine each other’s work, and iterate until the job is done right.

From Four Agents to a Swarm

What started as four distinct agents evolved into a swarm—a scalable network of specialized agents able to handle increasingly complex projects. By shifting to a directed graph model, One Model made it possible for multiple agents to work in parallel, dramatically reducing project time. Chris shared an impressive example: A task that two senior data engineers estimated would take twenty days was completed by AI in just 45 minutes. Another key takeaway is that the more specialized the agents work, the higher quality the output. Therefore, having more specialized agents is better than a few multi-purpose ones.

What Does This Mean for Productivity?

The implications are huge. AI-driven workflows mean fewer manual tasks, faster data processing, and a deeper focus on insights that matter. Companies can double down on their core missions while relying on AI to handle tedious, data-intensive work. Chris predicts that enterprise AI will become the interface we use to ask questions and get work done—a one-stop assistant that pulls insights from different tools and presents them in a digestible way.

Dashboards Are Dead—Almost

In the future dashboards as we know them may become secondary. Instead of static reports, enterprise AI will generate dynamic, on-demand insights and even make recommendations. Dashboards will still exist, but they’ll be an interface controlled by the AI—just one of many tools in the box. The first point of interaction will be the AI itself, which will decide what tools to use to provide you with answers.

Securing the AI Frontier

Chris also highlighted a critical concern: securing enterprise AI. As these AIs gain more access to tools and data, the risk of improper usage grows. HR and People Analytics teams need to partner closely with IT to ensure that the right security measures are in place—because once access is lost, it’s hard to regain control.

Welcome agentic workflows to the team.

Agentic workflows are reshaping the future of work. The enterprise AI of tomorrow won’t just assist employees; it will be an active participant in getting work done—faster, smarter, and more securely. Are you ready to work with your new AI teammates?

 

Are you thinking about using AI? You'll need a solid data platform.

Learn why that is so critical and see how you can achieve success by reading our whitepaper.