How to get the best results from your new AI colleagues
▼ Summary
– Teams will increasingly include both human colleagues and AI agents that can autonomously handle tasks from operations to decision-making.
– Gartner projects AI agent software spending will rise from $86.4 billion in 2025 to $376.3 billion by 2027.
– Fanatics benchmarks AI tools by surveying staff on time savings and task types, finding routine reporting automation frees workers for strategic work.
– Whoop’s Matt Luizzi notes agents excel at disruptive, repetitive questions, allowing teams to focus on proactive, revenue-impacting analysis.
– Synopsys’s Sriram Sitaraman says agents handle junior-level data tasks, enabling smaller teams to use data-driven insights for higher-value decisions.
Your future workplace will include more than just human colleagues. As organizations race to meet ambitious goals, teams are increasingly blending human talent with autonomous AI agents, marking the dawn of what experts call the autonomous business era.
In this new environment, roles once filled entirely by people,from routine operations to strategic decision-making,are being handled by AI agents capable of discovering, negotiating, and transacting independently. According to Gartner, global spending on AI agent software is projected to surge from $86.4 billion in 2025 to $376.3 billion by 2027, underscoring the rapid shift.
Three digital leaders at the Snowflake Summit 2026 shared how their organizations are already deploying agents in production. They emphasized three critical success factors: benchmarking agent performance, staying receptive to new approaches, and focusing on the right problems.
Benchmark Your Tools
Madeleine Want, VP of data at Fanatics, explained that her company actively tracks the value agents bring. Fanatics tests tools, compares features, runs previews, and builds design partnerships to measure impact. “We benchmark how you are using these tools, what tasks you use them for, how much time they save you, and what you do with that time,” she said.
The results show agents free up hours previously spent on routine reporting. “Every business analyst wishes they could focus on strategic work but gets bogged down in routine tasks,” Want noted. “We see staff getting that time back and reapplying it to more human, strategic work,the dream outcome.”
She stressed that agentic AI is still experimental. “This is not a traditional multi-year enterprise transformation. We’re in an experimental phase, so adopt early and try things, but hold it lightly. We need to stay agile.”
Stay Open to New Ideas
Matt Luizzi, VP of analytics at Whoop, saw the potential for agents to reduce time spent answering random business questions,tasks that consumed 50% to 60% of his team’s day. “People want to get those disruptions off their plates, and that’s where agents excel right now,” he said.
Introducing agents has allowed human colleagues to focus on strategic work that adds real value. “We’ve seen real revenue impacts,people identifying issues proactively, root-causing them with AI, and taking action faster,” Luizzi added.
He believes great ideas can come from anywhere. “Some organizations are bottom-up, with junior workers taking risks and adopting new tech. Others are top-down, with leaders identifying solutions that solve team problems.”
Find New Problems to Solve
Sriram Sitaraman, CIO at Synopsys, highlighted how agents are boosting human capabilities by handling large volumes of data. “The concept of the next best action used to be a conversation among humans. Now, AI enables truly data-driven, profitable actions,” he said.
Agents can take over tasks like running queries and creating graphs, freeing junior employees for higher-level work. “You don’t need a team of people having the conversation,just a smaller team looking at a large amount of data,” Sitaraman explained.
He sees agentic AI as a hierarchical progression. “Models will keep pushing tasks downstream to AI, and the complexity they manage will increase. In six months, AI will solve different types of problems,not the same ones as today.”
(Source: ZDNet)
