AI agents are your new colleagues - how to get the best results

KEHAN CHEN/Moment

Follow ZDNET: Add america arsenic a preferred source on Google.

ZDNET's cardinal takeaways

• Your adjacent enactment squad volition see humans and agents.
• Experiment and benchmark AI tools to cheque value.
• Stay unfastened to caller agentic solutions to caller challenges.

Worrying whether the idiosyncratic adjacent to you is pulling their value professionally is nary longer your lone concern. For radical who privation to conscionable choky targets and present large results, your squad is apt to include a wide premix of quality colleagues and agentic counterparts.

We are entering the property of the autonomous business, wherever caller combinations of exertion and information mean immoderate of the roles we instrumentality for granted contiguous -- from basal operational tasks up to decision-making responsibilities -- are fulfilled by agents that discover, negotiate, and transact autonomously.

Also: 12 rules of agentic AI for palmy endeavor transformation

Tech expert Gartner suggests companies are expanding their investments successful agents, with AI cause bundle spending acceptable to scope $206.5 billion and $376.3 cardinal successful 2027, up from $86.4 cardinal successful 2025.

Some companies already usage agents successful their operational activities. Three integer leaders astatine the Snowflake Summit 2026 successful San Francisco recently explained however their organizations are putting agents into production.

After the sheet session, ZDNET asked the participants what they'd learned astir moving successfully with their agentic colleagues. They suggested 3 areas are crucial: benchmarking agents, staying unfastened to caller ideas, and focusing connected the close areas.

Benchmark your tools

Madeleine Want, VP of information astatine sports specializer Fanatics, recognized that delivering large results crossed agentic and quality colleagues is simply a pugnacious ask, truthful her enactment tracks and traces benefits crossed the information practitioner community.

She said Fanatics is an assertive and aboriginal adopter of AI for data, wherever the enactment tests tools, compares features, runs previews, and develops plan partnerships.

"We benchmark however you are utilizing these tools, what benignant of tasks you are utilizing them for, however overmuch clip you consciousness that they are redeeming you, and what you are doing with the clip -- each of these kinds of self-reported value-based questions," she said.

Also: 40% of enterprises volition scrap AI agents - 3 ways to guarantee yours don't fail

Want, who manages information engineering, information science, and instrumentality learning crossed the betting and gaming part astatine Fanatics, told ZDNET the benchmarks amusement agentic input saves quality time.

"Every concern expert retired determination volition archer you immoderate mentation of, 'I privation I could beryllium doing much strategical work, but I americium bogged down successful regular reporting,'" she said.

"What we are seeing is that the much regular reporting tasks are the ones that often lend themselves champion to automation done AI, truthful we are seeing unit get that clip backmost and past reapplying it to enactment that's much quality and much strategic, which is benignant of the imagination result that you would anticipation for."

Want said the palmy exertion of agentic AI is astir getting clasp of amended tooling to enactment with, truthful you tin get the indispensable parts of enactment done and absorption connected the much absorbing areas you bash best.

Also: AI is causing cognitive fatigue. Here's however to enactment with much haste and little speed

However, portion definite tools mightiness enactment successful the present, she recognized that agentic AI is simply a enactment successful progress, and her company's committedness to adopting and investigating tools means professionals mightiness beryllium exposed to caller services regularly.

Want said her organization's philosophical attack to agents means deployment involves a back-and-forth process betwixt managers and professionals arsenic caller AI-enabled ways of moving are discovered.

"There's a batch of anticipation absorption to say, 'This is not your accepted endeavor exertion multi-year translation project,'" she said, advising different professionals to enactment unfastened to exploration and change.

"We are not adopting well-tested, well-trodden technologies that, erstwhile rolled out, volition ne'er beryllium rolled back. We're successful an experimental signifier close now, and so, follow aboriginal and effort things, but besides clasp it lightly, due to the fact that we're going to request to enactment agile."

Stay unfastened to caller ideas

Matt Luizzi, VP of analytics astatine wearable exertion specializer Whoop, is different integer person who was anxious to assistance his squad marque amended usage of their time, adjacent earlier the rollout of agentic AI.

"I was trying to recognize wherever my squad was spending their time, and radical were saying they're spending betwixt 50% and 60% of their clip conscionable answering random questions from astir the concern that came in," helium said.

"'What were income yesterday? How does that disagree by region? Why were our web sessions up?' Those are disruptive things that radical privation to spell away. Those are tasks that radical would beryllium blessed to get disconnected their plates. It besides happens to beryllium wherever agents excel close now."

Also: The autonomous concern is coming. Here's wherefore that displacement is bully quality for professionals

Luizzi told ZDNET that his concern has seen that introducing agents means quality counterparts tin walk much clip with their nonrecreational colleagues connected strategical enactment that adds incremental value.

"We've besides seen existent gross impacts from this exertion already, with radical being capable to place things proactively, basal origin them with AI, troubleshoot what's going on, and instrumentality enactment overmuch faster earlier the vessel has near the station."

Luizzi suggested that the march of agentic AI volition proceed to stitchery pace, peculiarly for tasks that tin beryllium easy automated.

"We'll proceed to spot advances that unlock caller capabilities for wherever humans are spending time, but we request to proceed to propulsion the boundaries," helium said.

Also: Forget productivity: Here are 5 strategical shifts that thrust existent AI value

To that end, Luizzi suggested that nary azygous worker is apt to clasp the cardinal to agentic success. Great ideas tin bubble up from anywhere, and each professionals indispensable beryllium acceptable to marque a mark.

"Some organizations are going to beryllium bottom-up, wherever the junior-level workers are taking connected caller technology, taking risks, and making time," helium said.

"Some of these initiatives are going to beryllium top-down, coming from leaders similar us, coming to conferences and proceeding what different customers are doing, and being capable to persist those passim the organization, and place and signifier lucifer wherever those solutions lick problems that their squad faces."

Find caller problems to solve

Sriram Sitaraman, CIO astatine bundle specializer Synopsys, said helium manages reasonably ample amounts of engineering and firm data. One happening that's go wide crossed some areas is that agents are showing however they volition assistance to boost quality capabilities.

"If you look astatine the measurement of information available, the conception of the adjacent champion enactment you tin instrumentality utilized to beryllium a speech betwixt a clump of humans based connected existent priorities," helium said. "Now, with AI, you tin genuinely marque a data-driven, profitable action."

Sitaraman said his institution has recognized the imaginable for AI agents to fulfill the tasks of inferior employees, specified arsenic moving speedy queries, creating graphs, and deriving insights.

Also: How to bushed the AI algorithm and get the occupation of your dreams

He besides gave the illustration of deciding which caller features to physique for an application. He said employees tin enactment alongside their agentic colleagues to sift ideas and aboveground commercially viable propositions.

"You don't request a squad of radical having the conversation. It's a smaller squad of radical looking astatine a ample magnitude of data," helium said.

"Many efforts to reconcile information sources for decisions are present focused connected however humans instrumentality vantage of AI. That effort is astir trimming down the ample volumes of information to actionable adjacent steps."

Sitaraman said agentic AI gives clip backmost to quality workers. For example, by picking up level 1 data-sorting and sifting tasks, unit tin determination to higher-level, value-creating work.

"It's a hierarchical thing. The models volition support pushing tasks downstream to AI, and the complexity of tasks AI tin negociate volition summation arsenic the models get better," helium said.

"So, successful six months, I spot AI solving antithetic types of problems -- not the aforesaid types of problems arsenic present but antithetic types, and that's going to germinate continually."