How Is Agentic AI Changing the Workforce?

Today, AI agents can control your smart-home devices, or help you find products and compare prices, monitor your health metrics, remind you to take your medications and suggest lifestyle changes based on your personal health data.

But an topic I hear in closed rooms, events and see in top publications is Agents as part of the Workforce. Agents, not Assistants, not RAG, not a custom GPT nor multiple Digital Twins. Agents. In fact, I am hearing that they have now matured. They are yesterday's conversation in certain circles.

And the next big thing is autonomous agents. To develop now. To launch in 2026. To double the existing workforce. 

And this is not claimed by a Founder of an overhyped startup presenting a pre-recorded miracle software promising a $1 billion revenue generating business with a single employee.

This is coming from Fortune 500 behemoths that are in charge of the deployment.

Big Tech Has Gone Big on AI

Some of the big players using agentic AI include software giant Microsoft (MSFT), Alphabet's (GOOGL) search engine Google, and IBM, are integrating it to automate tasks, improve customer experiences and accelerate decision-making.

Google is banking on the ecosystem angle. “You need a comprehensive and integrated platform that brings all your company’s data, tools, and people together in one secure place.” They are differentiating themselves vs single tools with a single use vs the pre-existing base they have that they would only need to connect. And yes, some of it is not as good as some single tools out there... yet.

AI promises to radically remake the entire shopping experience with agentic commerce - shopping powered by AI agents acting on our behalf - represents a seismic shift. “It moves us toward a world in which AI anticipates consumer needs, navigates shopping options, negotiates deals, and executes transactions, all in alignment with human intent yet acting independently via multistep chains of actions enabled by reasoning models.”

To some, this sounds futuristic. To Marketers, impossible without a million caveats. To technical teams, a nightmare. And to CEOs, a dream, and a boat they want to get on asap. And if you are a Performance Marketer that worked on Automation 14+ years ago knowing how intelligent scraping works, data enrichment, data structuring, etc. you would go far and beyond to learn who packaged this all into the doomsday bubble that is being discussed today. What a Marketer - when only 10% of AI agents are new technology.

Overall Impact on E-Commerce

By 2030, McKinsey said, the US business-to-consumer retail market alone could see as much as $1 trillion in revenue from agentic commerce, with global projections reaching as high as $3 trillion to $5 trillion.

ll kinds of businesses - brands, retailers, marketplaces, logistics and commerce service providers, and payments players - "will need to adapt to the new paradigm and successfully navigate the challenges of trust, risk, and innovation.” “Agentic commerce requires a fundamental rethinking of how value is created, captured and delivered".

But where do you start?

1. Map tasks and outcomes, not just roles

Start your workforce planning by defining the outcome you need. Be specific - what do you want humans to do vs agents to do. Break it down by tasks. Then build the right Agent vs Human combination to deliver on them.

2. Understand what you actually need

   
   

Map those tasks to tools. If you are not part of communities or groups where you can ask, find them. Don't risk wasting your budget. Avoid multi-year contracts, unless you know for certain they will develop product changes and features on demand (and you really need to understand how your requests will be prioritized before you sign). Understand every single line of cost and research non-variable models.

3. R&R needs to be crystal clear

If you don't know what a RASCI is, you may want to investigate it to better understand how AI can change the workforce. If lines blur as a human organization scales, roles and responsibilities need to be defined accordingly, or you will find a lot of challenges along the way. And it will definitely be exacerbated with AI Agents in the mix. You may have the best technology in place, but if it is not adopted the right way, it can be a detriment to ROI.

You need to stack the cards right, because I have also seen cases where AI agents are deployed and "pushed" to humans without providing any upskilling support. Just like any other MarTech, if you want 100% adoption, training and "champions" are essential.

4. It's not always clear how AI agents are represented in financial reporting

In some cases they sit under the Tech stack, and in some more recent insights, they sit under headcount. They are presented as part of the org. We are starting to get ready for what in some cases is being called the "robot tax" as over the past year, legislators of all political leanings have proposed ideas to mitigate the potential harms of AI to workers.

• Requirements for major companies and federal agencies to report artificial intelligence-related layoffs to the U.S. Department of Labor.

• DOL to then compile data on AI-related job effects and publish a report available to the public and Congress.

• While these actions remain in the early stages of the legislative process, several states have moved more quickly with New York and California requiring employers to take actions such as conducting bias audits, creating AI risk management policies or notifying applicants about their use of AI in hiring.

5. Keep a close eye on how AI is changing the workforce from a legislation perspective

Many companies, just like it happened with Social Media, have jumped on the opportunity of massive demand with no regulation. But it won't take long until proper guardrails are put into place. Stay updated and work closely with your HR and Legal partners to ensure you stay ahead of the curve when you implement AI Agents in your organization, at Business level, in the Marketing org or other departments.

6. Track everything

When you have a background in Performance Marketing, you know very well by now that you need to track everything. Not every single detail is necessary, but I have surprised myself many times tracking metrics regularly that weren't a must-have and be tremendously glad I did because I have needed them later on. I had plenty of historical data to support an ask or push for change that had a significant impact at Business level.

Track everything.

Monitor performance, stay close to model training and align new updates to the strategy - within measure - as the technology evolves.

7. You don't have a failing Business. You have a failing Strategy.

No organization - especially when we are talking about hundreds of team members - will adopt a tool perfectly overnight. None. Adoption takes time. Consumer cases can be faster, especially if the UI is intuitive. But perfect Business adoption, depending on the size, can take months. Influencers jumping on Consumer use cases, won't cover how to connect applications to an automation workflow builder, helping you find where API keys are, ensuring that filtering and prompts are optimal to have an output that doesn't need to be refined.

Invest in practical hands-on training for your teams, relationship skills and critical thinking. We are risking the next generation not knowing how to do something when they click on a button, and the tool doesn't work. We adopted smartphones but also need to be able to live without them. 

There is a Risk in AI Efforts

Venture capital investment in agentic AI between fourth-quarter 2024 and first-quarter 2025 more than tripled (+265%).

But months ago I had already heard significant skepticism on this topic from an Investor as we discussed ChatGPT wrappers.

Gartner said that by the end of 2027, more than 40% of agentic AI projects will be canceled. “The main reasons include escalating costs, unclear business value and inadequate risk controls”.

They are cautioning Business leaders to resist the temptation to deploy agentic AI indiscriminately and instead focus on cases in which agentic AI’s unique capabilities create measurable business value. They recommended that business leaders first evaluate where agentic AI is truly the best fit.

“Not every workflow or customer pain point requires the autonomy and complexity of agentic AI. In many cases, traditional automation techniques, machine learning, or even traditional software may deliver equal or greater value at a fraction of the cost and risk". 

While reading further about how AI is changing the workforce, you will find that in many cases, human input will save you a lot of time. And tokens or credits.