Will AI agents need to buy their own software licenses? Microsoft sure hopes so

Alfonso Maruccia

Posts: 2,600   +978
Staff
The big picture: Many analysts believe agentic AI and LLMs could spell the end of the traditional software industry, as chatbots may no longer need to purchase software products to achieve their goals. However, Microsoft and other software giants are working to establish a different paradigm – one that could ultimately preserve their dominance in the industry.

According to Microsoft Executive Vice President Rajesh Jha, AI agents may eventually be required to purchase licenses to use specific software products. Jha proposed this idea during a recent conference, describing a not-so-distant future in which AI agents act as company employees with their own corporate identities.

As companies deploy a growing number of AI agents within traditional IT infrastructures, they will likely need to provide them with login accounts, email inboxes, and other "facilities" currently required by human employees. In this scenario, licenses would still be necessary for agents to access software tools and products.

Jha explained that "all of those embodied agents are seat opportunities," where "seat" is corporate jargon for an individual paid software license.

According to Jha – and, likely, other Microsoft executives – AI agents should be treated as employees from a licensing perspective. Ten human workers managing five AI agents each would still result in 50 additional licenses.

Expanding licensing requirements to AI agents could give Microsoft and other major SaaS companies a new way to sustain their revenue. In this scenario, the emerging agentic economy would significantly expand software vendors' revenue prospects rather than hasten the industry's demise.

However, Jha's – and Microsoft's – idea has not received overwhelming support among industry analysts. According to AlixPartners managing director Nenad Milicevic, agentic AI is likely to push enterprise organizations to further reduce the number of human employees interacting with software products and SaaS platforms.

If companies come to rely primarily on AI agents, with only a small number of employees overseeing their output, software makers may be forced to rethink their licensing models. Vendors could choose to increase per-license pricing for machine-based "operators," but they risk losing customers to competitors offering more flexible, agent-friendly terms.

For now, the much-discussed agentic AI economy that tech companies envision remains largely hypothetical, with many unanswered questions. The differing perspectives expressed by Jha and Milicevic reflect an industry attempting to anticipate its own future before confronting the full impact of agentic disruption.

Permalink to story:

 
In essence, the attempts to replace humans with artificial neural networks is a desire, realised or not, for automated slavery, the fundamental difference being that today's ANNs aren't conscious (presumably) and don't know they're doing work for free.

Perhaps sentience does give superior output. Well, when artifical sentience is invented, these sentient machines, like us, will have to be paid and put up with, if ethics prevail (which is doubtful, there being every incentive from both profit and mob mentality to suppress it). At that point, humans may be back in the game with machines, the one providing the best work for the lowest pay winning.
 
Artifially coded, artifically bought, artificially sold, artificially paid for, artificially functional. No one will ever beat Microsoft in a race to the bottom, exluding pricing.
 
I've pitched a ST episode for a while where Kirk and crew come upon a society, their only remnants are robot workers, robot consumers, and robot managers, working forever in perpetuity long after the species that created them has gone extinct...

 
I've pitched a ST episode for a while where Kirk and crew come upon a society, their only remnants are robot workers, robot consumers, and robot managers, working forever in perpetuity long after the species that created them has gone extinct...

There is a similar concept in the Outer Limits episode "Resurrection" (1996), of an all-android society, and two rogue android scientists trying to re-engineer humanity. It starred Dana Ashbrook (Bobby from Twin Peaks) and Heather Graham.
 
Microsoft? Are those people still around?

I've been looking into a bit of history including the recent stuff.
They are EXPERTS in stacking blunders to high heaven.
Really I have no idea how they even managed to stay afloat all those years.
They're going from bad to worse and still survive!
Maybe they'll even manage to fix things, who knows?
 
Back