Why agentic AI will be a huge money-spinner for big tech providers

The multi-trillion-dollar global services market is ripe for change, according to new analysis from venture capital firm Sequoia Capital, and enterprises driving agentic AI are perfectly poised to capitalize on this opportunity.

With the rise of agentic AI gathering pace and pushing automation capabilities, the VC firm suggests the technology could have a transformative impact on the highly lucrative market.

Where the cloud transition was fueled by software-as-a-service (SaaS), Sequoia believes the AI transition will be fueled by ‘service-as-a-software.’ In other words, labor will be sold by software companies, rather than software to complete labor.

This will be made possible by agentic AI, an advancement of AI through which users can program the technology to complete entire tasks. Rather than requiring a human to continuously deploy or redeploy the AI, agentic systems are able to act autonomously.

If SaaS is taken as a test case, then this stage of the AI lifecycle could lead to some dramatic shifts in the tech market landscape – agentic AI can complete services independently, opening up the services market to software providers.

In 2010, the software market was valued at $350 billion, with cloud software constituting $6 billion. By 2024, the market had nearly doubled in value to $650 billion, with cloud making up a much larger $400 billion share.

Currently, AI software firms make up just $3 billion of the $10 trillion services market, though a continued push on agentic technology is likely to boost that and many firms are already dipping their toes in.

Salesforce recently rolled out Agentforce, for example, a platform that allows Salesforce customers to create customizable, autonomous agents that can independently complete tasks.

Speaking at the company’s annual conference in September, CEO Marc Benioff hailed AI agents as the next step in the evolution of AI, and appeared confident the launch of its Agentforce service will mark a significant change in how organizations leverage the technology in years to come.

Microsoft also unveiled a similar feature called Copilot agents in recent weeks. In much the same way, Microsoft’s tool will allow enterprise customers to fully automate tasks.

While it’s early days, Google has already claimed its own agentic offering delivers a strong return on investment (ROI). During the firm’s Gemini at Work event earlier this year, the firm cited customer Bell Canada, which has reportedly saved $20 million using Google’s platform.

How significant will agentic AI be? 

Agentic AI is starting to gain mainstream traction with announcements from big-name tech companies, though iManage business consultant Jack Shepherd thinks the industry needs to temper its expectations somewhat.

“The reality of what these tools can offer will probably not match the hype that currently exists around them. This is because, in the early stages of hype, it is common for people to get excited about the technology but forget about the real-world applications to which it can be put to use,” Shepherd told ITPro.

A stumbling block will be personal preference, Shepherd said, referring to lawyers as an example of those who typically want as much control over their workflows as possible and need personal preferences built into new technologies.

“The premise of agentic AI is that a user can enter a prompt, and then the system works out for itself how the user’s end goal should be achieved. However, in the case of lawyers, they often have a very specific idea of how they would like the end goal to be achieved. It is not apparent how these two things can be reconciled with each other,” Shepherd said.

In other industries there is a need for consistency, Shepherd added, and it may become difficult to predict outcomes and costs in certain organizations if workers are doing things differently.

“If agentic AI is deployed without any form of process engineering beneath it, there is a significant risk that businesses cannot operate effectively,” Shepherd said.

Dr Stefan Leichenauer, VP of engineering at SandboxAQ, told ITPro that it’s important to remember the limitations of agentic AI and bear in mind the necessity of humans.

“In the near term, it will be much more useful to view AI as a tool that enhances and complements human capabilities rather than an independent intelligence. Perhaps we will get to fully autonomous AI in the more distant future, but technologically, we are still a few breakthroughs away,” Leichenauer said.