By
The hype around AI agents in the enterprise has reached a fever pitch. After spending time at Gartner I have seen many Hype Cycles but this is different.
These autonomous software agents are touted as the next leap in productivity and automation. In fact, AI agents are becoming ubiquitous in tech circles, and “every single investor in Silicon Valley is bullish on startups building them”, as TechCrunch recently noted. From automating customer service to streamlining internal workflows, AI agents promise to handle tasks independently and at scale. It’s no surprise, then, that enterprise adoption of AI has accelerated dramatically with 72% of organisations now using AI in at least one business function, up from 50% in prior years.
Global AI investment is soaring as well, with spending surging to $13.8 billion in 2024 (a six-fold jump from 2023). Tech giants like Salesforce and Oracle have poured billions into AI agents, aiming to boost margins and optimise costs. Clearly, a new era of AI-driven “digital workforce” is on the rise in the enterprise.
However, amid this excitement, it’s also becoming evident that success with AI agents is far from guaranteed. As someone who works closely with enterprises on digital talent and AI, I believe the biggest hurdle ahead won’t be building or buying AI agents, it will be matching the right agent (or team of agents) to the right task. This is an old problem for tech leaders which is about to explode.
In this blog, I’ll explore the data behind the rise of AI agents in business, what Gartner and others say about why such projects succeed or fail, the emergence of AI agent marketplaces, and why matching AI agents to enterprise needs is poised to be a challenge as significant as hiring the right people. Finally, I’ll share how we at Gigged.AI are approaching this challenge.
An AI agent is essentially a software program that can make decisions and perform tasks autonomously using an AI model. Thanks to breakthroughs in generative AI and large language models, these agents are proliferating fast.
Gartner predicts that by 2028, at least 15% of day-to-day work decisions will be made autonomously by AI agents, up from essentially 0% in 2024. One-third of enterprise software applications are expected to include some form of agentic AI by 2028 (up from less than 1% today).
While vendors and startups race to label their offerings as “AI agents”, Gartner analysts warn of “agent washing,” noting that only about 130 of the thousands of purported Agentic AI vendors are truly delivering real agent capabilities.
According to Gartner, only about 48% of AI projects make it from prototype to production. Even when they do, it takes an average of 8 months to scale a prototype for real-world use. Other estimates suggest that over 80% of AI projects fail to achieve their intended ROI, roughly twice the failure rate of traditional IT projects.
Gartner also forecasts that more than 40% of Agentic AI projects will be canceled by 2027 due to escalating costs, unclear business value, or inadequate risk controls. Many projects lack clear objectives, face data quality issues, and suffer from cost overruns.
43% of organisations cite poor data quality or readiness as a major obstacle to AI success. Many pilot projects never scale because the data is siloed or unreliable, or the business case isn’t solid.
In 2020 we saw large enterprises over-hire in tech talent. This then contributed to mass layoffs in 2022 which are still going on today. I predict something similar with AI Agents as companies build and buy thousands of AI Agents but do not actually have the demand for them. Matching a project or task to the right skill is super important, whether that skill comes from a human or AI agent.
Amazon Web Services (AWS) recently announced a dedicated AI agent marketplace, allowing startups to offer agents directly to AWS customers. Partners like Anthropic will let enterprises browse and install AI agents through the AWS ecosystem.
Google Cloud, Microsoft, Salesforce, and ServiceNow have all launched similar marketplaces, aiming to become app-store-like hubs for AI agents.
These marketplaces aim to solve distribution issues by centralising agent discovery and integration. But the success of these platforms, especially for niche or high-skill enterprise use cases – remains to be seen.
Enterprises won’t struggle to find AI agents, but they will struggle to deploy the right agent for the right task. It’s not unlike workforce planning: finding someone with the right skills, experience, and fit is hard.
As AI agents become part of the digital workforce, companies will need to treat them like human workers – understanding their capabilities, limitations, and how they fit into broader workflows. Success will depend on the orchestration of skills across both humans and machines.
At Gigged.AI, we’ve spent the last 4 years helping enterprises intelligently match tasks with the right human talent. We have our own agent, Sia (launching out of Beta very soon), that helps our clients get the answers they need.
Now, we’re extending that capability to include AI agents.
Our Skills Intelligence Platform uses AI to understand task requirements and match them to internal employees, freelancers, or AI agents. By applying the same profiling to AI agents as we do to human workers, we help enterprises build hybrid teams optimised for delivery.
For example, a finance reporting task might be handled by Agent X (data analysis), Agent Y (report drafting), and Jane from the finance team (review and context). Our platform facilitates this orchestration by treating all resources, human and digital, as part of the same talent ecosystem.
AI agents are here, and they’re becoming a central part of enterprise operating models. But buying or building them isn’t the hard part. Matching them effectively and integrating them into real workflows will be the real challenge.
As Gartner’s data shows, success hinges on execution, not hype. Enterprises that treat AI agents as part of their talent strategy, and invest in intelligent matching systems, will lead the way. At Gigged.AI, we’re building the Skills Intelligence infrastructure to make this future possible.
Gigged.AI helps enterprise technology leaders map skills, mobilise their people & AI agents and fill critical gaps with on-demand talent.
Hire vetted developers, data scientists, AI specialists and more in days. Use outcome-based or T&M SoWs to keep projects moving without adding permanent headcount.
Map every skill and agent in your function to a comprehensive framework. See real-time capability and deploy teams to projects faster.
Join 25,000+ freelancers using Gigged.AI to find remote work with global enterprises. Create a free profile, get matched to gigs automatically and get paid securely.
Something else? Please submit your contact details using this form and we’ll be in touch.
