Follow ZDNET: Add us as a preferred source on Google.
ZDNET’s key takeaways
- AI will enter a new phase in 2026, analysts said.
- Businesses will better leverage the tech and see results.
- AI agents and commerce opportunities will be key.
The AI hype fueled by the launch of ChatGPT at the end of 2022 has only accelerated. Organizations, however, have yet to see much ROI on their mounting investment in the technology — but experts say that wait may be over in the new year.
Based on promises of AI’s potential to dramatically optimize operations through new developments in the space, including models that are smarter, cheaper, multimodal, better at reasoning, and even autonomous, business leaders have funneled money into related expenses. Global corporate AI investment reached $252.3 billion in 2024, and US private AI investment hit $109.1 billion, according to Stanford data — it’s safe to assume those numbers will only continue to grow.
Also: Why AI agents failed to take over in 2025 – it’s ‘a story as old as time,’ says Deloitte
But a look back at 2025 reveals a common thread: AI’s potential to dramatically optimize operations has not yet been realized across the board. Most memorably, a now-infamous MIT study found that 95% of businesses weren’t seeing an ROI from their generative AI spend, with only 5% of integrated AI pilots extracting millions in value. While the criteria for returns are narrowly defined, which partially explains the high percentage, it is still indicative of a wider trend.
“So far, a small group of leaders have converted AI into outsized value — new revenue pools, new business models, and real valuation premiums — while most others have settled for ‘respectable but modest’ returns,” said Dan Priest, US chief AI officer at PwC.
Yet, Priest adds that he thinks the new year will finally see AI’s value gap start to close, a position held by nearly every expert ZDNET spoke to.
A shift in ROI
Priest mostly attributed this forthcoming expansion to the precision that CEOs and other business leaders will have to bring to their AI projects by identifying a few high-impact areas where AI can “reshape the economics of the business” and pursuing those with focus.
Also: This company’s AI success was built on 5 essential steps – see how they work for you
China Widener, Deloitte vice chair and US TMT industry leader, echoed this sentiment, claiming that the upcoming year will shift from “heavy AI investment that remained stuck in pilots” to meaningful changes for enterprises.
“In 2026, competitive advantage will come not from simply adopting AI, but from orchestrating it — translating innovation into sustained ROI and new forms of business value,” said Widener.
Application over evolution
It is notable that in both of these predictions, experts highlight that the shift doesn’t lie in the evolvement of the technology itself, but rather in how business leaders approach implementing AI into their businesses. How will that get done? There are several key considerations for businesses, starting with the adoption of AI agents.
For instance, Widener suggests that embracing AI’s agentic capabilities will enable business leaders to meaningfully rethink how teams operate, as well as how they carry out work and generate growth.
Also: The fix for messy AI agent ecosystems might finally be here – and it’s open source
In theory, the value of AI agents for businesses is simple: these AI assistants can perform tasks that humans can, but without human limitations (such as needing breaks), while also collaborating with each other to efficiently carry out tasks. In practice, however, that reality is somewhat more challenging to implement.
AI agents
2025 was touted by many as the year of AI agents. Yet, as revealed by Deloitte’s Tech Trends report this week, the technology did not take off this year despite the hype and promise.
In particular, Deloitte’s 2025 Emerging Technology Trends study, which surveys 500 US Tech Leaders, found that 30% of the surveyed organizations are exploring agentic options, with 38% piloting solutions and only 14% having solutions ready to deploy. The number of organizations actively using the systems in production is even lower, at 11%.
Also: AI could double the US economy’s growth rate over the next decade, says Anthropic
Gartner has released similar data saying that over 40% of agentic AI projects will be canceled by the end of 2027, due to factors such as escalating costs, unclear business value, or inadequate risk controls. Even still, Gartner analyst Arun Chandrasekaran coined 2026 as the year of “Operationalizing AI agents.”
“While AI agents are becoming increasingly common as pilot projects, most enterprises are struggling with moving them into production,” said Chandrasekaran. “Ensuring a robust control plane for managing agent lifecycle, instituting governance to secure, red-team, validate and observe agents and building stateful multi-agent systems are all major goal posts for the industry to improve on in 2026.”
Also: AI agents are already causing disasters – and this hidden threat could derail your safe rollout
The firm is also bullish on the value AI agents will bring to businesses, predicting that at least 15% of day-to-day work decisions will be made autonomously through agentic AI by 2028, up from 0% in 2024.
Agentic commerce
AI agents have the potential not only to optimize internal business operations but also to enhance how people execute everyday tasks. For example, one of the most buzzworthy topics related to AI agents is AI for commerce.
In their simplest use case, AI agents can help users select the product they need and add items to their cart. In their ideal state, AI agents will be able to complete transactions on users’ behalfs, which could come in handy when purchasing a product at a certain price point or when avoiding tedious tasks like travel booking.
Also: Should you trust AI agents with your holiday shopping? Here’s what experts want you to know
The latter, more advanced use cases may just be possible in 2026, according to Ken Moore, chief innovation officer at Mastercard.
“In 2026, two powerful forces will converge — AI-driven autonomy and the evolution of trust — as agentic commerce moves from early adoption to scale,” said Moore. “Consumers will shift from manual operators to strategic orchestrators, delegating routine decisions to AI like replenishment or travel booking.”
Education and upskilling
Beyond agents, a central puzzle piece in how businesses will successfully implement AI is proper education. Forrester predicts that by 2026, 30% of large enterprises will make AI fluency training mandatory to lift AI adoption and reduce risk.
This is a major departure from what we have seen thus far. Deloitte found that only 7% of AI spend goes to changing the culture and training, and learning. An October 2025 Wharton study also found that investment in training is softening, dropping eight percentage points year over year.
Also: The great AI skills disconnect – and how to fix it
This lack of adoption is an impediment to successful AI implementation, with Forrester data showing that 21% of AI decision-makers cite employee experience and readiness as a barrier to adoption. Kim Herrington, a Forrester senior analyst, added that an improperly trained workforce is a recipe for risk.
“AI runs on data, and employees shape that data every day (often without realizing it),” she said. “Poor literacy and fluency lead to poor inputs or behaviors, which cascade into flawed decisions or poorly trained AI models that can rapidly scale access to misinformation.”
Herrington said that mandatory training will help remind employees that AI outputs are capable of making mistakes, as well as how to best use them, which can also build their confidence in using the tools.
Timeline
While a lot of AI delivery predictions for 2026 seem to rely on AI agents, it is worth hedging expectations, as the change won’t happen overnight or be seamless.
Also: 5 ways to prevent your AI strategy from going bust
“Agents will still be imperfect, and that’s okay,” said Priest. “The difference in 2026 is that more companies will have real benchmarks, clearer guardrails, and a repeatable playbook. Combined with a tighter, top-down focus on where agents are deployed, that’s what will turn agentic AI from experimentation into real enterprise transformation.”