Prepare for the age of frumpy, but, functional AI!

In 2025, the buzz about agents was to the world of data, analytics, and AI what the engagement of Taylor Swift and Travis Kelce was to the world of entertainment —undeniably sexy. Yet, just as that power couple will inevitably fall into a routine of quotidian domesticity, so too will AI lose its sheen, trading its tiara for a hard hat in 2026. 

Hype and hyperbole will give way to dowdiness and drudgery as the AI bubble begins to burst, companies prioritize AI governance and literacy, and agents take on data grunt work. In 2026, the art of the possible will succumb to the science of the practical as we predict:

* Enterprises will delay 25% of AI spend into 2027. AI value is failing to land.
Only 15% of AI decision-makers reported an EBITDA lift for their organization in the past 12 months, and fewer than one-third can tie the value of AI to P&L changes. 

The payback expectations are high — 85% of C-level AI decision-makers expect a positive return within three years to consider an investment successful. Given this, we believe thar CEOs will pull more CFOs into AI deals in 2026. Finance-gated decisions will slow production deployments and decimate proofs of concept, leading enterprises to delay 25% of their planned spend into 2027. 

The disconnect between the inflated promises of AI vendors and the value created for enterprises will force a market correction. As demand slips, utilization will lag, cost per useful inference will remain high, and providers will chase fill rate with discounts and oversized commitments. Savvy buyers should capitalize on this supply side frailty by manipulating the levers of AI cost while refocusing investment on top- and bottom-line impact.

As firms employ agentic data and analytics, data teams will reduce head count by 25%. 
While there is debate about whether AI agents are coming for developers, there is another, more surprising group at risk of obsolescence — data professionals. 

Our data indicates that plans to hire data engineers, data scientists, and analysts to support AI are slowing. The percent of AI decision-makers who reported that their IT teams planned to hire more analysts this year to support AI projects was 9% less compared to 2024, and the percentage planning to hire more data scientists and data engineers fell by 6% and 5%, respectively. 

This trend will accelerate as data leaders infuse AI agents into the orchestration of the entire data, analytics, and insights lifecycle, giving rise to a new world of agentic data and analytics. As these AI agents autonomously discover, clean, and analyze data across silos, data teams will become more productive with fewer resources. Those who remain don’t need to turn in their spurs — prepare them to transition from data wranglers to agent rustlers.

Vendor fragmentation will force majority of enterprises to compose “agentlakes.”
Hyperscalers and data platform and automation vendors can’t claim agentic AI dominance just yet. Vendor fragmentation will cause a majority of enterprises to build composable agent architectures. 

These agentlakes will manage and orchestrate fractured AI agent deployments and enable complex multiagent use cases. While support for Model Context Protocol (MCP) is expanding rapidly as a universal standard for agent-tool communication and has the potential to also meet the needs of agent-to-agent communication, there are many other protocols at play such as Agent2Agent (A2A) and NANDA. 

Rapid advances in agentic coding tools that integrate fully into infrastructure and deployment tools are proliferating at an extreme rate. The underlying data supporting context engineering for agents will drive the need to compose real-time support for multimodal, multisource data, capabilities not easily found in one platform. 

Be ready to enable automation streams to flow into the agentlakes through composable and interoperable agents.

Sixty percent of Fortune 100 companies will appoint a head of AI governance.
While “governance” may be the frumpiest word in the lexicon of business, it’s essential to the development of safe, scaled AI. In 2026, the complexities of navigating the patchwork of legislation in the US and regulations such as the EU AI Act abroad will cause 60% of the Fortune 100 to hire or appoint a designated head of AI governance. 

Sony, Bank of America, and UBS have already done so. Look to tap someone internally who is deeply familiar with both internal policies and the regulatory environment, such as your chief risk officer or chief information security officer. 

Or, attempt to attract someone from an AI vendor that has long established dedicated ethics, safety, and governance teams. However, once you find this individual, they will not succeed alone. Quarterback a cross-functional team that embeds governance into the entire AI lifecycle — not just to comply but to ensure alignment with company values and objectives.

To lift AI adoption and reduce risk, 30% of large enterprises will mandate AI training. 
Poor AI literacy continues to erode trust in AI and leads to poor adoption. Our data shows 21% of AI decision-makers cite employee experience and readiness as a barrier to AI adoption. Companies like Johnson & Johnson are responding by rolling out mandatory GenAI training for 25% of their workforce. 

Further, 27% of AI decision-makers indicated their organization plans to or will require responsible AI training for technical roles within the next 12 months. Both AI adoption and risk management are intrinsic to AI maturity. 

That’s why enterprise-wide AI literacy not only improves the organization’s AIQ, but also protects the company from liability, particularly in regulated industries. Mandating AI literacy training will simultaneously improve adoption and mitigate risk. Consider partnering with an AI service provider or your existing technology providers to offer formalized training and set clear measures of success to gauge AI knowledge and use.

Summary
Every bubble inevitably bursts, and in 2026, enterprise ROI concerns will exceed the tensile strength of vendor hyperbole. In the face of this market correction, enterprises will prioritize function over flair. 

CFOs will get pulled into more AI deals, companies will distribute their bets across agentic ecosystems, and savvy enterprises will invest in governance and AI fluency training to squeeze real value out of AI.

-- Source: Forrester Research, USA & Australia.