Agentic AI projects abandoned by 54% of firms, SnapLogic report finds
New SnapLogic research covered by TechRadar says unclear business value, not weak AI models, is why more than half of enterprises have dropped agentic AI projects. Poor data foundations and spending that outpaces oversight are the main causes.

Key takeaways
- SnapLogic data reported by TechRadar puts the abandonment rate for agentic AI projects at 54%, blamed on unclear business value.
- 93% of enterprises said they have AI agent policies and governance frameworks in place.
- 64% reported moderate or significant AI overspending, despite 88% saying their budgets were clearly defined.
- SnapLogic says weak data foundations, not limited AI capability, explain the disappointing returns.
In this article
What the SnapLogic report found
More than half of large companies have dropped agentic AI projects because they could not show a clear business case. TechRadar reported the figures on 18 November 2026, drawing on new research from the data integration firm SnapLogic. The study puts the abandonment rate at 54%.
Agentic AI describes software that can plan and carry out multi-step work with limited human supervision, using company data and calling other applications. SnapLogic's argument is that the models were not the weak link. It traces poor returns on investment back to the data those systems were handed.
Policies exist, but data foundations lag

Earlier surveys often criticised businesses for having no clear AI direction. This report found the reverse. Some 93% of enterprises said they had already written their own policies and governance frameworks for AI agents. Governance here means the rules, checks and named owners that decide how a system may be used.
Access is broad as well. Three in four respondents (77%) said their agents already have, or would be given, full or extensive reach into company data and business systems so the tools can work properly. Nine in 10 said they trust, or would trust, what agentic AI produces. SnapLogic's point is that this confidence sits on foundations that are not yet solid.
Spending is running ahead of oversight
The pace of adoption stands out as a risk in the data. More than a third of respondents (36%) admitted that AI investment is moving faster than their ability to govern the systems they are buying. Two-thirds (64%) reported moderate or significant overspending in the category, even though 88% said their agentic budgets were clearly defined.
Pressure from the top appears to be part of the picture. One in two IT decision-makers said they feel pushed to keep investing in AI agents, often because of demands from senior management. A similar share, 48%, said they struggle, or expect to struggle, to see the results they were promised.
What the report recommends next
SnapLogic's chief technology officer, Jeremiah Stone, wrote that putting governance, infrastructure and money in place is not enough on its own. In his words, "those foundations alone do not guarantee business outcomes". He argues that companies must connect AI to the right data, applications and workflows, and get a clearer view of performance and cost, before spending becomes measurable value.
For IT buyers, the practical reading is to check data quality and access before scaling agent deployments, and to track cost per project next to results. SnapLogic sells data integration software, so it has a commercial interest in this conclusion. The TechRadar coverage did not set out the sample size or method behind the survey, so the numbers are best treated as the vendor's own research rather than independently verified findings.
Sources
Spotted a mistake? See our corrections policy. Read how News by OnDevice works in our editorial policy.



