TL;DR — Key Takeaways
- Gartner predicts that up to 70% of enterprises will abandon agentic AI systems built through vendor-assisted models by 2028 as maintenance costs rise and internal teams struggle to manage them independently.
- Forward-deployed engineering can accelerate deployment, but Gartner warns that poorly structured engagements may create vendor lock-in and long-term dependence on outside engineers.
- Fewer than 20% of FDE engagements are expected to turn custom client requirements into standard vendor product features through 2028.
Corporate enthusiasm for autonomous artificial intelligence (AI) is hitting a costly structural wall.
By 2028, up to 70% of enterprises will abandon agentic AI systems built through vendor-assisted models as maintenance expenses soar and internal teams struggle to modify the technology independently, according to a report by market research firm Gartner.
The chief catalyst for the impending shift is forward-deployed engineering (FDE), a high-touch model where tech vendors embed their own engineers directly within customer organizations to build customized software. While FDE offers rapid initial deployment, Gartner analysts warn that it risks creating long-term dependence on expensive third-party firms, leaving enterprises unable to manage or update their own systems once vendor engineers depart.
“Success starts with getting the engagement structure right, from scope and incentives to governance, ownership, and exit,” said Mukul Saha, senior director analyst at Gartner. “The best-scoped FDE engagements have clear guidelines on governance, business value delivery, IP ownership, project co-ownership, knowledge transfer, and an exit strategy from day one.”
Saha noted that the hype around AI has led to widespread “FDE washing,” where traditional professional services, implementation, and IT consulting are rebranded with strategic labels to command premium fees without offering equivalent operational depth.
Through 2028, Gartner projects that fewer than 20% of FDE engagements will successfully convert custom client requirements into standard features in a vendor’s core product, signaling that many tailored solutions may become costly dead ends.
To avoid vendor lock-in, Gartner advises enterprise engineering leaders to restrict FDE models to highly complex projects requiring deep product specialization or tight architectural integration. For routine implementations, standard partner or consulting frameworks often deliver more predictable, cost-effective outcomes without sacrificing internal control.
The warning adds to a growing chorus of skepticism regarding the true return on investment for enterprise AI projects.
Earlier this year, Gartner forecast that over half of generative AI initiatives would exceed their budgets due to poor architectural decisions and a lack of operational expertise. Furthermore, standard AI agent deployments are expected to face a 40% reduction or complete decommissioning as organizations grapple with governance and safety hurdles.
Beyond software architecture, Gartner also cautioned that knee-jerk corporate restructuring tied to AI automation may backfire. The firm recently predicted that nearly one-third of employees laid off due to AI integration will eventually need to be rehired at significantly higher costs to fill critical knowledge gaps.
