Gartner predicts 30% of GenAI projects to be abandoned before 2026

Despite the widespread hype around generative AI, Gartner predicts that 30% of projects involving the technology will be abandoned after the proof-of-concept stage by the end of 2025.

Organisations are expected to cancel a significant percentage of AI projects due to difficulties associated with poor data quality, inadequate risk controls and escalating costs of unclear business value, the research firm said.

A major challenge for organisations arises in justifying the substantial investments needed in generative AI for productivity enhancement, which can be difficult to directly translate into core business benefits. Deployment costs associated with AI can be significant, ranging from US$5 million to US$20 million, the company said.

For example, Gartner estimates that embedding generative AI APIs in custom apps can have upfront costs ranging from US$750,000 to US$1 million, and recurring costs of up to US$1200 per user per year. Extending generative AI models via data retrieval can have similar upfront costs and per user costs ranging from US$1300 to US$11,000.

Customising generative AI models for applications such as custom virtual assistants can have upfront costs of US$5 million to US$6.5 million, and per user costs of US$8000 to US$11,000 annually, and building custom AI models from scratch can cost US$8 million to US$20 million upfront, and US$11,000 to US$21,000 per user per year.

But the experience of early adopters in the space demonstrates that the payoff can be significant, Gartner said. A recent survey conducted by the company found that respondents reported on average a 15.8% revenue increase, 15.2% cost savings and 22.6% productivity improvement.

As a result, Gartner is recommending organisations analyse the business value and total costs of GenAI business model innovation to determine both the direct return on investment and any future value impact.

The cost of deploying GenAI. For a larger image click here.

“Unfortunately, there is no one size fits all with GenAI, and costs aren’t as predictable as other technologies,” Gartner distinguished VP analyst Rita Sallam said at the Gartner Data & Analytics Summit in Sydney. “What you spend, the use cases you invest in and the deployment approaches you take all determine the costs. Whether you’re a market disruptor and want to infuse AI everywhere, or you have a more conservative focus on productivity gains or extending existing processes, each has different levels of cost, risk, variability and strategic impact.”

Top image credit: iStock.com/WANAN YOSSINGKUM