Dive Brief:
- Leaders relying heavily on AI to carry out mainframe exit strategies should prepare for missed expectations, according to Gartner. More than 70% of projects started this year will fail due to overestimation of generative AI capabilities.
- Businesses risk cost overruns, technical debt and critical failures by focusing on AI, according to Gartner. The gap between the marketed promise of generative AI and its actual ability to migrate legacy code is widening, Alessandro Galimberti, VP analyst at Gartner, said in a press release.
- Vendors are embedding AI into offerings regardless of the technology’s ability to improve outcomes, Galimberti said. “When this is combined with the ‘too-big-to-fail’ nature of mission-critical mainframe applications and the accelerating loss of experienced talent, infrastructure and operations leaders face a perfect storm of risk that makes poorly planned exit strategies increasingly untenable,” he added.
Dive Insight:
AI is infiltrating nearly every aspect of business operations, including mainframe overhauls as enterprises and governments alike look to modernize their tech stacks.
While analysts caution that the technology alone isn’t a magic fix for notoriously complex migrations, companies that rely heavily on mainframes — especially in the financial services sector — have leaned into AI’s potential.
Chase used generative AI to speed its mainframe modernization, implementing governance and safeguards to reduce risks associated with modernizing critical applications. Meanwhile, Morgan Stanley built an in-house AI tool called DevGen.AI that has updated millions of lines of legacy code across the banking giant’s mainframe, according to a spokesperson.
Vendors such as AWS and even large language model provider Anthropic tout the benefits of using AI to modernize legacy applications. However, mainframe giants including IBM and Kyndryl have reinforced the role of mainframes as a modern platform for enterprises. The North Carolina Division of Motor Vehicles selected Kyndryl, an IT services provider, for its mainframe modernization project earlier this year.
Enterprise mainframe strategies depend on the complexity of their environments, Galimberti said in the press release. For many customers, generative AI can be used to modernize their mainframes rather than speed migration off the mainframe, he said.
Gartner estimated that 75% of mainframe exit vendors will adjust or end their business models by 2030 as demand for one-size-fits-all offerings decreases.
“Organizations must balance strategies focused on optimizing existing mainframe investments, while limiting full platform exits to select, case-by-case scenarios, as these efforts require high-risk transformation and often result in suboptimal outcomes,” Galimberti said.
