BofA software survey: AI’s share of budgets rises to 9.4%; returns and costs determine the pace of expansion
Software budgets are expected to grow 5.3% in 2027, with AI receiving both new funding and money reallocated from existing budgets. But only 17.2% of respondents have achieved measurable enterprise-wide returns; how much revenue cloud, security and application vendors capture depends on deployment costs and actual benefits.
Original source：BofA Global ResearchFirst published on Unhidea：Original source date：
Key takeaway
BofA believes AI is expanding and redistributing software budgets, but the key for vendors turning deployment interest into revenue is delivering measurable business returns at an acceptable cost.
Is AI funding coming from new budgets or taking money from existing projects?
BofA’s survey of 309 technology budget decision-makers shows that, using a weighted calculation based on the budget ranges selected by respondents, software budgets will grow 5.3% in 2027; AI’s share of software budgets will rise from 8.3% in 2026 to 9.4%. Overall budgets are expanding, and AI’s share is also rising.
Around 61% of respondents fund AI primarily with new budget allocations, while around 38% rely primarily on reallocations from other IT categories. AI is bringing incremental funding while also competing for money allocated to existing projects; ERP and finance, and human capital management, rank in the bottom two places among 2026 budget priorities.
Deployment intentions are high. Why have returns become the next hurdle?
84.5% of respondents agree they will deploy AI agents over the next 12 months, but only 17.2% have reported measurable enterprise-wide returns on investment; another 31.4% expect returns within 12 months, while 30.7% expect to wait one to two years. There is a time lag between deployment plans and actual benefits.
88% of respondents cite improving productivity and reducing costs as uses for AI. BofA therefore believes continued budget increases will increasingly depend on whether those benefits can cover the costs; 46.6% have concerns about usage-based pricing, and uncertainty over costs as usage expands will constrain investment.
Applications attract the most attention. Can traditional software vendors charge more?
Around 40% of respondents chose AI applications as the area with the largest increase in AI spending in 2027, above the 20.1% for data platforms. Back-office software is considered for AI investment by 67.6% of respondents, but human capital management remains a low overall budget priority: customers wanting to use AI to improve efficiency does not mean they are willing to increase budgets for existing software.
BofA believes IT service and operations management, the field in which ServiceNow operates, is better positioned to generate revenue through process automation and cost reduction. Traditional vendors have data, processes and business context, but AI labs can access that data through APIs and take over the user interface; retaining the full customer experience and offering competitive prices and quantifiable results are more conducive to capturing value.
How will cloud and security capture AI-related spending?
According to BofA’s range-weighted calculations, public cloud budgets will grow 8.4% and cybersecurity budgets 6.9% in 2027, both faster than software overall. 91.3% of respondents say their AI computing usage is higher than a year ago; BofA believes this benefits Snowflake’s data processing and AI workload revenue and will also increase demand for Datadog’s system monitoring. Spending expansion still depends on whether projects can be put into production at an acceptable cost.
Security both constrains deployment and receives new funding: 53.7% of respondents have slowed AI deployment because of security concerns, while 60.8% identify AI security as a priority for incremental security budgets. BofA favors CrowdStrike and Palo Alto Networks to meet demand for AI and identity protection.
Microsoft’s advantage lies in its ability to charge at multiple stages: intent to use its AI cloud infrastructure stands at 88.4%, while 76.7% currently use or plan to put Copilot into production. BofA believes Azure captures underlying computing consumption, while Copilot supports enterprise deployment through governance, security and workflows.
