BofA: AI Agents to Unlock $155 Billion Software Spending

Bank of America predicts AI agents will unlock a staggering $155 billion in software spending by 2030, dwarfing current forecasts. These autonomous digital colleagues are expected to drive significant workforce productivity improvements.

The financial world, often prone to both exuberant highs and cautious retrenchment, is currently grappling with the tangible returns on its massive artificial intelligence investments.

While the initial splash of generative AI tools, from sophisticated chatbots to image generators, certainly captured headlines, their direct, quantifiable impact on corporate bottom lines has, for many, remained elusive. CEOs report that only 25% of AI projects deliver expected ROI.

The market, perhaps suffering from a touch of AI fatigue after the initial hype cycle, has adopted a wait-and-see approach, leading to what Bank of America Securities (BofA) terms an “underwhelming generative AI adoption or monetization cycle” so far.

But beneath this perceived lull, a deeper, more transformative shift is brewing, one that BofA believes will unlock a staggering $155 billion in software spending by 2030.

This isn’t merely a bullish projection; it’s an estimate that dwarfs most current industry analyst forecasts by a factor of three.

The catalyst for this monumental surge? The quiet, yet profoundly powerful, emergence of agentic functionality.

This isn’t just about smarter chatbots; it’s about autonomous digital colleagues capable of performing specific tasks, making decisions, and even initiating actions with minimal human oversight.

BofA’s “deep dive” into this nascent field posits that these AI agents are the true key to unlocking sustainable, measurable, and material workforce productivity improvements – the holy grail of corporate investment.

The analyst’s methodology for arriving at this eye-popping $155 billion figure is as insightful as it is ambitious.

Leveraging a top-down approach, BofA focused on the “services” Total Addressable Market (TAM), essentially the vast pool of human salaries and wages.

Their calculations estimate global knowledge worker wages across seven major occupation categories, including sales, finance, and IT, at an astonishing $18.6 trillion annually.

The assumption that follows is that by 2030, agents will perform a mere 10% of the workflows for this massive cohort of employees.

Even this seemingly modest penetration translates into an automation-driven value of $1.9 trillion.

Here’s where the monetization model becomes clear: BofA projects that software vendors can capture roughly 8% of this immense value, representing a compelling 12 times customer return on investment for agentic spending.

This powerful value proposition – a clear, quantifiable ROI – is what sets agentic AI apart from the more nebulous benefits often associated with early generative AI deployments.

It’s a shift from novelty to utility, from exploration to measurable efficiency.

The current landscape, however, reflects the nascent stage of this revolution.

BofA’s surveys indicate a fascinating paradox: while a significant 64% of organizations anticipate pursuing agentic AI initiatives in 2025, actual deployment remains largely in the exploratory phase.

As of January 2025, 53% of organizations were still in exploration, 25% in pilot stages, and a mere 6% had reached full production.

This suggests a strong strategic intent, but also highlights the inherent challenges and complexities of integrating autonomous agents into existing workflows and enterprise structures.

Unsurprisingly, the initial battlegrounds for agentic technologies are expected to be in high-volume, repetitive, or data-intensive functions.

Customer service, marketing, sales, and software development are flagged as the first significant job functions poised for widespread adoption. Imagine AI agents handling routine customer queries, personalizing marketing campaigns at scale, qualifying sales leads, or even generating code snippets and debugging in real-time.

The efficiency gains could be transformative.

So, who stands to benefit most from this impending wave?

BofA has identified several key players within its coverage, companies already positioning themselves at the forefront of this shift.

Microsoft Corp, with its powerful Azure cloud platform and deep OpenAI partnership, alongside its ubiquitous Microsoft 365 product cycles, is an obvious contender.

Salesforce, with its “Agentforce” initiatives, aims to embed agentic capabilities directly into its vast CRM ecosystem.

ServiceNow’s “Now Assist” is designed to automate IT and customer service workflows.

Intuit’s “Intuit Assist” for QuickBooks and TurboTax promises to revolutionize small business finance and personal tax preparation.

HubSpot’s “Breeze AI agents,” GitLab’s “Duo,” and OneStream’s “SensibleAI” round out the list of firms deemed to possess the critical ingredients for success.

These aren’t random picks; BofA’s calculus hinges on a few critical ingredients.

Foremost among them is access to vast, proprietary customer datasets – the lifeblood of any truly intelligent AI.

Companies like Salesforce, with its sprawling CRM data, or Intuit, with its intimate knowledge of small business finances and individual tax data, possess an invaluable competitive moat.

Coupled with robust data management and governance capabilities, these firms are uniquely positioned to train, deploy, and scale agentic solutions responsibly and effectively.

Furthermore, these companies already have publicly available or in-beta agentic AI features or products, signaling their readiness to capitalize on the coming boom.

While the market remains wary of the immediate AI monetization cycle, BofA’s report serves as a powerful reminder that the most significant shifts often begin quietly, beneath the surface of prevailing skepticism.

The ramp-up in monetization is expected to begin in earnest in calendar year 2026, setting the stage for a period of profound technological and economic restructuring.

The future of work, driven by the quiet efficiency of AI agents, is not just coming; it’s already being built, poised to deliver a productivity revolution that could reshape global industries.