IBM and OpenAI announced a broad enterprise partnership aimed at converting powerful AI models into actual business value through automated workflows, application modernization, and cybersecurity improvements across core operations. The collaboration combines OpenAI's frontier models, including GPT-5.6, Codex, and ChatGPT Work, with IBM Consulting's technology and industry expertise to help organizations transform how they work.

Why Can't AI Just Automate Everything on Its Own?

The partnership reflects a critical shift in how enterprises view AI adoption. While business leaders are convinced that AI models are powerful, the real challenge lies in converting that intelligence into functional workflows that actually run the business. According to Michael Healy, Managing Partner of Offerings, Assets and Gen AI at IBM Consulting, "Enterprises no longer need convincing that the models are powerful. The challenge now is turning that intelligence into agentic workflows that actually run the business".

This insight aligns with a broader workforce reality: federal agencies and enterprises alike are discovering that technology adoption fails not because the software is inadequate, but because organizations underestimate the human side of change. Research indicates that roughly 60% of technology adoption challenges are people-centered, with user proficiency accounting for nearly 40% of the struggles organizations report.

What Does This Partnership Actually Do?

The IBM-OpenAI partnership will focus on several key business areas. Initial sectors of focus include financial services, government, telecommunications, and retail, with specific applications in finance, procurement, customer operations, and human resources. The companies will create a dedicated OpenAI Practice, with thousands of IBM consultants and engineers obtaining expert-level certifications through the OpenAI Partner Network.

IBM will also launch specialized forward-deployed units of engineers and consultants trained to work directly with clients on complex workflows and highly regulated environments. The partnership begins by helping companies understand where AI can make the biggest difference, since many enterprises struggle with technical debt and unclear processes. As Healy explained, "Before you can automate a workflow, you need to understand how it actually works. In many enterprises, that isn't always clear due to technical debt and human hand-offs".

As Healy

How to Successfully Deploy AI Across Your Organization

• Start with a baseline assessment: Before rolling out enterprise-wide AI solutions, conduct a lightweight diagnostic to understand what your organization can realistically absorb. This prevents expensive rework and ensures smarter sequencing of priorities.
• Match AI capabilities to specific mission needs: Technology adoption succeeds when a tool solves a problem that end-users actively recognize as their own. An automation capability that transforms one workflow may be irrelevant when applied to different work processes.
• Build reinforcement infrastructure into operations: Behavior change requires 60 to 90 days of active reinforcement to become habitual. Designate internal champions with credibility and give supervisors visibility to catch process regression early, ensuring sustained leadership commitment.
• Position technology as a retention tool: Frame AI solutions as force multipliers that eliminate burnout-inducing administrative tasks, allowing high-performing employees to refocus on mission-critical work that drew them to their roles.

Why Workforce Capacity Matters More Than Model Power

Federal agencies face a particular challenge: half of federal leaders surveyed state they are not confident that their agency's current staffing levels can meet mission objectives. When capacity shrinks, the natural institutional reaction is to look to technical solutions to automate the gap. However, decades of research on technology adoption reveal a consistent truth: most implementations fail not at the technical layer, but because the human side of change does not receive the same behavioral rigor, empathy, and structural alignment as the technology itself.

Retaining top talent has emerged as the clear top priority for federal leaders at 27%, followed closely by implementing new technology and automation at 21%. This suggests that agencies view technology not as a replacement for people, but as a tool to preserve institutional knowledge and prevent burnout-driven departures.

The challenge is compounded by competing priorities. Federal workforces routinely navigate, adopt, lead, or absorb five to 15 change initiatives at once, and that saturation compounds every adoption challenge. When employees appear to resist a new technology, they are usually just rationally managing limited bandwidth. More communication campaigns or mandatory training events cannot fix a workforce that is simply full to capacity.

What Makes Government AI Adoption Different?

Federal IT modernization operates under unique constraints that commercial change models were not designed to navigate. Rigid security authorization timelines and governance structures routinely decouple technology readiness from workforce readiness. The workforce is often briefed on a tool months before it is authorized, or they gain sudden access to a platform they were never meaningfully prepared to use.

Additionally, appropriations cycles mean long-term sustainment budgets look drastically different from initial implementation funding. This structural disconnect often starves the reinforcement process of resources the moment a system goes live. The program managers who succeed today are not those with the largest software budgets, but those who start the workforce transformation concurrently with technical design.

How Will IBM and OpenAI Modernize Applications?

Application modernization and product development form another major part of the partnership. Codex and ChatGPT Work will be combined with IBM Consulting Asset technology and expertise to help clients modernize applications and accelerate software development. Codex, functioning as a coding assistant, will accelerate the application development modernization cycle in combination with IBM Consulting Advantage's coding harness and OpenAI's models.

The partnership also expands IBM and OpenAI's work on cybersecurity. After IBM joined the OpenAI Daybreak Cyber Partner Program, the companies plan to combine OpenAI's frontier AI capabilities with IBM Autonomous Security, a service that uses multiple AI agents to coordinate security decisions, response, and intelligence.

For example, a manufacturer might manually compare its information with a competitor's products to identify possible substitutes. AI can more quickly align those product specifications and connect the results with supply chain and manufacturing processes, helping companies pull forward new product innovations and manage inventories more efficiently.

As models continue to improve at a faster rate, the capabilities available to enterprises will expand. Healy noted that "models are at a turning point where they are becoming much more intelligent at a faster rate, and we will see additional advances from those models that we don't have today".

Healy