IBM Is Certifying Thousands of Consultants in OpenAI

It’s Not Just IBM Consultants Getting OpenAI Certified

On August 14, IBM announced a strategic partnership with OpenAI. The plan is to integrate GPT-5.6, Codex, and ChatGPT Work into the IBM Consulting Advantage platform, and to push into finance, government, telecom, and retail—along with corporate finance, procurement, customer operations, and HR functions.

What caught my eye in the press release wasn’t the technology. It was the line saying IBM would build a dedicated OpenAI organization and have thousands of consultants and engineers earn expert certification through the OpenAI Partner Network. Through this collaboration, IBM also rose to OpenAI’s “elite partner” tier.¹

Reader, here’s why these two sentences matter. In the enterprise market, OpenAI isn’t just selling models—it’s expanding the pool of people certified to work with its technology. The IBM deal is a case of that strategy being executed at the scale of a major consulting firm.

---

IBM Built a Dedicated Unit Staffed with OpenAI-Certified Consultants

IBM made three promises in this announcement: turning legacy operations into AI-executable workflows, accelerating application modernization and development, and layering in security and AI risk management. So far, this is what every major SI (systems integrator) always says.

What’s new is the shape of the organization built to deliver on it. IBM created a dedicated unit called the “OpenAI Practice” and defined that unit’s identity as a group of certification holders. It’s not the first time a consulting firm has had its staff collectively earn certifications for a specific vendor’s technology. But this time, that fact isn’t a side effect — it’s the headline of the partnership.

Andy Baldwin, SVP of IBM Consulting, put it this way:

“The challenge isn’t accessing AI technology. The challenge is integrating AI safely — and at scale — into complex enterprise environments.”

I think this single sentence captures the state of the enterprise AI market in 2026 more precisely than anything else. What’s blocking adoption isn’t model performance — it’s figuring out internal approval processes and who’s accountable for what.

The free training for customer-side staff follows the same structure

A few days ago I had a chance to read through the facilitator script for a training session OpenAI runs called “Activator Labs 101.” Unlike the partner-side program with firms like IBM, this one is free training aimed at staff working inside customer organizations.

The session divides AI champions into three tiers: executive sponsors who set direction, transformation leaders who design company-wide rollout and governance, and agent activators who redesign repetitive work for a specific team. The training targets this third tier.

One sentence defining the activator stood out to me. What separates them from a power user isn’t skill — it’s scope of responsibility. A power user improves their own work, but an activator builds workflows “that others can use and the organization can sustain.” That means this person must consult with the workflow owner, security, legal, and IT, and does not hold final approval authority themselves.

The hands-on example is deliberately simple. Take a single task — intake and routing of internal requests. The AI is only allowed to classify the request and suggest an owner. If information is missing, it must stop rather than guess, and anything customer-facing or hard to reverse gets routed to a human. Assignment only executes after the intake lead gives explicit approval.

Attending the session earns you a badge. Submit a real workflow with supporting evidence and you get a second badge — “Agent Activator.” The champion community already has roughly 8,000 members.²

On the partner side, thousands of IBM consultants earn professional certifications. On the customer side, staff earn badges. The two programs share the same architecture. Alongside selling the model, OpenAI is cultivating, on both sides at once, the people who will be responsible for adopting that model inside their organizations.

Personal use by employees is 90%, official company subscriptions are 40%

Why go this far? The MIT NANDA research team’s findings offer an answer.

The MIT NANDA team’s “State of AI in Business 2025” became famous for the line “95% of generative AI pilots failed to generate meaningful returns.” The report synthesizes over 300 public case studies, 52 structured interviews, and a survey of 153 senior leaders. It’s not a peer-reviewed academic paper, and the sample isn’t large, so it’s hard to generalize this figure at face value.

What I find far more important in this report isn’t the 95% — it’s a different pair of numbers.

Companies where employees use personal AI tools for work: 90%. Companies that purchased an official company-wide subscription: 40%.

This 50-percentage-point gap is what’s known as shadow AI³. And this is OpenAI’s real problem — and its real opportunity.

OpenAI already has a service with 900 million weekly users. Its business customers have surpassed 1 million, and ChatGPT for Work seats⁴ have exceeded 7 million. In April 2026, the company revealed that enterprise revenue had surpassed 40% of the total and would match consumer revenue within the year.

So the bottleneck isn’t brand awareness or model performance. Individual employees are already using it — what’s stuck is the step of converting that individual use into an official company contract. To get past this step, someone inside the organization has to say:

“I’m going to change this workflow this way. Here’s how we got security review, here’s where the approval gate sits, and if something goes wrong, I own it.”

Because no one is there to say that sentence, an organization’s AI adoption stalls. IBM’s certified consultants and the customer’s activators are, in the end, the people who say that sentence on someone else’s behalf.

Salesforce pioneered this playbook 20 years ago

If this structure feels familiar, you’re right. Salesforce has been running this exact playbook with Trailhead⁵ for nearly 20 years.

Salesforce led with certification rather than direct product sales. It taught for free, handed out badges, and made those badges valuable in the job market. The result: even when a company had no reason to adopt Salesforce, individuals now had a reason to learn it and push for its adoption — because it built their own career capital.

According to IDC estimates, this ecosystem is projected to generate a net 11.6 million jobs and $2.02 trillion in revenue between 2022 and 2028. That’s not Salesforce’s own revenue — it’s the number generated around it. Of course, this was a vendor-commissioned estimate, so it should be read with an awareness of how sensitive it is to its underlying assumptions.

OpenAI is running the same playbook, only much faster. In December 2025, it launched its first certification program with a goal of certifying 10 million Americans by 2030, lining up Walmart, John Deere, Lowe’s, BCG, and Accenture as initial pilot partners. Coursera and Credly handle issuance. Where Salesforce built up from the community, OpenAI started by locking in the large enterprises that actually hire the workforce as its early partners. This IBM deal simply adds a major systems integrator as a distribution channel on top of that.

The workflow blueprint can migrate to another model, but the person who learned the methodology stays

One objection is possible here: “Still, isn’t the lock-in weaker than what Salesforce had?”

That’s a fair point. Salesforce’s structure was one where switching costs grew as customizations piled up. OpenAI, by contrast, is teaching people to produce workflow blueprints — documents that spell out what information to use, what’s off-limits, and where a human needs to sign off. In principle, these are portable to any other model. Requirement specifications travel well.

That’s exactly why certification matters.

When technology alone can’t hold onto a customer, what’s left is people — the person inside the company recognized as the owner of the AI workflow, the one who’s internalized the methodology through practice, the one who’s negotiated with security and legal teams using that exact vocabulary. Whatever tool that person proposes to expand next quarter is likely to be a tool designed around the very methodology they were trained in.

The moment thousands of IBM consultants get OpenAI-certified, the default reference architecture they bring to client companies is effectively decided too. And it’s no accident that the reference implementation is ChatGPT Work. This feature, launched in July 2026, connects Slack, Teams, Gmail, Drive, Salesforce, and SharePoint to autonomously execute multi-step tasks — while requiring human approval before sensitive actions. The principle drilled into every Activator training session — “AI prepares and recommends, but humans approve” — is implemented directly as a product feature.

This isn’t a case of training coming first and the product following. OpenAI is distributing its product manual in the form of methodology and certification.

Oswarld’s Lens

I used to build and run Notion’s Korean community. I learned one thing from that job: a community starts working as a distribution strategy not when people like the product, but when the ability to handle that product becomes someone’s social status. Once a person who’s good at building Notion templates starts getting called “the person who’s good at that” inside their company, the distribution engine shifts — it’s no longer the company pushing it, it’s that person.

I saw the same pattern over and over while designing GTM strategy at Gamma. Things moved faster not when we repeated product demos, but when a specific point person emerged inside the client organization to drive adoption.

So I think OpenAI’s strategy here is well designed. But there are two things companies adopting it need to watch for.

First, the Activator is a role where responsibility easily outgrows authority. The training script itself says as much: Activators coordinate and recommend, while actual change authority sits with the workflow owner and governance lead. On paper, that’s balanced. But in practice, if the company doesn’t properly design the approval gates and exception paths, the person left holding the bag when something goes wrong is the one practitioner who built the workflow. I see this pattern constantly in consulting — organizations tend to change the work while leaving the accountability structure untouched.

Second, a certified advisor is not a neutral advisor. IBM has long positioned itself as “model-neutral.” An elite partner tier and a dedicated practice is a decision that nudges that position a little. That’s not to say it’s a bad thing. But when you receive an architecture proposal from that kind of organization, you have to separate, at least once, whether it’s a technical judgment or a byproduct of the partnership structure.

If you’re at a Korean company, add one more question to that list. Korea’s major SI firms will soon bring similar dedicated units and certification programs of their own. When they do, the question to ask isn’t “how many people got certified” — it’s “who decides which points in our workflow require human approval.”

Closing

Here it is in three lines.

• What IBM struck with OpenAI is a technology agreement, but it’s also a personnel certification agreement. Thousands of OpenAI-certified consultants become the channel through which this technology reaches client companies.
• OpenAI runs the same structure on both sides at once: expert certification for partner-firm consultants, and Activator badges for client-side practitioners. Its read is that what blocks adoption isn’t model performance — it’s the absence, inside the company, of someone to own that workflow.
• A workflow design document can be carried over as-is to a different model. So what actually keeps a customer locked in isn’t the technology — it’s the person who has learned that methodology and drives adoption inside the company.

If you happen to be leading internal AI adoption, or you’re on the receiving end of an SI proposal, I’d suggest checking one single question at your next meeting: “Does the document have a name attached — someone who’s accountable if this workflow goes wrong?” If it doesn’t, you’re not at the adoption stage yet. You’re still at the experiment stage.

💬 Have you ever been the one pushing internal AI adoption forward, or been on the receiving end of a proposal that leaned on vendor certification? Tell me in the comments where the gap between authority and responsibility broke down hardest. I’ll pick this up again in the next issue.

---

💬 Share your thoughts or experiences on this topic in the comments · 📨 Know a colleague wrestling with internal AI adoption? Send this their way

---

The draft looks accurate and complete—no Hangul remains, all numbers match, and the structure (headings, footnotes, links) mirrors the source exactly. No corrections needed.

References & Further Reading

Primary sources

• IBM Newsroom, “IBM partners with OpenAI to accelerate secure AI deployment for enterprises across core operations”, 2026. 8. 13. Link ··· This is where today’s piece starts. Read the paragraphs on the “OpenAI practice” and certification before the technical integration section.
• OpenAI Academy, “Champions” community. Link ··· This is where the activator program actually runs. You can check the scale yourself.
• OpenAI, “Launching our first OpenAI Certifications courses”, 2025. 12. 9. Link ··· Contains the 2030 target of 10 million certifications and the pilot partner list.
• OpenAI, “The next phase of enterprise AI”, 2026. 4. 8. Link ··· Covers the enterprise revenue share crossing 40% and the consulting-partner strategy. Good background reading for the IBM story.
• MIT NANDA, The GenAI Divide: State of AI in Business 2025, 2025. Link ··· Look past the 95% figure to the 90% vs. 40% gap. This is the core evidence behind today’s piece.

Background

• OpenAI, “1 million business customers: the fastest-growing business platform in history”, 2025. 11. 5. Link ··· The original source for the numbers: 1 million business customers, 7 million seats.
• Salesforce·IDC, “Salesforce Economy will create 11.6M jobs and $2.02T in revenues between 2022 and 2028”, 2023. 9. 11. Link ··· A precedent showing how large a certification-based ecosystem can grow. Keep in mind this is a vendor-commissioned estimate.
• SiliconANGLE, “OpenAI debuts ChatGPT Work, an agentic tool for automating business workflows”, 2026. 7. 9. Link ··· Worth reading alongside the education methodology to see how it dovetails with product features.
• Sify, “95% Companies Failing with AI? An MIT NANDA Report Misread by All”, 2025. Link ··· A rebuttal on how the widely-cited 95% figure should actually be read. Worth reading alongside the original report.

---

The author is Oswarld (Kwangseob Ahn). Current roles: Adjunct Professor at Sejong University, Strategy Consultant at INLEVEL9. Career, research, books, and recent work are kept current on the About page. Latest · July 2026: HEMA-2: A Consolidation-Aware Tri-Memory Architecture with Multi-Channel Scheduling for Lifelong Conversational AI.

📝 Glossary

Footnotes

1. Elite Partner tier: The top-tier partner classification within OpenAI’s partner network. It’s awarded based on the scale of certified staff, joint go-to-market capacity, and enterprise deployment capability. ↩

2. Champions community: A learning and networking space OpenAI runs for people driving AI adoption inside enterprise customer organizations. As of August 2026, roughly 8,000 members had joined. ↩

3. Shadow AI: The situation where employees use AI tools for work through personal accounts without formal company approval or contracts. Productivity rises, but risks remain around data leakage and audit trails. ↩

4. Seat: In enterprise software, one seat means one account, i.e., one user’s license. 7 million seats means the company is paying for 7 million accounts. ↩

5. Trailhead: Salesforce’s free online learning platform. Completing courses earns badges, and Salesforce built a structure where those badges function as real credentials in the hiring market. ↩