IBM Just Named the Bottleneck. Watch Where It Put the Answer.
IBM Consulting says the constraint isn't model capability, it's that most enterprises can't see how their own work runs. Then it put the answer to that inside its own delivery platform. That move is worth more attention than the partnership it came wrapped in.
- AI implementation
- IBM
- OpenAI
- systems integrators
- workflow discovery
- agency selection
Michael Healy runs offerings, assets and gen AI for IBM Consulting. Here's what he told IBM's own newsroom this month, describing the firm's new partnership with OpenAI: "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."
Read that again with an eye on who's saying it.
IBM Consulting has just embedded OpenAI's frontier models, Codex and ChatGPT Work into the platform it uses to deliver consulting work. It's standing up a dedicated OpenAI Practice, putting thousands of consultants and engineers through OpenAI Partner Network certification, and building forward-deployed engineering units for complex and regulated environments. Every commercial incentive in that announcement points toward saying the models are the thing. Healy said the opposite, twice. "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."
That's the execution argument, on the record, from a firm that would do very well if the argument were access.
The part the coverage skipped
Trade press filed this as a model vendor meeting a systems integrator, which it is. The move worth your attention is quieter and it sits a few paragraphs later in the same interview.
IBM's answer to the workflow-visibility problem is two products. Context Studio codifies industry and enterprise knowledge. Process Studio analyzes a company's standard operating procedures and workflows to find the inefficiencies. Both live inside IBM Consulting Advantage, the proprietary platform IBM uses to deliver its own consulting.
So the discovery step gets named as the binding constraint and then folded into the delivery stack of the firm doing the delivering. The thing you can't see about your own business becomes a feature of the product you're buying to fix it.
This is a structure, not a scandal
Worth saying plainly: there's nothing improper here, and firms have worked this way for as long as professional services has existed. An assessment whose recommended remedy happens to be a service the assessor sells is the default arrangement in this industry, not an aberration. It's often the right arrangement. The firm that diagnoses you carefully is frequently the firm best placed to fix what it found.
It's also not isolated. Three weeks before the IBM announcement, AlixPartners, a restructuring firm whose whole product is an independent read on what's broken, bought Artium, an agentic AI consultancy that builds enterprise agents for BNY Mellon, Mayo Clinic and eBay. It will run as a named sub-brand. Same shape, one tier down, and the trade press filed that one as capability expansion too.
What's changed isn't the ethics. It's what you're able to compare.
What you lose when the method becomes a product name
You can compare prices across three proposals. You can compare timelines, team composition, references, industry experience. What you can't compare is how each firm decides which of your workflows are worth rebuilding, because in every case that decision now lives inside something branded.
Ask three partners how they'd scope your engagement and you'll get three product names and three confident descriptions of what those products do. None of them will hand you the rule. And the rule is the entire game, because the difference between an AI programme that pays and one that quietly gets shelved at month nine is almost never the model. It's which twelve workflows went into scope and which forty didn't.
We've made a version of this argument about certification: a badge tells you someone understands the platform, not whether they diagnosed the work correctly before building on it. A platform name is that problem with the volume up. At least a certification is a public standard you can go read. Process Studio is a black box with a good pitch behind it, and so is every competing one.
The thing to ask before you sign
Ask a partner to walk you through the decision, not the tool. How do they choose which workflows enter scope? What has to be true about a workflow for them to recommend leaving it alone? What did they measure, on whose work, before they made that call last time?
You're listening for one thing: whether the answer is a method you could learn and apply yourself, or a product you'd have to keep buying. Both can be good. Only one of them leaves you able to check the work, and only one of them is still yours when the engagement ends.
Here's what makes this uncomfortable. Almost everyone credible now agrees on the diagnosis. IBM says the constraint is that enterprises can't see how their own work runs. Gartner is telling CFOs the same thing in different words. The analysts, the integrators and the model vendors have converged on it.
Nobody is in a hurry to hand you the glasses.
If you want that question answered for your specific situation, the Forge Playbook does it. Answer a few questions about your business and we'll put together a tailored outline of which workflows are worth automating and what a realistic budget looks like for each. Free, no obligation, takes about three minutes.
