Your real competitor is the unused licence you already pay for

Who are we actually competing with?

Not another AI trainer. Publicly listed enterprise learning seats sit in a band of roughly US$360–779 per user per year: Pluralsight lists business tiers from $399 to $779, and Coursera for Teams is listed at $399 (Pluralsight via G2, 2026; Coursera pricing via TrainingCost, 2026). Most enterprises we talk to already hold seats like these, plus Copilot, and most of those seats are dormant.

That dormant seat is our competitor. When a transformation owner weighs Chokmah against "doing nothing," doing nothing is not free. It is a licence already on the books, producing no measurable change. The question is not whether to spend more on capability. It is why the capability spend already made returned nothing.

MIT gives the shape of the answer: 95% of enterprise GenAI pilots produced no measurable P&L return, and the cause it identifies is organisational, not technical (MIT, July 2025). The licence was never the missing piece. The workflow was.

Pluralsight lists business plans at US$399, $579 and $779 per user per year; Coursera for Teams is listed at $399. >Pluralsight (G2) and Coursera (TrainingCost), publicly listed plans as of July 2026 (2026)

These are indicative list prices; enterprise deals are quote-gated and discounted. We cite the list, not a negotiated figure. The point is not that platforms are expensive. It is that the money is already committed and the workflow is unchanged.

What are catalogue platforms genuinely good at?

We are not here to disparage them, and this matters: the argument is about scope, not quality. Coursera, Pluralsight and their peers are excellent at delivering standardised concepts to large populations asynchronously and cheaply. If you need 500 people to understand what a large language model is, what a token is, or how retrieval works in principle, a catalogue seat is the right tool and we will tell you to keep it.

They are also good at credentialing at scale, at giving people a structured on-ramp, and at covering foundational literacy that would be wasteful to teach in person. None of that is in dispute.

What they cannot do is the next paragraph.

The three things a catalogue structurally cannot do

It cannot name your workflow. A course is written once for everyone. Your invoice-exception process, your EDI partner mappings, your RFQ drafting path: none of these exist in any catalogue, because they exist only inside your organisation. And workflow redesign is exactly where MIT located the 95% failure. A generic course cannot redesign a process it has never seen.

It cannot leave a working thing behind. A completion certificate is not a shipped agentic system. The gap between "our people took the course" and "the exception queue is now handled" is the entire distance a pilot has to travel, and the catalogue covers none of it.

It cannot measure absorption. Platforms report logins and completions because that is what they can see. They cannot see whether the work changed. That is the difference between adoption and workflow absorption, and it is the only difference that shows up in the P&L.

When should you keep the platform, and when cancel it?

Keep it where utilisation is real and the goal is genuinely foundational literacy across a large group. That is the job it does well, at a price per head that in-person delivery cannot match. Cancelling a well-used literacy platform to fund a workflow project is a false economy.

Cancel it (or at least stop renewing seats) where the licence was bought as a substitute for changing the work and utilisation has decayed to near zero. A dormant seat is not capability in reserve; it is a recurring cost attached to a decision that did not work. Reallocating it is often the cleanest first source of funding for a workflow project, and it comes out of a budget you have already justified.

The honest audit is unglamorous: pull utilisation, separate the seats doing real literacy work from the seats bought to avoid a harder conversation, and act on the second group. That is the first hour of a licence-utilisation audit, and an illustrative example of the downstream build is our engagement scenario for vendor invoice reconciliation.

What does this mean for a transformation owner?

Stop framing the decision as more training versus less. The catalogue seat and a workflow engagement are not substitutes competing for the same budget line: one is L&D and foundational, the other is transformation and specific. Fund the literacy from L&D, keep it if it is used, and fund the workflow from transformation, where the return is a measurable change in how the work runs.

Then hold the workflow project to the standard the catalogue can never meet: a named process, a measured baseline, a shipped system your team owns, and a number that moved. If the engagement cannot promise those, it is just another seat you will not use.

Disclosure: Chokmah is a new practice with no completed client engagements. This argument rests on published list prices, on MIT's research, and on method, not on results we are claiming. Hold us to the same standard: what workflow, what baseline, what number.

Audit the licences before you buy another

A two-week diagnostic includes a licence-utilisation audit, which seats are doing real work, which are funding avoidance, and which workflow to change with the money you free up.

Book an Adoption Diagnostic · Why training is not the product

Frequently asked questions

Are enterprise learning platforms worth it?

For foundational literacy across large populations, yes: a listed seat at roughly US$399–779 per user per year delivers standardised concepts more cheaply than in-person training ever could. What it cannot do is change a specific workflow that exists only inside your organisation. Keep it for the first job, not the second.

What is a typical corporate course completion rate?

Completion rates vary widely and are not a metric we would build a case on, because completion measures activity, not changed work. A team can complete every course and run the workflow exactly as before. The number that matters is whether the work now costs less in cycle time or errors.

Should we cancel our Coursera or Pluralsight subscription?

Only where utilisation has decayed and the seat was bought as a substitute for changing the work. Keep seats that are genuinely used for foundational literacy. The right move is a utilisation audit that separates the two, not a blanket cancellation or a blanket renewal.

What is the difference between training and enablement?

Training delivers concepts to people; enablement changes how the work runs. Training's output is a more knowledgeable person and a completion record. Enablement's output is a workflow that costs less than it did, plus the people who can maintain it. They belong to different budgets and answer different questions.

Does certification predict capability?

Weakly, if at all, for the work that matters here. A certificate shows someone completed structured material; it does not show they can redesign your exception queue or stand up an evaluation harness. We measure capability by what a team can build and maintain, not by what they have completed.