I have spent a long time in this catalogue documenting a specific and dispiriting pattern: vendors give away the courses and charge two to five hundred dollars for the exam, because the courses are marketing and the credential is revenue. Okta charges $250 with $100 retakes, which is on the fair end. Appian charges $250 per attempt with no bundled retake. Tenable is around $450.
The Okta Certified Technical Architect exam is $5,000. Against that background, Boomi is doing something notably different. Every self-paced course is free. Every certification exam is free.
The only paid items are optional live instructor-led classes, which are clearly identified as optional rather than being a disguised prerequisite. There are over a hundred courses and more than twenty certifications. That is not a free tier attached to a paid programme, it is the whole programme, and it places Boomi at the top of the short list of genuine counter-examples I have been maintaining, alongside Mendix, Qualys, New Relic University and Cribl. The exam design is the part I find most interesting, and it is a deliberate philosophical choice rather than laziness.
The exams are open-book, open-platform and untimed. You can consult the documentation. You can have the platform open. You can take as long as you need.
The conventional objection is immediate, which is that this makes the exam easy and therefore meaningless. I think that objection is mostly wrong and worth arguing with, because it reveals what we actually believe certifications are for. In real work, nobody memorises connector configuration syntax. They look it up, because looking it up takes eleven seconds and misremembering costs an afternoon.
An exam that tests recall of things everyone looks up is testing a skill that does not exist in the job. An open-book, untimed exam can instead pose a problem and ask whether you can construct a working solution, which is the thing you are actually hired to do. Whether Boomi's exams fully exploit that possibility is a separate question, and the format at least permits better assessment than the closed-book alternative. The honest counterweight is about signalling rather than learning, and it matters for anyone treating a certification as a career investment.
Hiring managers price scarcity, and a free, open-book, untimed credential is abundant by construction. If you hold a Boomi certification, an interviewer who knows the format will discount it, and they are not being unfair to do so. My view is that this should change how you present it rather than whether you get it: the certification is worth having because it structured your learning and cost you nothing, and what you should be prepared to do in an interview is describe an integration you built and the failure modes you designed around, which is evidence the badge cannot supply. The strategic question is the one I would want a prospective learner to think about seriously.
iPaaS as a category emerged to solve enterprise integration for people who were not going to write code, and it did that well for fifteen years. The ground is now moving. Code-first orchestration tools have made integration accessible to developers who would previously have found enterprise middleware unbearable. Workflow engines and agent frameworks are absorbing parts of the same problem space.
And the visual, drag-and-drop paradigm that made iPaaS approachable carries a cost that becomes apparent at year three, which is that a complex integration expressed as a diagram is considerably harder to review, diff, test and reason about than the same logic expressed as code in version control. None of this means iPaaS is finished, and large enterprises with hundreds of systems and a mixed-skill integration team have real reasons to stay. But a learner should know they are entering a category under pressure rather than one on the way up. The integration fundamentals underneath are what transfer, and they are genuinely valuable: idempotency, retry semantics and why naive retries cause duplicates, error handling and dead letter patterns, orchestration versus choreography, schema mapping and the pain of nullable fields, throttling against third-party rate limits.
Learn those here for free and they will serve you on any platform or none. The two omissions are the ones I now expect. Integration sprawl gets no treatment, and it is the entirely predictable consequence of making integration easy: three years in there are four hundred processes, nobody knows what a third of them do, several duplicate each other, and the one that silently stopped running in March was discovered in September. A vendor whose revenue scales with usage will not teach you to build fewer integrations.
And exit costs go unmentioned, which matters more here than almost anywhere, because an integration platform ends up holding the business logic that connects your systems, expressed in a proprietary visual format that exports to nothing useful. Migrating is a rebuild. Ownership is worth a footnote too: Boomi was spun out of Dell and is private equity owned, which is not a criticism but is a factor in how much confidence to place in a five-year roadmap. Four point four.
Genuinely free courses and genuinely free certifications with a fair and thoughtfully designed exam format, in a category facing real pressure, with the usual vendor silences on sprawl and exit costs.