I Went Looking for Boomi's Obituary. I Found a Different Story.
A consultant told me recently that he has no clients on Boomi anymore. His theory was the familiar one: private equity bought it, squeezed it, and the product went downhill. I nodded along, because it matched a story I already believed about what PE does to software companies.
Then I did what I tell everyone else to do before a platform decision. I checked.
The obituary nobody can source
Here is what the record actually shows. Boomi has been a Leader in Gartner’s Magic Quadrant for iPaaS twelve years running, and in the 2026 report it was positioned highest of anyone on Ability to Execute. It reports over 30,000 customers, up roughly fifty percent in about three years. And far from harvesting the install base, it has been on a buying spree: data integration, managed file transfer, API management, an AI gateway. That is expansion behavior. I looked for layoff reports and found none on the public trackers; headcount data shows slight growth. The strongest negative evidence I could find was anonymous employee reviews, which is not nothing, but it is not an obituary either.
So why does my consultant have zero Boomi clients? Probably the honest answer: the complaint that actually holds up is cost. Third-party procurement data puts Boomi’s per-connection pricing in the range of nine to nineteen thousand dollars a year, per connection, with reported annual escalators of a few percent at renewal. For a mid-market company with a dozen integrations, that math pushes you elsewhere fast. Boomi did not get bad. For a certain size of company, it got expensive, and consultants follow their clients’ budgets.
That distinction matters, because “the product is dying” and “the pricing model does not fit you” lead to very different decisions.
Stop comparing logos. Compare meters.
The real lesson from a week of digging: every platform in this space bills on a different unit, and the meter predicts your costs at scale better than any feature list.
Boomi meters connections. Workato meters tasks on top of a platform edition fee. Zapier meters tasks too, and here is the trap: a task is every action step that runs, not every workflow, so a five-step automation burns quota five times as fast as the plan name implies. Make meters credits, one per operation, and renamed the unit this year without changing how it works. MuleSoft historically metered vCores at four figures per vCore per month and is reportedly moving new customers onto flows and messages. Celigo meters integration flows and transaction volume. None of the enterprise vendors publish list prices, so every number you read, including mine, is third-party reporting, not a quote.
Run your actual volumes against the meter before you fall in love with a demo. Cheap at ten thousand tasks a month can be ruinous at ten million, and the reverse is also true.
The honest 2026 shortlist
NetSuite-centric mid-market: Celigo. Purpose-built for the NetSuite world, prebuilt integration apps that get you live fast, and reviewers keep putting it at the top of the satisfaction charts. Reported entry pricing runs somewhere in the low tens of thousands a year, well under the enterprise platforms. It gives up some enterprise plumbing that Boomi has, like master data management and heavier API governance, which is exactly the trade a mid-market NetSuite shop should usually make.
Modern enterprise automation: Workato. Eight straight years a Gartner Leader and furthest in vision three years running, with the strongest story for letting business teams build alongside IT. Watch the task meter at high volume, and ask about their flat-rate options for high-volume recipes before you assume the worst.
Big, Salesforce-heavy, well-funded: MuleSoft. Still the heavyweight for API-led enterprise architecture. Budget honestly: third-party estimates put real year-one cost for a mid-market deployment in the mid six figures once implementation, training, and support are counted, often two to three times the subscription itself. If that number made you flinch, it is not for you, and that is fine.
Boomi: still a legitimate enterprise choice, especially for hybrid estates that want integration, master data, and API management from one vendor. Just price the connection model against your integration count before you sign, and negotiate the renewal escalator while you still have leverage.
Light business ops: Make or Zapier. Zapier is the easiest on-ramp with the widest connector catalog, running about twenty to thirty dollars a month at the entry paid tier. Make does comparable work for noticeably less per unit, with a free tier generous enough for real experiments. Both meter per step, so cost climbs linearly with volume: fine for back-office automations, wrong for data pipelines.
Self-hosted and personal: n8n. My pick for tinkerers and cost-sensitive teams with engineers, and it is no longer a fringe bet: SAP invested this spring at a valuation north of five billion dollars and is embedding n8n inside its own agent tooling. Two eyes-open notes. It is fair-code, not open source, licensed for your internal use but not for reselling as a service. And the free community edition leaves out the high-availability and governance features an enterprise would demand, so run it on Postgres with queue mode if you mean it, and know what you are not getting.
The NetSuite reality check
If NetSuite is your hub, two facts matter more than any vendor comparison.
First, the ceiling is usually NetSuite, not the platform. Oracle governs concurrency at the account level: a base service tier allows fifteen concurrent requests, shared across SOAP and RESTlets, and each SuiteCloud Plus license adds ten more. Blow past it and you get throttled with a 429, or worse, a misleading login error that sends you debugging authentication while the real problem is volume. There is even an API for checking your own limits, and any integration worth its fee should call it. No iPaaS purchase fixes a ceiling that lives in your NetSuite contract.
Second, there is a clock running. NetSuite has published a deprecation path for SuiteTalk SOAP web services that, per multiple implementation partners, ends with SOAP integrations shut off entirely around 2028, with new SOAP integrations disallowed well before that. If you have integrations older than a few years, some of them are almost certainly SOAP. The next platform conversation is coming either way; the schedule is the only choice left. Verify the dates against Oracle’s release notes, then inventory what you have.
And the AI agent stuff?
Every vendor in this market now leads with agents, and the numbers are big: Gartner expects a third of enterprise software to include agentic AI by 2028, up from almost none in 2024. I use AI heavily in my own integration work, so I am not the skeptic in the room.
But I will leave one more Gartner number next to those: they also predict that more than forty percent of agentic AI projects will be canceled by the end of 2027. Buy the integration platform for the integrations. Treat the agent features as upside, not as the reason.
The bottom line
I started this post expecting to write a takedown and ended up writing a correction, including to myself. The platforms did not sort into good and bad. They sorted into meters and fits: Celigo if NetSuite is your center of gravity, Workato for modern enterprise automation, MuleSoft when you have the budget and the API estate to justify it, Boomi if its connection model matches your shape, Make and Zapier for light ops, n8n if you want to own the thing.
Before you sign anything, four questions. What is the meter, and what does my real volume cost on it in year three? Who maintains this after go-live? What happens at renewal? And which of my existing integrations are quietly running on borrowed time?
The consultant was not wrong about his clients. He was wrong about the company. Check the meter, check your volumes, and check the story you already believe before you let it make the decision. That last one is most of what I do for a living, and it is also how you avoid paying the integration debt tax twice.