What happened to PredictionIO after Salesforce bought it
The room usually agrees PredictionIO should be implemented properly on Salesforce, because it is still on an old diagram and two partners have replied with plans. That agreement is the problem: it is not a current product.
Platform & ArchitectureAn architect finds PredictionIO on an old diagram and asks for a project to implement it properly on Salesforce. Two partners reply with plans. Both assume a current product, a licence, and a sync. The room agrees the plans are close enough to compare. That agreement is the problem. Neither plan can point at a screen a buyer can open this quarter. The 2016 report will not support those plans. If the open-source project is still run in-house, that is a different estate. Do not glue the estates together because they once shared a name.
The sentence that draws the split is the one to quote: "PredictionIO’s tech will remain open source and free to use, while the team works on evolving the project into a self-sustaining community." Salesforce bought PredictionIO in 2016. Open source was to stay free. The team’s work was folded into Salesforce’s own machine learning. The software was not described as becoming a paid Salesforce edition. The people were described as taking work into Salesforce’s own machine learning. A buyer who wants "PredictionIO inside Salesforce" is asking for a blend the report does not offer. Free software can be run by the organisation that operates it. Salesforce products that exist now can be bought under their current names. PredictionIO cannot, on this source, be ordered as the current machine-learning product.
"Self-sustaining community" is what the report said the team would work towards for the open-source project. It is not a service-level promise from Salesforce in 2026, and it is not a named support contract. If a project that is run in-house is depended on, the dependence is the maintenance that can be seen in that project. Do not tell an audit that Salesforce operates it because a team was acquired in 2016. The present tense belongs to the report. "Will remain" and "works on" are 2016. They are not a status page for this year. Quote them as history. Do not promote them into a current entitlement.
| Question | Who has to answer it | What a bad answer usually means |
|---|---|---|
| What did the 2016 report split in two | The architect who found the old diagram | Free open source and Salesforce machine learning are being ordered as one SKU |
| Who operates a copy, if one is still running | Whoever runs that process, by name | "Salesforce owns it" is offered as cover for a repository that was cloned |
| What does 19 February 2016 date | The person adding a 2026 release to the roadmap | The report date is used as a sunset, a relaunch, or a merge with Data 360 |
| Where a price or a licence would appear | Procurement, on the order form | "Free to use" is briefed as a Salesforce tier with a fee of zero |
| Where did the tied product end | Whoever names SalesforceIQ as the destination | A later retirement is treated as a screen that can still be configured |
What did the 2016 report split in two?
Keep the split. The open-source side was to stay free to use. Free to use is not the same as "Salesforce will host it". The machine-learning side was folded into Salesforce’s own work. Folded in is not the same as "still available under the old name". It is not a current product to implement. Why Salesforce acquired it, on this source, is the machine-learning work the team was joining, with the open-source project to remain free. That is the motive the report can carry. Do not add a platform strategy from a later decade and backdate it. Why the project was sold, or why maintainers accepted the deal, is not a statement available here. Do not borrow one from a profile.
A good answer sounds like the quoted sentence, the year 2016, and a named refusal to put PredictionIO on a statement of work next to a cloud licence. The person comparing the two partner plans can say both plans assume a current product the report does not offer.
A bad answer sounds like the diagram has been continuous for a decade, so the name should stay so the architecture looks unbroken. Continuity of a box is not an entitlement. Strike the line. If the real need is a current data or agent design, name the current product. Do not keep the 2016 name so the picture looks continuous. Impact on customers splits the same way. If the open-source engine was never run, a PredictionIO screen was not gained when the company was bought, and there is not one now. If it was run, whatever was kept is still operated by the organisation that kept it. Salesforce owning the company in 2016 does not patch that server and does not move training data into a CRM object.
Who operates a copy, if one is still running?
Salesforce bought PredictionIO. Yes, as an acquisition in 2016. What that ownership means is the split above, not a SKU. Do not tell a security review that "Salesforce owns it, so the library is covered". Ownership of a 2016 acquisition is not a support contract for a repository that was cloned. Cover, if it is needed, is an arrangement signed for the estate that is actually run.
A good answer sounds like a named operator, a statement of what data the process may read, and a statement of whether any prediction is allowed to write back to a record. That estate is not a Salesforce implementation, even if a Salesforce org is one of the systems it touches. Touching would be an integration that has to be designed. It is not an entitlement from 2016.
A bad answer sounds like Salesforce owns the company, so the cloned library is in support, and a sync can be added because the names used to match. A partner who puts PredictionIO on a statement of work next to a cloud licence has mixed the free software with a contract vehicle that does not sell it. Staff the system that is actually run, with owners who can explain what it reads and what it writes. Do not staff a SKU the order form cannot list. Staffing a PredictionIO team inside a Salesforce programme, in 2026, is the spend to refuse.
What does 19 February 2016 date?
The acquisition is 2016. The report cited here is dated 19 February 2016. "When did Salesforce acquire PredictionIO" is 2016, with that report date as the day to point at. There is no later completion notice in the style of a modern close sentence. Do not invent one. Do not use 19 February 2016 as a sunset of the open-source idea. The report said the tech would remain free. The date is the report of the acquisition, not a switch-off. Equally, do not use it as the day an org gained a model feature. An org’s features have their own release notes. This report is not one of them.
A good answer sounds like 19 February 2016 labelled as the report date, with no 2026 release beside it, and Data 360 kept as a separate name. Data 360 is a later name, from 14 October 2025, for what had been called Data Cloud. It is not the 2016 project. Do not merge them because both words concern data. Delete a roadmap line that shows PredictionIO landing in Data 360.
A bad answer sounds like the report date is the sunset, or a 2026 release restores the old name inside a current cloud. A timeline that adds a 2026 release has left the record. There is no relaunch date here. SalesforceIQ, which that work was tied to, was later retired. The retirement is a later fact about a different product name. It closes a path. It does not open a new SKU under the old name.
Where a price or a licence would appear
The cited page does not publish a purchase price or a licence fee. How much Salesforce paid, how much PredictionIO was sold for, and the acquisition price are not answered here. Do not fill them. Free to use, in the quoted sentence, is about the open-source tech. It is not a published acquisition price of zero, and it is not a Salesforce licence fee of zero for a current cloud.
A good answer sounds like procurement stating there is no PredictionIO line on a Salesforce order form, and that "free to use" attaches to the open-source tech as described in 2016. It does not attach to a button in a setup menu. Finance is told there is nothing to buy under this name. Finance is not handed a reconstructed 2016 price.
A bad answer sounds like the acquisition included a free Salesforce tier called PredictionIO, because the report said free. A free trial inside Salesforce is the wrong object. Worth is not scored here. A thread that argues the deal was cheap or expensive needs a figure this record does not publish. Leave the argument. A DVD, a Reddit thread, a YouTube video, a TikTok, a course, a book, or a certification is not a source for the date, the price, or a shutdown. A certification study guide will not make PredictionIO a current Salesforce exam topic on the strength of this report. Spending, in a clip, is the same unpublished number.
The facts that will stand are the year, the free open-source line, the fold into Salesforce machine learning, and the later retirement of SalesforceIQ.
Where did the tied product end?
It is not a current product. Do not buy it, staff it, or migrate onto it. That is the practical end state. There is no dated sunset here for the open-source project, and there is no Salesforce retirement notice titled PredictionIO. What there is, is the later retirement of SalesforceIQ, the product that work was tied to. SalesforceIQ was later retired. The 2016 path cannot be followed to a screen a seller can open. Do not swap one retired name for the other and call that a destination.
A good answer sounds like "not a current product", said before any partner demo is booked, and SalesforceIQ named only as a product that was later retired, not as a place to configure. The person who still runs the engine can say the decision to keep or leave it is theirs, on the estate they operate. It is not an order in the 2016 report.
A bad answer sounds like users must move onto PredictionIO, or PredictionIO users must move onto a product this record has not named, because the acquisition implies a destination. Do not force anyone onto PredictionIO. Product decline is not a chart available here. Users forced onto Salesforce assumes a user base being moved onto a current SKU. There is not one to name. A demo of PredictionIO inside a Salesforce org would be a demo of a product the source does not offer. Do not ask a partner for that demo. A case study, a roadmap, and an integration guide book are not documents to hand over for a current build.
If the diagram is still on the wall, take the box down before the next partner meeting. A box with no screen is an invitation to invent a screen.
What to do with the answers
Read the five as a set. Specific answers mean the room can keep a 2016 split from becoming a 2026 order. General answers mean the decision rests on a shared feeling that an old name on a diagram is a product, which the report does not say.
Putting the name on an order form, or describing an open-source process that is still running, is set out in PredictionIO stayed open source, and it is not a current SKU. When the path they name is SalesforceIQ, the retirement is set out in RelateIQ has no org left to implement. Do not migrate onto either name. Do not staff either name as if the menu were merely hidden.



