What happened to MuleSoft
The decision is usually taken in a room that has already agreed MuleSoft became Data 360. That agreement is the problem: it is easy to reach from an October sentence, and it fits a rename of a different product, which is not what carries the order.
Platform & ArchitectureThe release usually enters the room already stopped. Somebody says MuleSoft has become Data 360, and the conversation starts at how to freeze the flow rather than at which product the October sentence renamed, because the agreement feels settled: the integration platform is gone. The flow is still there. The agreement is the problem. It is easy to reach, and it fits a rename of a different product, which is not what carries an order.
On the Thursday in question the order feed was a MuleSoft flow that posted orders from the finance system onto the account. It had a named owner, a key, and a queue for failures. The fault sitting in that queue was a mapping error, the same class of fault the feed had been throwing before anyone opened the pack. The steering pack cited October 2025 and drew one box with both names on it. The integration owner froze the change window rather than deploy a flow into a product they had been told no longer existed. The orders kept failing in the old way while the meeting argued about a label.
The distinction that matters is short enough to write under the diagram. Salesforce bought MuleSoft in 2018, and MuleSoft kept its name. Salesforce owns the product. It is an integration platform: a message leaves a system of record, a flow decides whether Salesforce may create or update a record, and errors wait for the person who owns the queue. Data Cloud took the name Data 360. A seller can still read a subset of the order in Salesforce while the finance system, the warehouse, or the bank remains the owner of the order. Those are different sentences. Only the first can pause a Thursday release, and only when someone has actually changed the flow.
What follows are the questions that separate the sentences. A good answer names a person, a year, a contract, or a queue. A general answer at this stage becomes a forced move later, with the mapping error still unfixed.
The questions, and what a bad answer signals
| Question | Who has to answer it | What a bad answer usually means |
|---|---|---|
| 1. What happened to the product | The person who wrote the steering pack | A rename of Data Cloud is being treated as the end of the flow |
| 2. Who owns the runtime | The integration owner, by name | The sandbox is being treated as if it already contains the flow |
| 3. Which date sits on which line | Whoever drew the timeline | 2018 and 14 October 2025 have been merged into one event |
| 4. Where the price is written | Procurement, from the contract | A figure from memory is about to enter the business case |
| 5. Whether there is an end date | Whoever proposed taking the flow down | The purchase year is being read as an expiry |
1. What happened to the product, and is that the same event as the rename?
The question is not whether the slide looks tidy. It is what changed for a customer who already runs a flow, and what the sentence in the pack actually says.
Salesforce bought MuleSoft in 2018, and MuleSoft kept its name. That is the event. The product you ask for in a support ticket is still MuleSoft. The name on a purchase order can still say MuleSoft. Customers who already run flows keep running flows. The acquisition changed the owner of the product. On the facts used here, it did not hand you a new integration method, and it did not switch customers onto a different product by the deal itself. If a programme is moving a flow, that is because someone chose a new design. The choice needs an owner, and it can be tested.
The help page cited here is about Data 360. It states the rename in one sentence: "As of October 14, 2025, Data Cloud has been rebranded to Data 360." Use that sentence to correct a label. Data Cloud took the name Data 360. The sentence does not mention MuleSoft. It does not say MuleSoft was folded in. It does not say an integration flow became a Data 360 pipeline. It also does not say why Salesforce acquired MuleSoft, or why MuleSoft sold. A diagram that draws one box and writes both names on it is the defect. Separate the boxes before anyone argues about cost.
A good answer sounds like the integration owner, in the Thursday meeting, saying the flow that posts orders from the finance system onto the account is still MuleSoft, bought in 2018, name kept, and the October sentence is the Data Cloud rename quoted above.
A bad answer sounds like the pack saying MuleSoft became Data 360 in October 2025, so the change window stays frozen until the runtime is redrawn as a pipeline. Every clause may feel current. None of them is the sentence on the help page, so none of them can be checked against it.
If both products are genuinely in the programme, write one sentence for the messages MuleSoft may carry and one sentence for the data set Data 360 may hold. If the sentences match, you have one job wearing two names.
2. Who owns the runtime, and who owns the flow?
Salesforce owns MuleSoft. You do not own the platform because you bought a Salesforce org. Your organisation owns the contract you signed, the credentials you issued, and the flows you asked someone to build. Those are different from the vendor's ownership of the product. Say the difference out loud when a risk register treats them as one line.
Is MuleSoft owned by Salesforce? Yes. Did Salesforce buy MuleSoft? Yes. Who owns MuleSoft? Salesforce. Say the three answers the same way in the steering pack, the vendor form, and the support template. A form that still lists MuleSoft as an independent company, with a security review that ignores the Salesforce relationship, is stale. A form that lists MuleSoft as a feature inside the CRM sandbox is also wrong. Ownership of the company is not the same fact as inclusion in an org.
Your system of record can still sit outside Salesforce. The acquisition did not move the finance system, the warehouse, or the bank into the CRM. MuleSoft is the platform that can carry an agreed message between them. Who owns the order remains a design choice you write down. Salesforce owning MuleSoft does not make Salesforce the owner of every system the flow touches. If a supplier asks who owns the product, give them Salesforce, the year 2018, and the statement that the name was kept.
A Salesforce sandbox does not include MuleSoft. Refreshing the sandbox does not deploy an API and does not replay yesterday's messages. You point a separate runtime at that sandbox on purpose. Anyone who says the sandbox already includes MuleSoft should open it and show the flow. The absence is the finding. The acquisition did not change that boundary. Buying Salesforce did not turn the runtime on. Keeping the name means you still ask for MuleSoft by that name. It does not mean every org has a runtime. Your contract says whether you do.
A good answer sounds like a named integration owner who can show the key, the queue, and the contract that says whether this organisation has a runtime, and who can open the sandbox and show that the flow is not in it.
A bad answer sounds like "the sandbox already includes it, because Salesforce owns both." That sentence collapses a vendor, a contract, and a runtime into one login. The order feed will still fail in the queue nobody is watching.
Copying rows into Salesforce and calling the paste MuleSoft is the wrong job. An integration is a contract, an error path, and a system of record. If the plan says to migrate everything, ask which event. If nobody can name the event, you are looking at a paste. You still call the product MuleSoft. You still need a runtime the sandbox does not contain. You do not receive Data 360 because you have MuleSoft, and you do not lose MuleSoft because Data Cloud was renamed.
3. Which date sits on which line?
When did Salesforce acquire MuleSoft? In 2018. There is no day and month to add on top of that year from the record this guide follows. The parent record uses 2018. If a slide needs a day, it needs a document that states the day. A timeline that looks precise and is guessed is worse than a timeline that says 2018 and stops.
Put 2018 on the timeline as the acquisition. Put 14 October 2025 on the timeline as the Data Cloud rename, and label that line Data 360. Do not draw an arrow from the 2025 line back onto MuleSoft. The dates sit near each other in a search box. They are not one event. A reader who merges them will freeze a release for a rename that never touched the runtime.
The cited page is the Data 360 help page. It is not a 2018 press release. State the year. Send a reader who wants the legal instrument to the contract. Your own go-live, if you have one, is a third date, and it is yours.
A good answer sounds like two labelled lines: MuleSoft, 2018, name kept, and Data 360, 14 October 2025, with the quoted rename sitting only on the second line.
A bad answer sounds like a single arrow from MuleSoft to October 2025, with a day in 2018 that nobody can point to in a document. The day will be repeated until someone treats it as a notice.
4. Where is the price written?
How much did Salesforce pay for MuleSoft? The page this guide cites does not publish a MuleSoft purchase price or a licence fee. How much was MuleSoft sold for? Same answer. A Salesforce MuleSoft acquisition price does not appear on that page. A blank is more honest than a borrowed number with a confident tone.
The deal price, even if a historian later shows you one in another document, is not your licence. A historical figure would not tell you what you pay next year. Your licence is the contract in your drawer. If the contract is silent, the honest line in the pack is that the price is not in the paper you are holding. Procurement can ask the account team for the commercial paper. That paper is the price that binds you.
What you pay a partner to connect two systems belongs in a statement of work. That figure is not the 2018 purchase price. A quote that leads with a historical deal number, and never names the event the flow carries, has answered the wrong question.
A good answer sounds like procurement, having read the order form, stating the term and the fee your organisation actually pays, or stating that the order form is silent and the account team has been asked for the commercial paper.
A bad answer sounds like a business case that pastes a purchase price from memory, a video, or a forum next to the year 2018, and treats that number as next year's licence. The cited page will not back the figure, and the contract was never opened.
A blank price is a blank. It is not an invitation to average three rumours. Say the limit in the meeting in one line. The cited page does not publish a purchase price or a licence fee. Then talk about the flow.
5. Is there an end date, and what is still running?
Was MuleSoft shut down or given a sunset date? No end date is stated here. An acquisition is not a shutdown. A rename of Data Cloud is not a shutdown of MuleSoft. A keyword that pairs the product with sunset is a question. The answer is that no end date is stated. Write that in the column, and leave the column otherwise empty.
If a vendor tells you to move off MuleSoft because Salesforce bought it, ask for the retirement notice. A year in a search box is not a notice. 2018 is the year of the purchase. It is not an expiry. Customers still run MuleSoft as MuleSoft. The name was kept so that you can still ask for it by that name. A support template that has already swapped the name for Data 360 will send the ticket to the wrong queue.
Confirm your own contract term, your renewal, and the person who owns the runtime. Those are your dates. They are not a product sunset. A forced move onto a Salesforce object, because someone said the platform was finished, will strand the order feed. Leave MuleSoft up until you have a written replacement for the messages it carries, tested, with an owner. That replacement is a project you choose. It is not a conclusion you can draw from 2018 or from October 2025.
A good answer sounds like the contract owner reading the renewal date from the order form, and the integration owner naming the messages the flow still carries, with the failure queue still assigned.
A bad answer sounds like "we should come off it, the acquisition was 2018, and October 2025 finished the rest." That is two dates doing a job neither date was given. The mapping error will still be in the queue on Friday.
A DVD, a Reddit thread, a YouTube clip, a TikTok, a course, a book, or a certification is not a source for the date, the price, or a shutdown.
Where the decision goes next
The build, the system of record, the event, the key, and the person who reads the failure queue live on the MuleSoft implementation page. Read that when the pack can say the owner, the year, the missing price, and the missing end date. The order feed is a different conversation from the acquisition.



