How is Zoomin licensed now it sits inside Data 360?

Zoomin was a knowledge and documentation platform before Salesforce bought it. Its job was to take the documents a company already has, product manuals, help articles, release notes, policy PDFs, and make them findable and usable by customers, employees and, increasingly, AI agents. After the acquisition that job moved into Data 360, which Salesforce renamed from Data Cloud, and into a packaged capability called Enterprise Knowledge. If you are budgeting for it, you are budgeting for Data 360 consumption, and the interesting questions are about what drives that consumption.

#Where Zoomin sits in the price list

Salesforce's announcement described Zoomin as "a leading data management provider for unstructured data" and recorded that the purchase completed on 1 November 2024. Its own product blog is explicit about the lineage: "Enterprise Knowledge powered by Data Cloud is a powerful suite of unstructured data capabilities that connects and processes your organization’s enterprise knowledge across various sites and systems and activates it in Agentforce." The same blog calls these capabilities the first milestone in integrating Zoomin's technology. On the Service Cloud pricing page, Enterprise Knowledge appears as a line marked "Available for purchase", and Data Cloud appears as "Limited Data Cloud features", which tells you the entry point for a service team is a modest allowance with a paid upgrade.

We read the Data 360 pricing page on 13 October 2026 and found no line named Zoomin. That is the expected result. The commercial unit for unstructured work is the Data 360 action, paid for in credits or, for profile-based work, inside a per-profile price. Salesforce also offers a way to try the platform: "Salesforce customers can provision Data 360 for free with a limited amount of storage and consumption credits."

#The three Data 360 models side by side

The page offers three ways to buy. Flex Credits cost US$500 per 100,000 credits and can be applied to any Data 360 action; the page adds that credits can be moved between Data 360 and Agentforce. Profiles cost US$240 per 1,000 profiles a year and bundle the profile-building actions. Enterprise Profiles cost US$420 per 1,000 profiles a year and include more. Salesforce says there are pre-purchase, pay-go and pre-commit options for Flex Credits.

The row that matters for a Zoomin-style workload is the unstructured one. It reads "Tied to Usage" under every plan. In plain terms, choosing a profile plan does not buy you unlimited document processing. The page lists what you pay for under this heading as Unstructured Data Processing, described as analysing and extracting insights from documents, text, audio and other unstructured sources, and Intelligent Processing, described as AI-driven processing to interpret and enrich data. Both are metered.

#Why most knowledge workloads belong on credits

Profile pricing is built for a particular shape of use: you have a customer base, you want unified profiles and segments, and you want a predictable annual number. On the numbers, 500,000 profiles would cost US$120,000 a year on Profiles and US$210,000 on Enterprise Profiles, before any usage-tied actions. The page lists "1 Flex Credit per profile to scale & grow" for Profiles and 2 for Enterprise Profiles; confirm in writing exactly what those credits can be spent on.

A knowledge workload looks different. A software company grounding a support agent in its documentation may have a few thousand profiles and several hundred thousand pages. Its cost is set by documents, versions and refresh cycles, not by people. For that buyer, a profile plan pays for capability it barely uses, while Flex Credits let the spend track the library. The right test is to size both models from your own counts and compare, rather than defaulting to the plan the account team proposes first.

For scale, 10 million credits is US$50,000 at list. Whether that is generous or tight depends entirely on the credits-per-action figure on the Data 360 rate card, which the public page links to but does not reproduce. Ask for it, and ask what a "unit" of unstructured processing is: a page, a document, a chunk, or a megabyte.

#What drives the credits a library consumes

Four variables set consumption, and each can be measured before you sign.

The first is the size of the library, counted in documents and in their average length. A thousand short articles and a thousand 300-page manuals are different jobs. The second is the number of distinct sources, since each connector adds its own sync and mapping. The third is refresh frequency. Ingestion is free on the pricing page, but processing is not, so a library that changes daily and is re-processed in full each time can cost many times a library re-processed on change only. The fourth is what you do after processing: indexing for retrieval, summarising, extracting fields, and querying all draw on the same pool, so a design that retrieves more per answer uses more.

This is why a design decision can be a licence decision. If you decide to refresh only changed documents, restrict the index to approved content and keep stale material out, you reduce credits as well as risk. A workable pilot measures credits consumed for one source over two refresh cycles, then extrapolates.

#Two libraries, two very different bills

Without the credits-per-action figure we cannot turn this into dollars, but we can show how far apart two designs can sit in units of work, which is what multiplies by the rate.

Take a library of 10,000 documents. In the first design, the whole library is re-processed every night so that nothing is ever stale. Over a thirty-day month that is 300,000 document-processings. In the second design, the connector detects changes and re-processes only what changed. If about 200 documents change in a typical week, that is roughly 800 processings a month. The first design does 375 times the work of the second for the same answers, and it does so every month for as long as the contract runs. Even if the credit cost per document is tiny, a factor of that size is not.

Now add the initial load. Both designs process the 10,000 documents once at the start, so that cost is common. The difference lies in the steady state, and the steady state is what a three-year licence pays for. This is why the licence conversation and the design conversation should happen together, and why a vendor quote built from "number of documents" alone is incomplete. Ask for it to be built from documents processed per month.

Think also about audio and rich media. The page lists audio among unstructured sources. A call-recording archive, a webinar library or a video training catalogue can be larger than any text library you hold, and its processing is metered the same way. Decide early whether those sources are in scope, because adding them later changes the forecast by an order of magnitude.

#What Salesforce does not say, and what to ask

The announcement said Zoomin would help Agentforce give responses grounded in company knowledge, and that Service Cloud customers would be able to use enterprise knowledge to automate service interactions. It did not give a price, and it did not say that the capability would be included in any edition. Do not assume inclusion from a marketing sentence.

Questions for your account team should be concrete. Is Enterprise Knowledge purchased as an add-on, a bundle or an entitlement within an edition? What is the credit cost of each unstructured action in my contract? Are my existing Data Services Credits, if we have them, valid for the new processing, given the page says there is no forced migration to Flex Credits or Profile pricing and that moving means swapping all of your existing Data Services Credits? What happens to my rate if I overrun? And which of my current knowledge tools, including any Zoomin subscription I already pay for, stops being needed?

That last question is the one most often missed. If you hold a legacy Zoomin subscription, check its end date and whether it can be credited or co-termed. Paying for the old tool and the new consumption in parallel for a year is a common and avoidable cost.

#Rightsizing without guessing

Rightsizing a consumption licence is closer to capacity planning than to counting seats. Start with a source inventory: every repository, site and file store you might connect, with owner, document count, average size, language, and how often it changes. Rank the sources by value to the use case, and decide which are in scope for the first release. A first release that covers the three sources your agents actually cite is cheaper and more useful than one that tries to cover thirty.

Then measure. Use the free provisioning allowance Salesforce describes, or a sandbox, to connect one representative source and watch Digital Wallet for two refresh cycles. Record credits per thousand documents for the initial load and for a typical refresh. Multiply by your inventory and your refresh plan, and add a margin for growth and for the retrieval and query activity that follows.

Finally, decide the buying model with the measured number in hand. A steady, predictable library suits a pre-purchased block at a better rate. A library that is still being shaped, with sources added each quarter, may be better served by pay-go or a lower pre-commit with the right to top up. Whatever you choose, write the rate for additional credits into the contract now, because the price you negotiate under pressure in month ten is rarely the best one.

#Contract mechanics and monitoring

Data 360 has the same shape as Agentforce in one respect: it is consumption, and consumption needs a meter you can read. Salesforce's Digital Wallet is described as a free account management tool that gives near real-time data to monitor how you use consumption-based products, and it is the place to set alerts. The page also says billing depends on how your org is contracted and that a monthly usage summary is emailed to the billing contact. Make sure that person exists, and that someone with authority to act reads the summary.

Credits shared between Data 360 and Agentforce are a flexible feature but also a place where budgets blur. A knowledge pipeline and an agent drawing on the same pool will compete, and a spike on one can exhaust the other. Agentforce Flex Credits do not roll over at the end of a term, and if Data 360 credits behave the same way in your contract, a half-used pool at renewal is a loss. Check the order form for expiry on both products, and name an owner for the pool.

#Comparing offers

When you compare a Salesforce offer with an alternative knowledge platform, normalise to the same unit. Use documents processed per month, answers served per month and refresh cost per source. Include the cost of an editor or curator, since unstructured data does not govern itself, and the cost of retiring the old content. A cheaper licence that leaves stale pages in the index will cost you in wrong answers.

For the work around the licence, covering sources, owners, refresh, allow lists and deny lists, see Zoomin, knowledge, and grounding. The acquisition record is on what happened to Zoomin, the connector and object detail on integrating Zoomin with Salesforce, and the decision to move on should I migrate from Zoomin. For data platform work more broadly, see Data 360. Salesforce's completion notice reads: "Salesforce completed its acquisition of Zoomin on November 1, 2024." The Agentforce pricing page, read the same day, notes that "Unused Flex Credits do not roll over into subsequent subscription terms." which is a reminder to check expiry terms in any credit contract.