Use cases
Which content problems should you build on Sanity, and when should you buy a point tool?
Short answer
Sanity use cases are worth building when several sites, brands, channels or AI assistants read the same content, because one content model in Sanity, a hosted content platform, serves them all. Buy a point tool when one team needs one finished channel, such as a support chatbot with ticketing. Each use-case page carries Deploi's build, buy or depends call.
- Locking editors to one brand
- Enterprise custom rolesGrowth's built-in roles cover every dataset, and custom roles are available only to Enterprise customers, as of October 2026Sources: Sanity Docs (Roles); Sanity (Plans that scale with you)
- References between datasets
- Enterprise, with the feature enabledBoth datasets must belong to the same project, so cross-dataset references never cross projects, as of October 2026Sources: Sanity Docs (Cross-dataset references); Sanity (Plans that scale with you)
- Semantic search for AI answers
- Dataset Embeddings, on every planGenerating embeddings is included; semantic queries count against a monthly quota that Free stops at and Growth bills past, as of October 2026Sources: Sanity Docs (Dataset Embeddings); Sanity (Plans that scale with you)
- Documents in a merged estate
- No document overage on GrowthThe quota counts every dataset in the project, so CMSs merged past Growth's ceiling, quota add-on included, move to Enterprise, as of October 2026Sources: Sanity Docs (Plans and payments); Sanity (Plans that scale with you)
Which Sanity use case matches the problem you have?
Three Sanity use cases are live as of October 2026, and you pick one by asking what reads your content: several brands, several content management systems (CMSs), or people and AI assistants asking questions.
- You run several brands or regional sites: start with running several brands and sites from one Sanity setup.
- You run several CMSs and want one content platform: start with consolidating several CMSs into Sanity without a big-bang migration.
- You want people and AI assistants answered from the same content: start with building a knowledge base with AI answers on Sanity.
- You need help building it: talk to Deploi about scope before you choose a plan.
A problem that fits none of the three is usually one channel for one team, which means a tool purchase (Deploi's judgment).
Which plan gate does each Sanity use case turn on?
Two of the three live Sanity use cases can push you from Growth, Sanity's self-serve paid plan, to an Enterprise contract, and the knowledge base runs on every plan, as of October 2026. Brand isolation needs Enterprise because custom roles are Enterprise only (Sanity Docs, Roles, October 2026). A merged estate gets there once it passes the ceiling of Growth's Increased quota add-on, because Growth allows no overage above its included document quota (Sanity Docs, Plans and payments; Sanity pricing, October 2026).
| Use case | Experience label | Plan gate as of October 2026 |
|---|---|---|
| Several brands from one setup | Documentation | Enterprise custom roles to lock editors to one brand |
| Content management system (CMS) consolidation | Documentation | Enterprise once merged documents pass Growth's add-on ceiling |
| Knowledge base with AI answers | Production | None: Dataset Embeddings run on every plan, queries are metered |
How do you run several brands, sites or CMSs on Sanity?
Run several brands on one Sanity setup by design, or merge the content management systems (CMSs) you already run one source at a time: the two live use cases start from different places.
- Several brands by design. The multi-brand setup turns on what must stay apart, not on brand count. Growth's built-in roles cover every dataset, so locking editors to one brand needs Enterprise custom roles (Sanity Docs, Roles, October 2026).
- CMSs you already run. The consolidation program moves one CMS at a time and counts a source done only when it is switched off. The merged estate, not one site, decides when Growth runs out, because the document quota counts every dataset in the project (Sanity pricing, October 2026).
Either can also need references between datasets, which require an Enterprise plan with the feature enabled (Sanity Docs, Cross-dataset references, October 2026).
How do you put Sanity content behind search and AI answers?
Put Sanity content behind AI answers by storing each answer as a structured document and retrieving it with semantic search inside GROQ (Graph-Relational Object Queries), Sanity's query language. The knowledge base with AI answers builds the answers first and the chatbot second.
No plan gate blocks it, because Dataset Embeddings are available on all plans. Queries that use text::semanticSimilarity() count against a monthly quota that Free stops at and Growth bills past (Sanity Docs, Dataset Embeddings; Sanity pricing, October 2026).
The knowledge base is the one use case labeled production. deploi.ca's Shopify FAQ assistant scores answers with text::semanticSimilarity() in GROQ and falls back to keyword scoring, and the semantic search build shows those queries. Deploi hasn't measured the assistant's retrieval accuracy and quotes no figure for it.
What do the build, buy and depends verdicts on these pages mean?
Each use-case page carries Deploi's call in one or two of three verdicts: BUILD means build it on Sanity, BUY means buy a point tool, and DEPENDS names what to price or decide first. The verdicts are conditions, not a ranking, so a page can carry two.
- BUILD: the conditions under which the pattern belongs on Sanity, such as brands that share most of their content model and one team that owns the schema.
- BUY: the conditions under which a point tool wins. The knowledge-base page says to buy a support chatbot tool when you need agent handoff, ticketing and conversation analytics on day one, because Sanity stores and retrieves answers but ships no help desk.
- DEPENDS: what to price before you commit, such as Enterprise or a separate project per brand when each brand's editors must be blocked from the others' content.
Deploi’s point of view
- Our take
- Build on Sanity when more than one site, brand, channel or assistant reads the same content and one team will own the model; buy a point tool when one team needs one finished channel. Read the plan gate before the price, because locking editors to a brand and referencing across datasets are Enterprise features as of October 2026.
- Where we disagree
- The common pitch is that a headless content platform fits every content problem. A support desk with ticketing, handoff and conversation analytics is a product to buy, not a schema to build. The second wrong claim is that a workspace or a dataset per brand keeps brands apart: on Growth it separates screens, documents and assets, not editors.
- What we’ve seen
- deploi.ca's main pages run on other CMSs, and only its Shopify sections run on Sanity. There, one Sanity Studio runs three workspaces over two datasets in one project, the Shopify FAQ assistant retrieves answers with semantic search in GROQ and a keyword fallback, and Deploi moved its standalone FAQ project into the shared project on 26 September 2026, per its README. Two of the three use cases carry the documentation label because their verdicts rest on Sanity's documentation and public sources.
- What we refuse to promise
- A price or a timeline on a use-case page. Deploi quotes Sanity work by scope, after counting your brands, sources and documents, so cost questions go to a call, not to a figure here.
- What we'd tell you to do first
- Name everything that will read the content before you pick a use case. One reader is usually a tool purchase; several readers sharing one model is where a Sanity build pays off (Deploi's judgment).
When this answer changes
- Sanity makes custom roles or cross-dataset references available on Growth
- Sanity changes which plans include Dataset Embeddings or how semantic queries are billed
- Sanity allows document overage on Growth
Sources
Every source opened and checked on 25 October 2026.