Software vendors, data companies and analytics providers
DeepGarden for product and software companies
Your customers keep asking for analytics inside your product. Building that properly means a semantic layer, a per-tenant permission model, a query validator and an audit trail — none of which is your product, and all of which you would be maintaining forever.
The demo is a fortnight. The rest is two years.
What follows is the part that decides whether it can ship: tenant isolation that a security reviewer will accept, definitions that do not drift per customer, a guarantee that a generated query cannot write, and provenance strong enough that your customer can defend a number to their own board.
It is table stakes your customers are asking for, and every month spent building it is a month not spent on the thing they actually buy you for.
Where the answer is currently spread
- Your product's own data model, per tenant
- Your customers' data, where you already hold it
- External datasets you resell or enrich
- Your identity and tenancy model
Who is involved
- Chief product officer
- Time to a shippable feature, and what it costs to maintain.
- VP Engineering
- Tenant isolation, query safety and where the operational burden lands.
- Commercial lead
- Whether it can be packaged and priced as its own tier.
- Your customers' security teams
- Exactly what a security reviewer always cares about, one level removed.
Questions this makes routine
These all cross a system boundary, which is why they currently get asked once a quarter rather than once a week.
- Can we put our own brand on it end to end?
- How is one customer's data kept invisible to another?
- Do our customers get the evidence trail, or just the answer?
- Can we define the metrics once and have them apply per tenant?
- What do we own, and what stays yours?
What DeepGarden does about it
What changes, specifically, and what lands in front of the person who asked.
Three deployment models, clearly separated
Internal deployment under the DeepGarden brand, white-label under yours, or embedded where our governed query path sits underneath your interface. They differ on branding, on who owns the customer and on who owns the experience — see the licensing page for what each one actually means.
The governed path, inherited
Whatever sits on top, scoping happens in Roots before a question is planned, generated SQL is validated before execution, and evidence travels with the answer. Your product inherits the audit trail rather than having to build one.
Per-tenant provisioning
Definitions and access modelled per tenant, so a metric agreed once applies correctly across your customer base without being maintained separately for each.
Why the output holds up
The parts of the governed path that matter most in this particular setting.
- Tenant scoping is enforced in the foundation, before a question is planned.
- Read-only is enforced at the database role, so a licensee cannot configure their way past it.
- Evidence is not a premium tier — a white-labelled deployment gets the same traceability.
- Commercial, IP and support boundaries are written into the agreement rather than assumed.
How implementation would work
Scoped per engagement — this one genuinely varies
Start with the deployment model.
Tell us whether you want your brand on it, your interface around it, or both, and we will come back with terms rather than a brochure.