Case study 02 · for Planfocus Consulting

Coremetryx

Enterprise business intelligence, wrapped so a small business can actually use it — multi-tenant, branded dashboards built on Amazon QuickSight.

Coremetryx: a funds-under-management dashboard showing trend with forecast, distribution by asset class, and breakdowns by client category, age, risk profile and sector.

The problem

A weekend of spreadsheets

Business data lives across systems — a CRM here, a platform feed there, revenue in a finance package. Getting a single view means exporting all of it and reconciling it by hand, which is how reporting ends up being someone’s weekend.

Firms need reporting they can put in front of a client or a board: dynamic, branded, and current. Enterprise BI tools do this well, but they assume a data team, an analyst, and a budget that a small practice does not have.

What we built

Enterprise BI, operable by a small business

A multi-tenant analytics platform on Amazon QuickSight with complete data isolation per organisation. Six granular user roles from Reader through to Admin Pro. Natural-language querying and automated insights. Fully customisable, branded dashboards that export cleanly for a client meeting or a board pack.

Alongside it, a separate administration application for provisioning and managing QuickSight namespaces and users across every tenant — the operational layer that makes the whole thing maintainable.

The hard part

Making enterprise BI safely multi-tenant.

QuickSight is built for an enterprise running its own analytics, not for a platform serving hundreds of separate organisations that must never see each other’s data. Bridging that gap is the engineering story.

Namespace provisioning per tenant. Per-tenant identity mapped through Cognito. Role mapping that stays correct as people move between organisations. Dashboards embedded securely into a custom front end rather than QuickSight’s own console. And underneath all of it, a guarantee that no tenant can reach another’s data — enforced at the infrastructure layer, not by application logic that could be bypassed.

The Coremetryx dashboard library, listing available dashboards with badges for accessible, editable and AI-enabled.
Each organisation sees only its own dashboards, with per-user permissions applied.

Capabilities

What it does

Any data source

Track growth, identify seasonal variation, and project forward from current trends across whatever systems the business already runs.

Interactive visualisations

Dynamic and flexible on screen, exportable for management reporting and client presentation.

AI-powered insights

Natural-language queries and automated narrative insights, so a non-analyst can interrogate the data directly.

Six user roles

Granular access control from Reader to Admin Pro, so organisations can delegate safely.

Per-tenant branding

Every dashboard carries the organisation’s identity, not the platform’s.

Admin application

A dedicated interface for managing namespaces, users and provisioning across all tenants.

Stack

What it runs on

  • Front endReact, TypeScript, Vite, Radix UI, Tailwind
  • Back endNode and TypeScript, Lambda-deployed services
  • AnalyticsAmazon QuickSight, multi-namespace
  • IdentityAmazon Cognito, per-tenant user pools
  • InfrastructureAWS CDK, environment-separated stacks
  • QualityJest, Biome, typechecked CI

Multi-tenant is hard to retrofit.

If you are building a platform, it is worth getting right at the start.