Built and maintained with AI-MSL, delivered as a fully managed service on the AI-MSL Platform.
Some figures on this page are drawn from the engagement plan and will be updated with measured results at V1 acceptance.
Summary
Gemini Health delivered reports to its clients once a month, assembled by hand from internal extractions and Excel and sent through client engagement managers. CloudGeometry replaced that process with a secure analytics portal: clients log in, build reports in real time, see their report history and order new reports by prompt. Behind the portal, an AI-powered process built and maintained by CloudGeometry through AI-MSL pulls the right data from Power BI and the Gemini Health data warehouse and applies every security policy to each report as it is generated. Clients now get more data than before, faster, with Expert Supervision by senior CloudGeometry engineers on every change.
The Customer
Gemini Health is a healthcare benefits and data company serving provider and payer organizations. Its clients, hospital networks, physician groups, regional payers and insurance companies, rely on Gemini Health's reporting to manage cost, quality and compliance programs. CloudGeometry has been Gemini Health's engineering partner since modernizing its platform from a monolith to a multi-tenant SaaS on AWS, so the portal program started from a known system rather than a cold assessment.
The Challenge
Client reporting ran on a legacy monthly process. Client engagement managers delivered reports built from internal data extractions and Excel. Every report took analyst time, clients waited a month for answers, and what they could see was limited to what the process could assemble by hand.
Gemini Health wanted clients to serve themselves, without loosening its HIPAA posture:
- Self-service for clients. A portal with client login where clients build their own reports in real time, rather than waiting for the monthly cycle.
- Hard multi-tenant isolation. One organization must never see another's data, enforced in the portal's authorization layer and at the data-model level.
- Security applied to every report. Minimum-necessary access to protected health information, SSO and MFA, encryption in transit and at rest, and an audit trail of who accessed and who changed what.
- More data than before. Reports drawing on Power BI, the Gemini Health data warehouse and other source systems, not just what had been extracted for the monthly cycle.
The Solution
A client analytics portal. Clients log in to a dedicated portal, build reports in real time, see the history of every report they have run, and order new reports in plain language by prompt.
An AI-powered process behind it, built and maintained through AI-MSL. The portal is fed by an AI-powered process that CloudGeometry builds and maintains through AI-MSL. It retrieves the right data from Power BI, the Gemini Health data warehouse and other source systems, and applies all of Gemini Health's security policies to each report as it is generated, so access rules travel with the data rather than being checked once at login.
Why AI-MSL
Three properties decided it.
- Governance built in. Every unit of work is a Change that moves through a fixed lifecycle, from requirements and specifications through code, tests and documentation, with a human gate between stages. For a HIPAA-regulated build, the audit trail is a by-product of how the work happens, not a document written afterwards.
- Expert Supervision by senior CloudGeometry engineers. AI does the implementation; senior engineers hold the gates: solution selection, architecture review, scope validation and quality validation, and an AI Lifecycle Engineer signs off on every Change. Gemini Health gets senior-engineer judgment on every change, priced per Change rather than per retained headcount.
- A lifecycle, not a project. After V1, the same platform maintains the portal and the data process behind it, and new client requests enter the same governed pipeline.
Delivery
Most of the early effort went into requirements and security rather than implementation. Gemini Health's product owner resolved product decisions conversationally with AI-MSL, while the supervising engineers pressure-tested every HIPAA-relevant choice at the gates: minimum-necessary access, session and token lifecycle, and isolation at both the portal and data-model layers. Each decision flowed into the requirements, acceptance criteria and QA scenarios of the affected Changes. Tenant isolation is re-verified on every change that touches the data path, not once at launch.
Evidence by Design
For every Change, AI-MSL keeps a Chain of Custody: the requirement history, every gate decision with the approving engineer and time, and the specification and test evidence, in an append-only audit store. Engineering access to Gemini Health's environment runs through time-boxed, tenant-approved support grants, every operator action attributable and revocable by Gemini Health.
The Benefits
| Dimension | Result |
|---|---|
| Client reporting | Monthly manual delivery replaced by a self-service portal with real-time report building, report history and reports ordered by prompt |
| Data access | More data than before: Power BI plus the data warehouse and other source systems, not only the monthly extractions |
| Security | Gemini Health's security policies applied to every report as it is generated; tenant isolation enforced at portal and data-model layers and re-verified per change |
| Analyst time | Report assembly by hand through engagement managers and Excel replaced by an automated, governed process |
| Compliance posture | Change-level audit trail exportable for HIPAA assessment |
| Cost | Transparent per-Change pricing in DevCredits, no team ramp-up, no bench |
| Steady state | The portal and the data process maintained through the same governed pipeline, with new client requests entering as Changes |
What Gemini Health bought. Not a portal. A capability: the ability to change a regulated, client-facing system quickly, safely and with proof. New reports, new data sources and new clients enter the same pipeline as the original build, with the same gates, the same supervision and the same audit trail.
Key Features
Governed Changes. Every unit of work moves through a fixed lifecycle with a named human decision at each gate. Each Change is delivered as a tested branch; merging into main is Gemini Health's decision.
Expert Supervision. Senior CloudGeometry engineers hold the gates. AI executes; experts supervise.
Evidence by Design. Chain of Custody for every change, time-boxed and tenant-approved operator access, and an append-only audit store that a compliance team can export rather than reconstruct.
AI-MSL is delivered as a fully managed service on top of the AI-MSL Platform, an AI-powered SDLC with expert supervision built into it. CloudGeometry operates the platform, and senior CloudGeometry engineers review the gates and sign off on every Change. Client code and data remain in the client's environment. To see the platform run a Change end to end, ask for the live demo.
Contact us
See how AI-MSL can map your existing codebase, accelerate planning, and bring supervised automation to your development process.
Ready to Accelerate Delivery?
Transform fragmented development workflows into a governed AI-powered delivery pipeline with faster execution, better quality, and measurable outcomes.
.png)
.avif)
