AI Design: Unification of systems on a Multitenant and Whitelabel platform
Executive Summary
One of the country's largest Industry Federations faced a fragmented management scenario across two distinct ecosystems: a legacy proprietary system and a third-party SaaS system with a restricted architecture.
The coexistence of these solutions generated high costs, data fragmentation, compliance risks, and governance difficulties. The operational complexity of the proprietary system also led users to create parallel controls in spreadsheets, expanding the Shadow IT scenario.
To overcome this scenario, the organization sought Squadra with the objective of building a single, Multitenant, and Whitelabel platform capable of unifying the systems without replicating existing problems.
The main challenge was to understand the architecture and business rules of the SaaS, a true "black box," without access to the source code, API documentation, or database diagrams.
With the Genius platform, Squadra applied artificial intelligence to the Product Discovery and Design process, using multimodal analysis, reverse engineering, prototyping, and continuous validation via RAG based on a rigorously cleaned and validated knowledge base. .
The process transformed months of research, prototyping, and architectural validation into weeks, accelerating decisions, reducing rework, and creating a structured foundation for the development of the new platform.
The Company
One of the country's largest Industry Federations, with operations supported by different technological ecosystems for managing its activities and serving associated industries.
The Scenario
The organization maintained its management distributed across two systems: a proprietary one, which presented technical limitations and operational challenges, and a third-party SaaS, whose closed architecture restricted flexibility and customization capabilities.
This fragmentation impacted data governance and led to the use of parallel controls in spreadsheets to enable operational activities. Simultaneously, the closed nature of the SaaS ecosystem imposed additional technical challenges to unification, requiring a robust approach to risk mitigation.
The project therefore aimed at evolving towards a single, Multitenant, and Whitelabel platform.
The Challenge
Technical mapping of a SaaS with a closed architecture
The SaaS did not provide source code, API documentation, or database diagrams. It was necessary to perform reverse engineering to understand its business rules and functionality.
Fragmentation and Shadow IT
The operational complexity of the proprietary system led users to seek spreadsheet alternatives to enable their activities, impacting data governance.
Lack of customization
The SaaS did not allow visual customization, making it difficult to apply the organization's identity and that of the different companies served by the platform.
Risks related to data isolation
The absence of a clearly known Multitenant architecture increased concerns about information isolation between different tenants and potential privacy and competition risks.
Avoiding the reproduction of legacy problems
In addition to understanding the existing systems, it was necessary to ensure that the problems and limitations identified in the current scenario were not transferred to the new platform.
The Solution
Genius in Action
The Genius platform was applied to the Product Discovery and Design process to accelerate system understanding, define the new architecture, and validate decisions.
Genius's orchestrator agents used multimodal analysis to process screen recordings, meeting transcripts, and network logs, enabling the mapping of navigation flows, behaviors, endpoints, APIs, and business rules of the SaaS, as well as supporting the analysis of the legacy system's source code.
In defining the new platform, AI supported the structuring of a Multitenant architecture with logical isolation between different tenants, as well as a Design System based on Design Tokens to enable a Whitelabel experience, with adaptation of colors, logos, and typography.
To reduce reliance on parallel spreadsheets, experiences focused on high data density were also designed, including a Dynamic Data Grid with inline editing and a Drag-and-Drop builder for Self-Service BI.
Another differential was the use of RAG for continuous validation. The information gathered during Discovery was transformed into a knowledge base used by Genius agents to validate new screens, flows, and architectural decisions, helping to prevent problems from the legacy system from being reproduced in the new solution.
Results
Main Deliverables
Reverse engineering of the SaaS
Mapping of business rules, APIs, flows, relationships, fields, and records of the closed system, even without access to the source code.
Navigable and functional prototype
High-fidelity interfaces and navigable prototypes that simulated real use and allowed continuous validation with the client.
Multitenant and Whitelabel Architecture
Definition of an architecture prepared for multiple clients, with isolation between tenants and the possibility of customizing the visual identity.
Design System
Creation of a structure based on Design Tokens, allowing the adaptation of the platform's visual elements for different clients.
Solutions for high data density
Design of Dynamic Data Grid and Drag-and-Drop builder for Self-Service BI, increasing user autonomy.
Documentation package for Engineering
Generation of documentation with Story Map, Strategic Roadmap, User Stories, business rules, use cases, BDD acceptance criteria, flows, and architecture.
Results and Impact
With the support of Genius's artificial intelligence agents, the Discovery, UX Research, and Architecture process—a stage that would normally take months—was conducted in weeks, with up to 4x faster co-creation and an overall acceleration of 80% compared to the traditional cycle. This was possible thanks to the continuous reuse of knowledge mapped throughout the project, which reduced the time spent searching for already gathered information by 95%.
90% reduction in prototyping time: high-fidelity and navigable interfaces began to be produced in a much shorter cycle, allowing continuous validation with the client at each stage of the Design.
70% visibility into a closed system: even without access to source code, documentation, or diagrams, AI-assisted reverse engineering mapped business rules, APIs, flows, relationships, fields, and records of the legacy SaaS.
Automated handover to Engineering: the process concluded with the generation of a complete documentation package—Story Map, Strategic Roadmap, User Stories, business rules, BDD acceptance criteria, and architecture—supporting the transition to the development stage.
Want results like these?
Talk to our experts and find out how AI can transform your business.
