Case Study

Endesa SOI

Live URLPrivate enterprise pilot

Client: Endesa x Garaje de Ideas

Endesa SOI is an AI-powered research operating system built to centralize institutional knowledge, structure incoming research requests, and give teams instant access to validated insights through a conversational interface. By turning scattered documents and manual workflows into a governed research system, the platform helps accelerate decision-making without compromising rigor or traceability.

Next.jsReact.jsTypescriptSupabaseShadcn UIVercel AI SDKVercel AI ElementsRetrieval Augmented Generation

Endesa SOI platform dashboard with research workspace and knowledge retrieval controls

Endesa SOI documents repository view with indexed research files and filtering options

Endesa SOI conversational chat interface retrieving answers from internal research sources

Endesa SOI briefs dashboard showing structured research requests and status tracking

01 / Problem

The core blocker we had to solve

01 ContextResearch knowledge was there, but unreachable

The team was generating valuable research, but the system around that knowledge was fragmented. Past studies were buried across PDFs, folders, and disconnected assets, making it difficult to retrieve validated findings when they were needed. At the same time, new research requests frequently arrived with incomplete briefs, forcing long clarification loops before work could even begin. What should have been a strategic knowledge asset was instead creating friction, delay, and duplication.

02 ConstraintsEnterprise-grade rigor shaped the solution

This was not a generic AI assistant problem. The system had to work in an enterprise environment where trust, permissions, governance, and source traceability were non-negotiable. It needed to support multiple stakeholders, respect how research was already commissioned and delivered, and remain flexible enough to fit compliant infrastructure choices and future deployment inside Endesa's corporate environment.

03 ImpactThe workflow was slowing decisions down

The hidden cost of the existing workflow showed up everywhere: slower decisions, duplicated effort, preventable back-and-forth, and underused research. Teams were spending days clarifying study requirements, manually processing knowledge, and re-discovering insights that already existed. That reduced the visibility and leverage of Research at the very moment the business needed faster, evidence-based decision-making.

02 / Solution

From fragmented research to a single intelligence layer

We turned research into an operating system

Instead of solving for document retrieval alone, we defined the product as a full operating layer for research. The concept combined three core capabilities into one conversational experience: a brief builder that structures incoming requests, a knowledge center that makes validated past studies searchable in natural language, and synthetic personas that allow teams to explore hypotheses before launching new research. This shifted the value proposition from finding files faster to making research continuously usable.

We simplified access without simplifying the method

The product experience was intentionally designed around a single conversational entry point. Behind that simple interaction sat a more sophisticated system for knowledge retrieval, structured intake, and specialized agent behaviors. This made the platform approachable for business stakeholders while preserving the methodological standards, source discipline, and governance the research team needed.

We designed the MVP for immediate leverage

The implementation roadmap balanced immediate utility with long-term extensibility. The work covered workflow mapping, tooling audit, AI opportunity design, MVP definition, implementation, real-use-case testing, and knowledge transfer. That gave the team a practical first version capable of demonstrating value quickly, while also establishing the foundation for a broader research infrastructure that could evolve over time.

03 / Results

Impact delivered

Endesa SOI transformed research from a fragmented, manually intensive workflow into a faster and more accessible operating system for insight. By structuring intake, centralizing validated knowledge, and enabling conversational access to past research, the platform reduced operational friction for the research team while giving stakeholders a more direct path to trustworthy answers. The result was not just efficiency gain, but a stronger strategic role for Research inside the organization.

95%

Reduction in document search and knowledge processing time

95%

Reduction in research brief compiling time

24/7

Access to validated research knowledge through a conversational interface

+ Related Services

Relevant services for this outcome

RAG Pipeline Development Services | Retrieval-Augmented Generation

We build production-grade RAG pipelines on your documents, knowledge bases, and proprietary data. Semantic search, grounded AI responses, deployed in 4 weeks.

Know More

AI Process Automation Services

We replace manual, repetitive business workflows with AI-powered automations that run 24/7. Integrated into your stack, deployed in 4 weeks, from EUR8,500.

Know More

AI Agent Development Services

We build autonomous AI agents that qualify leads, handle support, and execute workflows without human oversight. Production-ready in 4 weeks. From EUR8,500.

Know More

+ Other Case Studies

Explore related projects

MVP Copilot

ADVANCED MVP PLANNING ASSISTANT

Next.jsReact.jsTypescriptSupabaseShadcn UIVercel AI SDKVercel AI Elements

View Case Study

Ancient Remedies

AI POWERED NATURAL REMEDIES APP

ExpoReact NativeTypescriptSupabase

View Case Study

Ready to ship your AI system?

Book a call. We will scope it in 30 minutes. What we can build, how long it takes, and what it costs. No pitch deck. No fluff.

Book a Call