AI-enhanced Enterprise Search Experience
This case presents the enhancement of The World Bank’s enterprise search as the first entry layer into a broader AI ecosystem, designed to introduce users to AI through their most frequently used tools. Built on existing search infrastructure using Google Vertex AI, it transforms keyword-based retrieval into AI-generated, grounded summaries with full source traceability. The experience also serves as a gateway into conversational AI via an option to continue the query in an AI Chat interface.
Product position at the World Bank's AI Ecosystem
Problem Statement
- Inefficient Keyword-Based SearchExisting enterprise search returned long, unstructured result lists, requiring users to manually scan, interpret, and synthesize relevant information.
- Disconnected from AI ChatLack of continuity between enterprise search and AI chat, limiting seamless transition from information discovery to deeper conversational exploration.
- Time-Critical Knowledge Access NeedsSlow access to critical information affecting time-sensitive decision-making workflows.
- Limited Reliance on Internal KnowledgeInconsistent reliance on external sources due to fragmented access to a unified internal knowledge base.
Previous Solution Overview
The World Bank’s enterprise search served as a centralized access point to institutional knowledge accumulated across decades of projects, research, documents, internal repositories, and organizational systems.
The underlying search infrastructure was built on Google Cloud, enabling large-scale indexing and retrieval across a broad network of internal databases and content sources.
AI-enhanced Product Vision
- Answer-First ExperienceEnterprise search evolves from document discovery to direct, structured answers, enabling employees to receive synthesized insights immediately after query submission.
- Time-Optimized Knowledge FlowTime-efficient knowledge workflows enabling faster access to actionable insights and improved decision-making speed.
- AI Search-to-AI Chat ContinuityAI Search is seamlessly connected with AI Chat, enabling users to continue exploring queries in natural language, ask follow-ups, and extend understanding within a unified AI ecosystem.
- Trusted First Point of SearchEnterprise search becomes the primary access point for World Bank internal knowledge, replacing users’ reliance on external search tools through AI-generated contextual summaries that reduce effort and increase adoption.
High-Level Timeline
High-Level Product Logic. Before vs After AI Integration
AI-enhanced Product
The updated experience was designed to preserve familiarity for existing users while gradually introducing AI capabilities into the established enterprise search workflow, maintaining a largely consistent interface to avoid disrupting frequent usage patterns.
At the same time, the product was reintroduced as Enterprise AI Search, highlighting the integration of Google Vertex AI as the underlying technology powering AI-driven capabilities.
Main Entry Point
2. Accuracy disclaimers were incorporated directly into the entry experience, encouraging users to validate AI-generated information before using it.
3. To support different levels of user familiarity with AI, the interface also introduced suggested prompts as a secondary entry point, allowing users to quickly select pre-defined queries and immediately navigate to relevant results.
AI-Generated Overview Layer (Projects)
The updated experience introduced an AI-generated overview layer positioned above traditional search results, providing users with a synthesized summary of the query alongside key data, related entities, and clickable references connected to internal knowledge sources. An integrated Ask AI Chat action allowed users to seamlessly continue the query in a conversational AI environment for deeper exploration and follow-up interaction.
5. Traditional keyword-based search results remained accessible below the AI overview layer, preserving continuity for existing enterprise search workflows while supporting deeper manual exploration when needed.
6. An integrated Ask AI Chat action allowed users to seamlessly continue the query in a conversational AI environment for deeper exploration and follow-up interaction.
AI Search Results (Person)
8. Accuracy disclaimers were incorporated directly into the entry experience, encouraging users to validate AI-generated information before using it.
9. To support different levels of user familiarity with AI, the interface also introduced suggested prompts as a secondary entry point, allowing users to quickly select pre-defined queries and immediately navigate to relevant results.
AI Search Results (WiFi)
2. Accuracy disclaimers were incorporated directly into the entry experience, encouraging users to validate AI-generated information before using it.
3. To support different levels of user familiarity with AI, the interface also introduced suggested prompts as a secondary entry point, allowing users to quickly select pre-defined queries and immediately navigate to relevant results.
Key Learnings
- 1. PromptingUsers experienced difficulties structuring prompts, which led to inconsistent output quality.
→ This led to the introduction of a prompt library designed to help users quickly select and reuse effective prompts instead of writing them from scratch. - 2. Onboarding / WalkthroughFirst-time users were often unclear about product capabilities and interaction patterns.
→ A guided onboarding experience was designed to help users quickly understand key functionalities and usage flow. - 3. File & Image InputThere was a strong expectation for the ability to upload and analyze files and images directly within the chat.
→ This led to prioritization of multimodal input capabilities to support document- and image-based workflows. - 4. Conversation HistoryUsers faced challenges in locating and reusing previous conversations and outputs.
→ This led to a redesign of conversation history into a more structured and searchable knowledge memory layer. - 6. Knowledge FilteringUsers needed the ability to narrow responses to specific domains and internal datasets.
→ This resulted in the introduction of structured knowledge filtering based on organizational context such as projects and datasets. - 7. Feedback LoopAlthough users were willing to provide feedback, the mechanism for doing so was not intuitive or easily accessible.
→ This led to embedding contextual feedback mechanisms directly within the chat experience. - 8. Multiple LLMsUsers expressed interest in using different AI models depending on task type and desired output style.
→ This led to the exploration of a flexible model layer allowing users to switch between different AI providers within the same chat experience. - 9. Translation NeedsUsers frequently worked with documents and workflows in multiple languages, which required constant switching to external translation tools.
→ This led to the consideration of in-chat translation capabilities to reduce reliance on external services.
Delivery Process
The product was developed over a 3 year period using an iterative, Agile-based approach, with continuous validation and close cross-functional collaboration.
1. Problem Framing & Product Vision
The process started with identifying key challenges around secure AI usage, fragmented tools, and limited access to internal knowledge.
These insights informed a clear product vision, defining the shift toward a unified, secure, AI-driven workspace for employees.
2. Roadmap Definition
and Alignment
Based on the product vision, a clear and transparent roadmap was established, outlining key milestones, priorities, and expected evolution of the product.
The roadmap was accessible across teams, enabling shared understanding of:
- Product direction
- Scope and priorities
- Team involvement and timelines.
3. Cross-functional Discovery
Early-stage exploration involved close collaboration across product, design, engineering, and architecture.
Work was driven through:
- Collaborative workshops (whiteboards, workshop sessions)
- Alignment discussions with technical architects
- Iterative exploration of possible solutions
- This helped translate complex requirements into feasible product directions.
4. Agile Delivery
via Azure DevOps
The delivery process was structured using MS Azure DevOps, with work broken down into:
- Epics
- User stories
- Tasks
This ensured clear traceability from high-level goals to implementation.
5. Sprint Execution
and Scrum
Development followed an Agile sprint cycle, including:
- Sprint planning
- Backlog grooming
- Daily syncs
- Sprint reviews
- Regular Scrum ceremonies ensured alignment, prioritization, and continuous progress across teams.
6. Iterative Development
The product evolved through structured iterations:
- MVP → validation of core concept
- Iteration 1 → usability and interaction improvements
- Iteration 2 → expansion into a scalable AI workspace
- Each phase built on previous learnings, allowing the product to scale in both functionality and complexity.
7. User Feedback and Validation
Continuous feedback loops were embedded throughout the process via:
- Usability testing sessions
- Internal feedback channels
- Direct user observation
- Insights were used to refine features, improve usability, and guide prioritization.
8. Continuous Alignment & Delivery
Ongoing collaboration across teams ensured alignment between:
- User needs
- Technical constraints
- Business goals
- This enabled consistent delivery of a complex product while maintaining clarity, usability, and strategic direction.
- 80%Reduction in Manual Discovery Time
Significant decrease in time spent manually navigating, filtering, and reviewing information across systems and result pages.
- 75%AI Search-to-Chat Engagement
A significant share of users continued their discovery flow into AI Chat, using follow-up questions and conversational exploration to deepen understanding beyond initial search results.
- 70%Reduction in Time-to-Answer
AI-powered response generation reduced the time required to identify, review, and synthesize information across enterprise-scale knowledge, with grounded answers delivered in an average of 3 seconds.
- 65%Increase in Enterprise Search Reliance
AI-generated grounded answers improved confidence in internal enterprise search, reducing dependency on external search tools and fragmented knowledge discovery.
Role and Recommendations:
Working closely with engineering and data architecture teams, user experience became deeply embedded in product decision-making, from defining interaction models and clarifying feature logic to structuring end-to-end user flows.
This resulted in a contribution that extended beyond execution into product thinking, cross-functional alignment, and active participation in shaping the product through continuous feedback collection and analysis, reflected in strong stakeholder recognition and recommendations. Acknowledged through outstanding endorsements from coworkers and leadership at The World Bank, including the Global CIO.
- Amy Jean Doherty I Global CIO at The World Bank"We appreciate your contributions and leadership!" (LinkedIn comment)
- Onika Vig I AI Solutions Product Owner at The World Bank"... Beyond her skills, Alice brings a positive, solution-oriented attitude that uplifts everyone she works with. I highly recommend Alice — she would be a tremendous asset to any organization looking for creativity, dedication, and impact. I love working with Alice ❤️"
© Fox Alice
