Designing AI search & agent experiences at Algolia
Algolia powers search and discovery for thousands of business-critical applications, processing more than one billion queries every five seconds.
During my time there, I’ve worked on products exploring how search is evolving in the age of AI — from machine-learning powered recommendations to conversational search and agent-driven workflows. This work sits at the intersection of search infrastructure, large language models, and product design, where the role of design is shifting from shaping interfaces to shaping how people interact with intelligent systems.
This page highlights a selection of those projects and how my design process has evolved while designing AI-native products.
Selected work at Algolia
Conversational AI
Ask AI
Evolving a mature documentation keyword search experience (DocSearch) to support natural language interaction, enabling users to ask questions and receive grounded answers powered by LLMs.
Focused on:
• Evolving a 10-year-old search experience without disrupting familiar workflows
• Designing hybrid interactions between keyword search and AI questions
• Shaping both the user interface and configuration for customers implementing the feature
Impact (March '26):
• 166 applications adopted, with 84% retention
• 94,503 search requests
• $529K influenced ARR
Agent platform
Agent studio
Evolving the Agent Studio experience to make it easier for customers to create and launch AI agent search and discovery experiences within the Algolia dashboard.
Focused on:
• Shifted agent creation from code-first to preview-first; making experiences tangible from the start
• Led vision for what agentic experiences we wanted customers to create and how they could create them
• Introduced guided onboarding; reducing time-to-value and improving activation
• Aligned four teams around a unified agent platform; driving cohesion across design, product, and engineering
• Transformed the builder into a scalable system; enabling configuration, triggering, and validation of agent experiences
Product intelligence
Optimization Recommendations
Designing a unified system for delivering feature and optimization recommendations to customers powered by machine learning, helping teams improve their search configurations.
Focused on:
• Unifying fragmented internal insights into a single recommendation framework
• Balancing contextual suggestions with a centralized recommendations hub
• Designing personalized guidance without creating notification fatigue
Impact from pilot (Jan '26):
• Up to 2.2× increase in dashboard engagement (5.4–11.3% vs 5% benchmark).
• Up to 3.8× increase in email engagement (21–28% vs 7.5% benchmark)
Agent capabilities
MCP creation
Designing workflows that enable AI agents to securely connect to Algolia data and tools via Algolia's first Model Context Protocol.
Focused on:
• Creation and management of MCPs for productivity (quick, pre-configured access) and public (configurable, shareable MCPs with tool and permission set-up) use cases.
• Integrating MCP capabilities into the Agent Studio experience
• Connecting agents to tools and datasets
• Defining permissions and capability controls for safe AI behaviour
Impact (March '26):
• 168 applications adopted
• 69,455 search operations
• $95.5M influenced ARR
How my design practice evolved
Working on AI-powered products at Algolia introduced new design challenges, from conversational interfaces to complex configuration workflows. Designing for these systems pushed me to adapt my process; not just in how I explored ideas, but in how I prototyped, collaborated, and worked closer to implementation.
Using AI as a thinking partner 🤔
AI tools became part of my early exploration process. I use them to structure initial thinking, map workflows, and pressure-test ideas before moving into more concrete design work.
Prototyping more realistic workflows 🔀
I began prototyping earlier and closer to the product environment to explore workflows such as Agent Studio and MCP set-up. Using tools like Cursor allowed me to experiment with layouts, hierarchy, and interaction patterns in more realistic contexts than static design tools.
Working closer to implementation 👩💻
I now review pull requests to evaluate UI behaviour and interaction details, and contribute small code changes when needed. This included submitting a PR for an animation in the recommended actions experience.
Exploring longer-term product directions 🔮
Alongside product work, I explored vision concepts for both the dashboard and customer-facing experiences powered by Agent Studio. This helped consider how AI-powered search and agent experiences might evolve as engineering teams build faster with AI-assisted development.