Case study

Gledva — AI Travel Companion

An intelligent travel workspace that keeps trip context and assists the traveler throughout the journey.

Role
Founder & Product Manager
Product
B2C AI travel platform
Platforms
iOS · Android · Web
Stage
0→1 → Launch → Iteration
Visit the product

The challenge

Travelers manage trips across fragmented systems — booking confirmations, itineraries, documents, calendars, packing lists, budgets, maps and travel applications. I identified an opportunity to build an intelligent travel workspace that maintains trip context and assists the traveler throughout the journey.

What I owned

Product strategy → user experience → requirements → AI architecture → engineering prioritization → launch → GTM → analytics → iteration.

Capabilities I defined and prioritized

AI itinerary planningshippedTrip timelineshippedBooking managementshippedTravel documentsshippedTasksshippedPackingshippedBudget trackingshippedSmart remindersshippedWeathershippedMapsshippedNearby discoveryshippedLanguage assistanceshippedTravel HubshippedGledva AI chatshippedAgentic orchestrationdesignedRAG retrieval layerdesigned

From AI chat to agentic travel

Moving beyond a standalone chatbot toward an architecture that understands context and supports multi-step workflows.

  1. User intent
  2. Context + trip state
  3. LLM reasoning
  4. Orchestration
  5. RAG / retrieval
  6. Tools + APIs
  7. Action
  8. Memory / state
  9. Guardrails + human approval
  10. Evaluation + observability

Traveler: “It looks like rain tomorrow. Reorganize my day.”

  1. 01

    Understand

    Identify destination, date, user intent and the requested change.

  2. 02

    Retrieve

    Pull tomorrow's itinerary, bookings, activity details and relevant traveler preferences.

  3. 03

    Use tools

    Obtain relevant external information such as weather and activity data through appropriate services.

  4. 04

    Reason

    Determine which activities are weather-sensitive and identify alternatives.

  5. 05

    Generate plan

    Create a revised itinerary considering timing, location and existing commitments.

  6. 06

    Ask for approval

    Present consequential itinerary changes before modifying confirmed plans.

  7. 07

    Execute

    Update approved itinerary elements, tasks or reminders through deterministic application functions.

  8. 08

    Maintain state

    Persist the updated trip so subsequent AI interactions use the latest information.

The product decision

Not the question

“How do I add AI to a travel app?”

The real question

“Which decisions should AI make, which actions should deterministic software execute, and when should the traveler remain in control?”

Designing AI autonomy

I designed the experience around levels of risk.

Low risk

Recommend nearby restaurants.

AI recommends autonomously.

Medium risk

Reorganize tomorrow's itinerary.

Generate proposed changes → request confirmation → apply changes.

High risk

Cancel a hotel or purchase a new booking.

Never execute from inferred intent → require explicit authorization.

Product principle: the higher the consequence and the lower the reversibility, the stronger the human approval requirement.

AI product architecture

Shipped capabilities and designed architecture, labelled honestly.

Reasoning

shipped

LLMs interpret natural-language requests and reason across ambiguous travel requirements.

Retrieval

designed

Traveler-specific information is retrieved from structured trip data and documents instead of relying on model knowledge.

Tool calling

designed

External and deterministic capabilities are exposed as controlled tools for the AI workflow.

State

shipped

Structured information for destinations, dates, activities, bookings, documents, tasks, expenses, packing and reminders.

Guardrails

designed

High-consequence actions require stronger controls and explicit user approval.

Failure recovery

designed

When reliable information cannot be retrieved, the product fails safely instead of allowing the model to fabricate an answer.

AI evaluation framework

I evaluate AI experiences across the complete workflow rather than only the generated text.

DimensionProduct question
Task successDid the AI accomplish the user's objective?
GroundednessIs the response supported by trusted information?
Tool accuracyDid the agent select and call the correct tool?
SafetyDid it stay within permitted actions?
LatencyWas the experience responsive enough?
CostWhat did successful task completion cost?
User satisfactionDid the customer find the result valuable?

Optimization objective: Quality × Reliability × Latency × Cost

The goal isn't to use the largest or most expensive model. The goal is to select the architecture that meets the required product experience.

Results

100+
Active users
acquired in the first quarter
3
Platforms
iOS · Android · Web
0→1
Concept to launch
product taken end to end
10+
Capabilities
unified in one travel workspace

Key learning

Building an AI product isn't simply adding a chatbot. The deeper product challenge is determining where probabilistic reasoning creates value, where deterministic software should remain in control, how the system should access trusted information, and how much autonomy customers should give AI.