Conversational AI for Workflows
Transformed High-Friction Workflows into Conversational Actions and trained an LLM on them
Problem Statement
The platform is integrating an intelligent conversational assistant to simplify multi-step operational workflows. By allowing users to retrieve data, perform tasks, and navigate dashboards through natural language, it aims to reduce reliance on click-heavy interfaces and make operations more intuitive. This large language model is set to release next month, with the goal of enhancing adoption and usability across warehouse and logistics teams.
Approach & Aim
Hypothesis
A structured persona system, enhanced with AI, can centralize tribal knowledge, surface gaps, and drive more equitable and effective product decisions.
Goal
To move from individual empathy to shared empathy across teams, ensuring personas reflect regional, cultural, and regulatory diversity and prevent costly misalignment caused by “universal personas.”
Scope
In scope: Persona creation, validation, tooling, adoption. Out of scope: Direct usability testing of product features.
Key Insights
8 user personas documented, based on interviews + field visits, enriched with GPT-driven solutions for emerging use cases. Shifted company from fragmented tribal knowledge to shared, documented understanding.
Personas mapped into quadrants showing commonality, autonomy, and edge-case triggers. Introduced contextual disabilities (low signal, no desk, weather barriers).
Focus sessions with TAMs, CSMs, PMs validated personas in design critiques. Personas integrated into GPT tooling for document validation and use-case simulation.
Created Slack bot to distribute persona nudges and structured a persona catalog. Built persona-centric A/B testing framework → increased relevance of experiment learnings.
Impact
Business Outcome
Personas informed product restructuring for warehouse roles. Defined metrics enabled PMs/Analytics teams to prioritize high-impact enhancements. Created foundation for persona-index dashboard and AI persona generator.
Team Outcome
Designers used personas in early-stage critiques. PMs integrated persona insights into PRDs. Analytics aligned test design to persona buckets.
Stakeholder Outcome
Leadership gained visibility into underserved personas → guided roadmap prioritization. CSMs/TAMs used personas to frame customer conversations.