Transforming AI-generated insights into a decision-driven system that reduced manual effort and improved operational speed
Project Snapshot
Company / Product: Evoke — Poker Ops, tournament optimization workflow (iterating an Excel-based tool)
Timeline: Oct 2025 – Dec 2025
Role: Senior UI/UX Designer
Stakeholders: Poker Ops + Business stakeholders
Platform / Tools: Excel-based ops tool (iteration-focused approach)
Scope (What I designed): prioritization aligned to weekly cadence, decision context, safe bulk actions with guardrails, integrated auditability/traceability
Outcomes (pilot): overlay -15–20%, profit/tournament +8–10%, weekly time -50–60%
Confidentiality: Pilot results are rounded / scoped.
Summary
This project focused on improving how tournament optimization decisions were made based on AI-generated insights.
Before this work, the system produced recommendations through a periodically generated spreadsheet. Users had to manually search for tournaments, interpret the data, and then apply changes in a separate platform. This process was time-consuming, error-prone, and highly dependent on individual interpretation.
I redesigned this workflow by transforming the spreadsheet-based output into a structured product experience. The new system automatically matched AI insights with real tournaments and allowed users to review, apply, or reject recommendations directly within the platform.
This shifted the workflow from manual interpretation to fast, decision-driven actions, while also creating a feedback loop to continuously improve the AI model.
Context
The AI system was already generating valuable insights, but the way those insights were consumed created friction.
Users had to move between tools, manually reconcile data, and decide what actions to take without a clear or structured interface. This introduced delays, inconsistencies, and limited the overall impact of the AI.
My Role
I led the UX/UI design, focusing on transforming a data-heavy, spreadsheet-based workflow into a usable and scalable product experience.
A key part of my work was understanding how users interpreted the data, what decisions they needed to make, and where errors or inefficiencies occurred.
From this, I designed a system that connected insights directly to real entities (tournaments) and supported fast, confident decision-making.
Problem
The core issue was not the quality of the AI insights, but how they were delivered and used.
Users had to manually:
- find relevant tournaments
- interpret AI-generated data
- apply changes in a separate system
This created high cognitive load, increased the risk of errors, and slowed down decision-making.
Approach
Instead of improving the spreadsheet, I focused on redesigning the entire interaction model.
The goal was to eliminate manual reconciliation and turn insights into direct actions.
This involved:
- designing a decision interface instead of a data interface
- connecting AI output to real tournament data
- structuring insights in a clear and actionable format
Key Product Improvements
The workflow was transformed from manual interpretation into a structured decision system.
AI-generated insights were automatically matched with real tournaments, removing the need for users to manually search and reconcile data.
Users could review each recommendation directly in context and either apply it instantly or reject it, providing a reason for their decision.
This created a clear decision flow while also generating structured feedback that could be used to improve the AI model over time.
The system reduced context switching, lowered cognitive load, and significantly improved the speed and consistency of decision-making.ile supporting operational confidence.
Approved Layout
Final layout designed and iterated by me, reviewed and approved jointly with Poker Ops and Business stakeholders.
Outcomes
- Reduced manual effort in tournament optimization workflows
- Faster decision-making through direct action interface
- Lower risk of errors caused by manual interpretation
- Creation of a feedback loop to improve AI recommendations over time
What This Demonstrates
This project demonstrates my ability to design products that bridge the gap between AI systems and real user workflows.
It shows how I transform raw data and insights into actionable experiences, reducing complexity while improving speed, accuracy, and scalability.