Designing an AI Phone Agent Platform to Power Scalable, Human-Like Customer Conversations
Helping businesses communicate, learn, and improve through AI-driven voice interactions Mawj is a B2B AI platform that enables organizations to deploy realistic AI phone agents for inbound and outbound calls at scale. The platform supports use cases such as market research, sales, reminders, verification, collections, and customer support—while providing deep analytics and automated insights from every conversation. A key competitive advantage is Mawj’s natural Saudi Arabic voice, enabling authentic, culturally aligned customer communication across the MENA region.
Context
Many businesses rely heavily on phone communication, yet traditional call centers are costly, hard to scale, and difficult to optimize.
Existing AI voice solutions lacked trust, visibility, and meaningful insights—offering automation without clear understanding of performance, customer intent, or improvement paths.
Teams needed a way to launch AI-powered calls easily, monitor performance clearly, and continuously improve agent behavior without technical complexity.
My Role | Senior Product Designer — (Part-time)
––––––
The Challenge
Design a platform that:
Makes AI phone agents understandable and trustworthy
Allows non-technical teams to launch and manage campaigns
Turns raw call data into clear, actionable insights
Supports multiple use cases without overwhelming users
Enables continuous improvement of agent performance
The Solution
We redesigned the product around clarity, control, and learning.
Campaign creation was simplified into a guided, prompt-based flow—allowing users to describe what they want in plain language and instantly generate AI call logic.
Analytics shifted from static metrics to insight-driven views, revealing topics, risks, trends, and improvement opportunities automatically.
Each agent gained a clear performance narrative, connecting conversations, outcomes, and recommended fixes in one continuous feedback loop.
The result is an AI platform that feels operational, transparent, and actionable—not experimental.
Outcome
Reduced campaign setup friction through prompt-based configuration
Enabled teams to understand why calls succeed or fail
Introduced continuous improvement loops for AI agents
Established a scalable UX foundation for AI-driven voice operations
Supported thousands of concurrent calls with full visibility
