Agentic UX / Product Strategy
Campaign Create Agent
A high-efficiency AI agent designed to transform complex ad campaign workflows into a transparent, conversational experience for industry experts. By bridging the "trust gap" through real-time visibility and manual overrides, the project has become Madhive’s top strategic priority for 2026, with the MVP's success leading to my appointment as the design lead for the entire AI vertical.
Company
Madhive
Duration
2 Months ~
Industry
AdTech / Enterprise AI
Scope of work
AI Agent Design
Conversational UX
Product Strategy
Problem
Ad campaign creation is a complex, high-friction process that requires manual extraction of targeting parameters from disparate documents. Users often experience "onboarding fatigue" due to a lack of context, leading to high drop-off rates when transitioning through critical stages like file uploading and campaign-to-line-item conversion.
Challenge
The primary challenge was designing a conversational AI workflow for a demographic of industry experts who, despite finding manual campaign creation tedious, harbor a higher degree of skepticism toward AI automation than younger users. Bridging this "trust gap" required architecting a transparent system that manages complex state transitions without stripping away professional control. I had to engineer a logic that handles unsaved data and intermediate processing states while ensuring every AI action, from targeting extraction to line-item conversion, remains verifiable and reversible to win the confidence of veteran power users.
Solution
I designed a "Transparency-First" agentic workflow that prioritizes user agency over "black-box" automation. When a user uploads complex documents like a Company IO, the system enters a real-time extraction state where users can visualize the AI’s work-in-progress as it populates the form. Crucially, I implemented a "Dual-Mode" interaction mechanism: users can observe the live process or pause the AI at any moment to manually edit and override specific fields. This hybrid approach allows the AI to handle the heavy lifting while keeping the expert user firmly in the driver’s seat, fostering trust through immediate feedback and flexible intervention.
Impact
While currently in the implementation phase, the project has already generated significant momentum. I presented the initial user flows and high-fidelity prototypes to our key enterprise clients, receiving overwhelmingly positive feedback that validated the "Human-in-the-Loop" design strategy. As a result, the AI integration initiative has been designated Madhive’s top strategic priority for 2026. Following the exceptional reception of the MVP, I have been appointed to lead the design and strategic direction for the entire AI vertical, overseeing its evolution as we move toward the official launch. This project remains a cornerstone of Madhive's enterprise expansion, directly aligning with our mission to drive scalability and long-term revenue growth.
Due to NDA restrictions with Madhive, only a limited portion of my work is publicly available. If you're interested in learning more, please feel free to contact me directly.
From MVP Acceleration to Future-Proofing
I architected a modular design strategy for the Campaign Create Agent, developing distinct design sets tailored to each phase of the product lifecycle, from the initial MVP (Minimum Viable Product) to the Final Production Version. By collaborating closely with the engineering and product teams, I prioritized core functional requirements to accelerate the launch of our first MVP. This phased approach allowed us to gather immediate user feedback while I simultaneously authored the architectural foundations for subsequent iterations. This ensured that our initial release was both stable and fast-to-market, while the high-fidelity final designs remained fully prepared for a seamless transition into advanced features in late 2026.
Design System Expansion - AI-Specific Components & Cross-Platform Theming
In response to the new AI vertical, I engineered a specialized suite of components, including advanced input fields and processing states, and integrated them directly into our Kit Design System. I architected these new assets to be fully compatible with our existing Multi-Theme infrastructure, ensuring they can seamlessly toggle between the DSP and MAV visual identities. This expansion not only supports the unique functional requirements of the AI Agent but also maintains absolute consistency across the entire Madhive ecosystem, allowing for a 1:1 design-to-code workflow in any platform environment.
COLLABORATIVE ITERATION & LIVE CODE REVIEWS
Following the initial design phase, I led live-coding sessions and design reviews with the front-end engineering team to bridge the gap between vision and technical execution. By inspecting code and adjusting design specifications in real-time, we collectively evaluated every feature based on our aggressive launch timeline. This deep cross-functional collaboration allowed us to strictly define the MVP (Minimum Viable Product) scope, prioritizing essential AI interactions for immediate release while strategically moving complex secondary features into subsequent development phases. This agile approach ensured a high-quality, high-performance launch that stayed on schedule without compromising the long-term architectural vision.