AI Leads
Turning a B2B marketplace into an intelligence platform, and building the pricing model to monetize it.
Impact
Overview
At the start of 2025, I led the end-to-end design of AI Leads, a capability that fundamentally changed what Matium was as a product. Before this, Matium's value was built around network access. Users had to search, filter, and message to find relevant opportunities. As the network scaled, it became clear that our data was the real differentiator, not access to it, but what we could surface from it.
Working directly with the Head of Product and Lead AI Engineer, I defined not just the interaction design but the product strategy: where AI-generated insights should live, how they should be gated to drive conversion, and how design could serve as the lever in a pricing model shift. Free members continued accessing Core. Paid Intelligence members unlocked full AI Leads. Free users saw a limited preview with blurred cards, creating a visible, tangible reason to upgrade.
The result was both a new feature and a new business model.
Problem
Matium provided a growing network for plastics and supply chain professionals, but finding the right buyers or sellers was still slow and manual. Users searched, filtered, and messaged repeatedly to uncover opportunities. This friction reduced the platform's perceived value and made it harder to justify upgrading from free to paid. We needed a way to surface the right opportunities at the right time, directly within the workflows where buying and selling decisions were already being made.
Solution
I partnered with the AI engineering and product team to design AI Leads, a proactive, data-driven feature embedded into the Directory, Haves, and Needs pages. AI Leads used intelligence from platform activity to automatically surface the most relevant trading opportunities for each user. Free members received a preview of five cards with one intentionally blurred, making the value of upgrading immediately legible. Paid members gained full access. Discovery went from a manual task to an integrated, conversion-optimized experience.
Surface Relevant Opportunities
Deliver AI-driven leads directly on the pages where users were already making decisions, so they never had to search for value. By embedding intelligence into core workflows, we reduced friction and accelerated the moment of insight.
Scale Trust Through Transparency
AI Leads were designed to feel clear, consistent, and easy to scan. Users needed to understand the recommendations they were acting on. Transparency in how leads were generated built confidence in the product and reinforced Matium's credibility as a data-first platform.
Drive Conversion Through Visible Value
The gated preview was a deliberate design decision, not a limitation. By showing free users exactly what they were missing, we created a conversion path grounded in product value rather than marketing pressure.
Over Q1 and Q2 of 2025, the AI engineering team built the data pipelines needed to make AI-powered insights possible across the platform. Working closely with them and the Head of Product, I explored how and where these leads could deliver the most value.
We identified two key integration points:
1. Directory Integration: Embedding AI leads directly into the directory so users could see high-value trading partners while browsing the network.
2. Haves and Needs Detail Pages: Surfacing AI-generated matches on individual listings, tailored to whether the user was buying or selling.
This discovery work shaped both the product strategy and the final experience, ensuring AI Leads felt like a natural extension of existing workflows rather than a feature added on top of them.
Map AI Data to User Workflows
Working with AI engineers, I mapped where predictive lead data could best serve users in context. The goal was to surface opportunities exactly where buying and selling decisions were happening, reducing friction and maximizing the relevance of each recommendation.
Integrate Across the Platform
AI Leads were embedded into the Directory, Haves, and Needs pages so opportunities appeared in the natural flow of user activity rather than in a separate or isolated experience.
Create Event Tracking to Measure Impact
I worked with the team to instrument key interactions with Statsig, capturing engagement patterns and upgrade behavior across AI Leads. This let us directly link feature usage to revenue impact and prioritize future iterations based on real conversion data.
AI Leads were introduced across three high-impact areas of the platform. In the Directory, they revealed new companies as potential trading partners. On Haves pages, sellers saw suggested buyers for their available inventory. On Needs pages, buyers were matched with relevant sellers. By embedding intelligence into these workflows, I ensured that actionable opportunities were surfaced exactly where decisions were made.
When AI Leads launched, it became the catalyst for rethinking Matium's entire pricing structure. Core, which had previously been the paid tier, shifted to free in order to expand adoption. Matium began charging not for access to basic functionality, but for the value of its intelligence.
Paid Intelligence members got full access from day one. Free members saw a page of five AI Lead cards with one blurred, a deliberate design choice to make the upgrade proposition tangible without requiring any marketing copy to explain it.
AI Leads did not just improve the product experience. It became a growth driver for Matium's business model.
Since launching in August 2025, AI Leads has changed how buyers and sellers on Matium discover opportunities. Embedding AI-driven insights directly into core workflows removed the friction of manual search and introduced a new layer of value for paid users.
What went well
The collaboration between design, product, and AI engineering was tight and efficient. With strong technical foundations already in place, we focused on crafting an experience that felt intuitive and native to the platform rather than bolted on.
What could have been improved
Under tight deadlines, we could not always guarantee the accuracy of the data powering AI Leads. Some contact information occasionally proved misleading, which limited the feature's full potential until data quality was refined in later iterations.