Retail AI Search

Adeptmind

Adeptmind is a Toronto based AI company focused on improving e-commerce search and product discovery. Its platform delivers personalized search results, guided recommendations, and contextual product discovery that bridges the gap between online and in-store shopping experiences, helping retailers increase engagement and conversions. Founded in 2016, Adeptmind serves hundreds of retail clients and focuses on making product search more relevant and intuitive with AI-driven insights.

Background

I was approached by Adeptmind, at a time when most retail search experiences were frustratingly limited and often drove customers back to Google or Amazon. Adeptmind was tackling this problem by building a more intelligent retail search engine that went beyond simple keyword matching. Instead of relying solely on a retailer’s product catalog, their platform crawled and learned from a broad set of web data to better understand how people actually describe and search for products. This allowed the system to interpret intent, modifiers, and context in complex queries, surfacing results that more closely matched what shoppers were looking for. The goal was to help retailers keep high intent customers on their own sites and convert interest into purchases by offering a search experience that felt closer to natural language understanding than traditional ecommerce search.

My Contributions

I initially focused on building internal tools to help train and improve Adeptmind’s search models using a human-in-the-loop approach. These systems allowed agents to review search results and correct relevance in real time, creating high quality labeled data that was fed back into the models to improve future searches. For example, if a query like “ripped blue jeans” returned a mix of ripped jeans and unrelated blue jeans, agents could downrank irrelevant results, helping the system better understand intent and modifiers over time. This work was done in 2018, before many of the recent advances in large scale language models and modern AI tooling.

As the product matured, I proposed shifting focus toward enterprise customers, with an initial emphasis on mall operators.

We partnered with Cadillac Fairview, one of Canada’s largest commercial real estate and mall operators, to define a clear product vision and strategy centered on reducing friction in the in-mall discovery and purchase journey. The goal was to enable shoppers to quickly find specific products across multiple retailers through a single mobile experience.

Following several weeks of collaboration and alignment on scope and outcomes, we signed a substantial contract with Cadillac Fairview. I worked closely with their internal stakeholders, a third party mobile development team, and Adeptmind’s in house search backend team to deliver the mall discovery app, Live.

Key Learnings

• Selling new and unproven technology to small and medium sized businesses involves high adoption friction
• Seed stage funded companies require disciplined prioritization, as runway can diminish quickly
• Be selective with bets, focusing only on problems with clear urgency and leverage
• Deep engagement with mall operators surfaced a genuine, high value pain point, enabling a focused solution grounded in real customer needs