one.com - AI Product Innovation

Role

I’m a Product Innovation Intern at one.com, currently serving as the Product Owner for the AI Guild. I represent our Dubai AI team to product leadership in Copenhagen, align AI initiatives with product strategy, and orchestrate fast PoCs that de‑risk ideas and earn a path on the roadmap.

How I work

  • Open ideation, on a cadence: I host recurring brainstorms where engineers, designers and PMs bring ideas. We document everything, cluster themes, and compare against our backlog and company priorities.
  • Quick research to reduce uncertainty: For promising ideas, we run lightweight market scans and competitor teardowns, then define success signals and risks.
  • Quarterly mini‑roadmap: We maintain a Guild‑level plan and sync it with the product team so discovery and delivery stay connected.
  • UX first, fidelity later: I start with flows and wireframes, plus short user studies to validate the core value—not pixel perfection.
  • Vibe code PoCs: In parallel I build clickable web PoCs using the same stack our devs would ship with. Stakeholders get something they can try; engineers get components and logic they can reuse.
  • Iterate, decide, move: We instrument demos, collect feedback, and either graduate the idea, loop once more, or archive it.

What I worked on

Below are representative initiatives I led or co‑led. Each started as a simple narrative, became wireframes, and shipped as a PoC that stakeholders could use.

Storefront Chatbot (PoC)

  • Initiated the wireframing and design process, shaping the conversational experience and user flows.
  • Collaborated on building the RAG (Retrieval-Augmented Generation) system, mapping user conversations to relevant website builder categories to guide users toward the right website solutions.
  • Led prompt engineering efforts to optimize chatbot responses and ensure accurate, context-aware interactions.

AI Image Improvement

  • Built a user-facing interface that leverages our AI image enhancement models, specifically tailored to benefit online shop customers.
  • Enabled users to submit simple prompts, which are processed by the Nano Banana model to generate improved product images.
  • Added a one-click background removal feature to streamline product image editing.
  • Designed this as a foundational step toward integrating AI-powered image editing across our broader product suite—including the website builder, bookings, and blogs.

Section Refresh

  • Developed a proof-of-concept for AI-powered section redesign using the LangChain agentic framework.
  • The system takes an existing website section and generates a refreshed design, leveraging AI to propose new layouts and content structure.
  • Serves as a foundational model for an AI website builder, where each section is treated as an individual, composable component.
  • Demonstrated how agentic workflows can automate UX improvements and accelerate iteration cycles for product teams.

Operating model

  • Stakeholders: AI Guild (Dubai), Product (Copenhagen), domain teams across builder/commerce/support.
  • Tooling: Figma (flows/wireframes), Notion (docs/decisions), Vercel/Cloud (PoC deploys), OpenAI/vision and text models, TypeScript/Node, Astro/Next, Tailwind.
  • Signals I track: time‑to‑first‑demo, stakeholder comprehension, technical feasibility, and the clarity of the path to production.

Why this works

Shipping real, clickable PoCs compresses the distance between idea and understanding. PMs and leaders can evaluate the value fast, and engineers inherit working slices that shorten time‑to‑production.