RiverHacks | RiverHacks

The ask: Make every venue and menu look local

Heineken’s trade-marketing teams dream of posting bar promos that feel shot on-site but generic pack-shots or AI mock-ups still look fake. Farm Frites faces the same headache: one hero recipe must be re-shot and re-written for each market, draining time and budget.

Together, these brands need visuals that are authentically photographed, instantly localised, and culturally spot-on without booking another photoshoot or translator.

The challenge:

Can your team build a “shoot-once, localise-forever” creator that slashes photoshoot costs and translation churn?

Use ChatGPT Pro, Vision prompts, and Manus actions to scrape venue images, place branded overlays, and auto-translate CTAs so marketers can pump out outlet-perfect, culture-perfect visuals before their coffee goes cold.

Client challenge: Valuation velocity

Design an AI application that automates essential investment manager tasks, such as market analysis and risk assessment, to deliver data-driven decision support. The solution should streamline workflows by leveraging AI to enhance financial screening and valuation.

The winning solution:

The team prototyped an AI solution automating key investment tasks, from identifying strategy-aligned opportunities and conducting comprehensive due diligence, to refining their investment focus. They utilized AI tools like ChatGPT for research, Lovable for prototyping and Gamma for their presentation, showing AI's ability to boost efficiency and precision.

Client challenge: The fashion co-pilot

Develop an AI assistant that empowers retail associates to instantly recognize loyalty members, access their purchase history, and deliver personalized recommendations in real time to enhance customer experience, boost sales, and foster loyalty.

The winning solution:

The winning team used AI to rapidly research, wireframe, prototype, and present a working proof of concept to address the 42% of customers who leave stores without purchasing. The solution focused on providing customer history to retail assistants, using facial recognition (or QR codes, depending on local privacy laws), to enable personalized in-store experiences. They used AI tools like Claude.ai for research, v0.dev for wireframing, Cursor for prototyping, and Beautiful.ai for presenting.

The ask: Kill the chaos, empower the store

AS Watson’s store teams need speed but right now, getting answers takes too long. When staff ask: “Where do I find that again?”- the answer is buried in old manuals, outdated SharePoint folders, or half-remembered instructions. It slows down service, creates risk, and frustrates everyone.

The challenge:

Can your team build a smart manual-search assistant that feels like asking a co-worker but works even faster?

Using tools like ChatGPT, Manus, Bolt, or Power Automate, you’ll combine clever prompting and retrieval techniques into a single chatbot that helps store staff move fast, stay compliant, and serve customers better.