Customer Service Automation · Marketing Personalization
yappyBuy needed an enterprise-grade AI stack for its Buddy shopping assistant so it could personalize search, product support, and checkout interactions while meeting European data-privacy and governance requirements.
Reported impact
yappyBuy accelerated development of its Buddy assistant, gained a governance- and GDPR-aware AI foundation for open-beta rollout, and built a platform it expects to use for personalized search, support, upsell, and checkout optimization across B2B and B2C commerce clients.
yappyBuy built its Buddy e-commerce assistant platform on the IBM watsonx portfolio to support personalized shopping journeys across online stores. IBM watsonx.ai serves as the enterprise AI studio and model runtime, where the team began with IBM Granite models and Meta Llama 3.1 for conversational and agentic flows. IBM Watson Discovery, watsonx.data, and watsonx.governance extend that core stack so Buddy can ground responses with retrieval-augmented generation and govern how the assistant is used as the startup moves from pilot to open beta. The platform ingests PDFs, JSON files, CSV files, Word documents, databases, and dynamic API responses to provide contextual answers, product guidance, and support during browsing and checkout. IBM also helped yappyBuy tune latency and cost, which matters for real-time commerce interactions, while preserving control over data-storage and processing locations needed for GDPR-sensitive European deployments.
Merchant documents and data sources such as PDFs, JSON, CSV, Word files, databases, and live APIs are fed into the Buddy platform, where watsonx.ai uses Granite or Llama models with retrieval-augmented generation to answer shopper questions, guide product discovery, and support checkout interactions with grounded responses.
Currently the primary IBM watsonx products of our platform include watsonx.ai, IBM Watson Discovery, watsonx.data and watsonx.governance.
We started off with the IBM Granite models together with Meta Llama 3.1, the next generation of Meta’s open LLM, for our conversational tool with agentic flow capabilities.
parts of our conversational agentic workflows now include extensive use of the IBM Granite models that decrease latency and optimize the user experience.
we analyzed over 1 million online comments
we can now successfully incorporate documents, such as PDFs, JSON files, CSV files, [Microsoft] Word docs and databases, or dynamic API responses to provide contextual information
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
OpenAI Sample Aug 3, 2026 Circles Circles needed to turn rich telco signals such as usage, billing, network activity, and location into real-time support and monetization actions instead of reactive, siloed customer journeys. 22% increase in ARPU in Singapore from AI-driven personalization AI Concierge · CareX · Codex 
Google Cloud Sample Est. · Jan 28, 2026 Sabre Travel buyers were stuck with a 30-year-old 'tyranny of the five boxes' search paradigm while content sat scattered across many sources, and travel sellers (e.g. airlines) were absorbing tens of millions of customer… $150M in annualized cost savings from cloud infrastructure modernization Gemini · Google Maps Platform · Travel Impact Model API OpenAI Sample Jan 26, 2026 Indeed Indeed needed to expand AI across job search, employer hiring workflows, and internal operations while keeping human decision-makers in the loop and building employee confidence in how to use the technology well. About 70% of sponsored applications come from AI-powered recommendations Career Scout · OpenAI APIs · Smart Sourcing 
