Accuracy at launch, with an instrumented path to 95%+ through continuous learning
Of dispersed technical expertise centralized into a single queryable knowledge base
Multi-interface platform with 4 data pipelines and 14 agent tools
.png)
A long-standing industrial distributor had built its competitive advantage on deep technical expertise accumulated over three decades. That expertise was distributed across PDFs of spec sheets, threaded email archives, ERP records, and the working memory of senior staff. Sales reps could identify the right part, price, and configuration, but only by knowing where in the business to look and whom to ask.
The result was a quoting process defined by manual lookup, inconsistent answers, and uneven turnaround. The exposure was structural. Experienced employees were transitioning out of the business, and institutional knowledge was leaving with them. New reps faced a long ramp. Customers expected faster, more accurate responses. The technical content the company depended on existed, but it was not modeled or indexed in a form any system could query with precision.
OneSix engineered Oracle, an AI-powered quoting and knowledge platform that consolidates the client's documents, emails, and ERP records into a unified, queryable knowledge base. Four data pipelines ingest and structure content from sources that had previously lived only in static files or in the working memory of senior staff. Decades of dispersed technical context are now modeled and addressable in real time.
On top of that foundation, Oracle runs a multi-agent system of fourteen purpose-built tools, combining OCR and large language models to handle intelligent part identification, configuration recommendation, and citation-backed responses. Reps and downstream systems access the platform through chat, API, or a purpose-built UI. Every answer is accompanied by its source, which is the precondition for use in front of a customer.
The platform launched at ~71% accuracy with a defined path to 95%+. Feedback and validation pipelines measure performance continuously, and the system improves with each interaction. What was a manual, expertise-dependent quoting process is now a production-grade, evidence-backed workflow that scales as senior staff transition out of the business.
More to explore
OneSix implemented a data warehouse solution for a commercial printing client, transforming IoT sensor data into real-time, actionable insights that enhanced operational visibility and reduced downtime.
OneSix developed a scalable horse workout tracking application using advanced computer vision, multi-camera arrays, and ultra-wideband tracking to monitor key performance metrics and detect early signs of injury.
OneSix implemented a scalable data warehouse solution for a digital health company, enabling efficient, accurate reporting and data-driven insights into member engagement across texting programs.