Home services marketing intelligence platform came to Armakuni for aI call and form intelligence: model benchmarking and fine-tuning, data pipeline modernization, GenAI use case assessment, and multi-POC agentic AI commercial intelligence layer.
The challenge
Armakuni won the initial engagement when the company was still Blue Corona, before its merger with Ryno Strategic Solutions. The engagement began with a focused AI call analysis PoC that produced production-quality results fast enough to convince the C-Suite. After a six-month hold, Armakuni maintained communication and re-engaged, growing the account across data platform modernization, model benchmarking, voice intelligence assessment, and four parallel AI POCs covering the full commercial intelligence stack.
Scale: 150+ employees performing manual call analysis daily prior to AI engagement; platform processes tens of thousands of inbound calls across home services clients including HVAC, plumbing, and electrical contractors.
What we built
Armakuni delivered aI call and form intelligence: model benchmarking and fine-tuning, data pipeline modernization, GenAI use case assessment, and multi-POC agentic AI commercial intelligence layer. The engagement ran as multiple fixed-scope SOWs: Data Audit and Modernization PoC (Jan 2024, Nov 2024), Nova Model Benchmarking and Fine-Tuning (Oct 2025), New GenAI Use Case Discovery Assessment (Jun 2025), Prompt Framework POC, Enhancement Detection Engine POC, Forecasting and Budget Optimization POC, Anomaly Detection and Proactive Alerts POC (all Nov 2025). AWS funded the work at $25,000.
The outcomes
5 measurable outcomes shipped across the engagement. The ones that moved the business the most:
Customer ready for production in under two weeks of a four-week PoC.. The engagement that proved Armakuni's capability began with a focused AI call analysis feasibility PoC. Ryno had attempted this internally and failed to reach production quality. Armakuni delivered results convincing enough that Ryno committed to production within the first half…
150-person manual call analysis workflow targeted for AI replacement.. The PoC demonstrated that AI could perform the classification, disposition labeling, and outcome extraction that 150 employees were doing manually each day. The platform built by Armakuni is the technical foundation for eliminating that overhead as capabilities move to…
Nova fine-tuning creates a path to materially lower AI operating cost.. The TCO analysis Armakuni produced gives Ryno a structured Go/No-Go framework for switching from Anthropic Sonnet to a fine-tuned Amazon Nova variant. The smaller fine-tuned model represents a significant reduction in per-call AI inference cost at equivalent accuracy if adopted.
Commercial intelligence layer created where none existed.. The Enhancement Detection Engine, Forecasting and Budget Optimization engine, and Anomaly Detection layer collectively add capabilities the platform did not previously have: identifying missed revenue per call, validating marketing ROI, and detecting performance issues before…
Account maintained through a six-month engagement hold.. Initial projects were paused for approximately six months. Armakuni maintained consistent communication throughout the hold and resumed the engagement when Ryno was ready, demonstrating the partnership depth that generates sustained multi-year account growth.
Built on AWS
Delivered under our AWS Generative AI Services, AI Services, Data Analytics competencies.
What's next
Production deployment of Phase 4 POC capabilities across the platform; potential voice intelligence implementation following the GenAI assessment roadmap; ongoing model fine-tuning and optimization as Nova performance data matures.
