Solo consultant builds local AI assistant to unify calls, emails, and missed leads

A solo consultant is developing a local-first AI assistant to consolidate client communications across phone calls, messaging apps, and email into a single daily report. The project is motivated by the high cost of missed calls, with industry data indicating small businesses fail to answer 40–60% of incoming calls. The system runs three AI models sequentially on a single server, using Whisper for local call transcription, bge-m3 for semantic search, and Gemma for image processing, keeping all data off external APIs. A key technical challenge involves entity resolution — linking contacts who reach out via different channels without a shared identifier, such as a caller who later follows up by email. The developer is currently designing the report format and contact-matching logic before building the final aggregation layer.