The challenge
Two automation phases for a financial professional: phase one turns a daily news email into a private audio podcast. Phase two built a full research engine that reads source articles and produces a branded PDF newsletter and a two-voice conversational podcast on any topic. No manual work required.
How we built it
Phase one ingests daily newsletter emails, uses ElevenLabs to generate a private audio briefing, and distributes it via Transistor.fm. Phase two built a full research engine that reads source articles in full, runs a three-section analysis with de-duplication, and outputs a branded two-page PDF newsletter and a two-voice conversational podcast, all controlled by a simple Google Sheets config layer so new topics can be launched without touching the code.
How the system works
The actual AI architecture we designed and shipped for Jeff Williams.
Specifically, what we built
- Daily Gmail intake to ElevenLabs audio via Transistor.fm podcast
- 3-section research model with full-article reading and de-duplication
- Branded 2-page PDF newsletter and 2-voice podcast generation
- Google Sheets config layer to launch new topics without a rebuild
The measured impact
- Reduced manual briefing prep by an estimated 80-90%
- 118 hours of AI automation delivered across 2 phases
- 4 reusable outputs: podcast, transcript, newsletter, and audio
Key result
80-90%
less manual briefing prep
Build something like this for your business
Book a free 30-minute AI audit and we will scope exactly what is possible for your specific workflows.
Client
Jeff Williams
Financial Services