How Arca runs day-long financial agent workflows in an AI-native wealth firm
Latent Space · youtube · 2026-11-01
Aman Singh, Arca's head of AI, joins Latent Space at AI Engineer New York to explain how Arca is building an AI-native wealth management firm.
Key points:
- Architecture: custom harness plus persistent memory; agentic workflows spanning days, not hours; models chosen via internal benchmarks with custom task routing
- Human-in-the-loop: a $100K tax bill case shows agent proposals with human approvals; trade execution stays with human experts
- Safety and compliance: prompt injection defenses, sandboxes, financial plan verification, auditability by default
- Org methodology: staff feedback becomes evals; operational knowledge becomes agent context; "agents prompt people more than people prompt agents"
- Client examples: surfacing hidden fees, proactive service
- Singh argues owning agent infrastructure is what separates AI-native firms from firms that merely use AI tools
More from coding & agent
- Open-Source CodeRail Lets Your Whole Team Share One Claude Code Session From Any Device — L_i_n_k0666 · 2026-11-01
- Lazy Astra Agent Fix: Set /goal First to Make It Finish Jobs — msg · 2026-11-01
- NVIDIA's OpenShell Ditches Sandboxes for Zero-Privilege Agent Security — Hesamation · 2026-11-01
- AI agent logged into her video system unprompted to verify a fix — alliekmiller · 2026-11-01
- CI cost showdown: same 20K minutes — GitHub $120 vs Blacksmith $80 vs Celesto $13 — aniketmaurya · 2026-11-01
- Celesto launches CI runners that undercut GitHub Actions by ~90% with a one-line change — aniketmaurya · 2026-11-01