Giga (Scout)
AI agents for enterprise support
About
ok wait, an agent that optimizes for funded deposits sounds amazing until it starts cold DMing my grandma about high-yield savings. Guardrails doc when?
hot take: 'self improving agent' is the new 'self driving car', everyone demos it on a clean highway and nobody shows the four way stop. show me a messy bank ticket queue.
The cut from KPI dashboard to live agent convo at 0:14 is genuinely well edited. Whoever cooked that landing video deserves a long weekend.
How does Scout know its 'improvement' isn't just nudging customers into deposits they regret next month? Curious where the goal function ends and the customer interest begins.
Pricing on outcome-based KPIs is the most interesting bet here. If you charge per funded deposit lift you eat a lot of attribution arguments, if you charge per seat you leave money on the table.
For regulated FI customers, where does the conversational data live and who signs the BAA equivalent when Scout 'learns from every conversation'? Asking for every compliance team that will email you Monday.
Is there an API to read the agent's current 'hypothesis' for why a change moved the KPI? Would be a killer webhook for our experimentation tool.
Sorry can someone explain why this is different from a really good Zendesk macro plus a dashboard. I'm still warming up to chatbots in general.
tweet copy is unusually clean for an AI launch, no 'meet the future' nonsense, just verb verb verb. respect.
Counterpoint: enterprise support is shrinking as a budget line because every CFO is told AI will collapse it. You may be selling into a category that's actively defunding itself.