1M-3MProduct launchExplainerSeries AB2BGlobalUSVertical AI
TLVC Rating
4.20
Hook
Editing / Creativity
Copy
Sentiment of launch
Distribution strategy
Community Rating
No ratings yet
to rate this launch.
Giga's Scout is a system for enterprise support teams that takes a target KPI, such as funded deposits from new bank customers, and works to move that number on its own. It builds agents around the goal, watches live conversations (especially the ones where a human rep had to take over), spots recurring failure patterns, and proposes specific fixes like wiring up an API so the agent can issue a promo code inside a chat. Small changes to wording or policy are tested on a slice of traffic and shipped automatically, while anything touching real money or external systems is routed to a named owner with the evidence Scout collected, then resumed once unblocked.
The launch matters because most "AI agent" products still leave the optimization loop to humans, who manually review transcripts, write prompts, and guess at what to change next. Scout closes that loop by tying agent behavior directly to a business metric the customer defines in plain English, with a command center that shows what was tried, what worked, and what is waiting on approval. It is aimed at operators running large support functions who care less about deflection rates and more about downstream revenue outcomes.
Giga is a San Francisco company founded in 2024 by Esha Dinne and Varun Vummadi , both graduates of IIT Kharagpur , with Vummadi as CEO and Dinne as CTO. The company builds AI support agents for enterprises and recently raised $61 million to expand its enterprise voice AI work, with DoorDash among its customers , which gives Scout an existing deployment base to learn from as it rolls out.
Comments (10)
Sign in to join the discussion.
Y
Yael BrennerJun 29, 2026
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?
N
Naveed ShaikhJun 29, 2026
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.
T
Tomasz KowalJun 29, 2026
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.
P
Priya VenkateshJun 29, 2026
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.
M
Marcus AdeyemiJun 29, 2026
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.
I
Ingrid SolbergJun 29, 2026
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.
C
Chewa MwaleJun 29, 2026
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.
B
Brigitte LaurentJun 29, 2026
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.
K
kenjiJun 29, 2026
tweet copy is unusually clean for an AI launch, no 'meet the future' nonsense, just verb verb verb. respect.
F
Femi AdesanyaJun 29, 2026
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.
Work With Us
Book an intro call to explore our suite of services. We might have a waitlist.
Giga (Scout)
AI agents for enterprise support
AI agents for enterprise support
Tags
TLVC Rating
Community Rating
No ratings yet
to rate this launch.
Giga's Scout is a system for enterprise support teams that takes a target KPI, such as funded deposits from new bank customers, and works to move that number on its own. It builds agents around the goal, watches live conversations (especially the ones where a human rep had to take over), spots recurring failure patterns, and proposes specific fixes like wiring up an API so the agent can issue a promo code inside a chat. Small changes to wording or policy are tested on a slice of traffic and shipped automatically, while anything touching real money or external systems is routed to a named owner with the evidence Scout collected, then resumed once unblocked. The launch matters because most "AI agent" products still leave the optimization loop to humans, who manually review transcripts, write prompts, and guess at what to change next. Scout closes that loop by tying agent behavior directly to a business metric the customer defines in plain English, with a command center that shows what was tried, what worked, and what is waiting on approval. It is aimed at operators running large support functions who care less about deflection rates and more about downstream revenue outcomes. Giga is a San Francisco company founded in 2024 by Esha Dinne and Varun Vummadi , both graduates of IIT Kharagpur , with Vummadi as CEO and Dinne as CTO. The company builds AI support agents for enterprises and recently raised $61 million to expand its enterprise voice AI work, with DoorDash among its customers , which gives Scout an existing deployment base to learn from as it rolls out.
Comments (10)
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.