Decagon (@DecagonAI) / X

Decagon

526 posts

Decagon

@DecagonAI

Powering concierge customer experiences for the most impactful companies in the world. Backed by @a16z, @accel, @baincapVC, @coatuemgmt, @indexventures

San Francisco

Joined January 2024

Pinned
Decagon
@DecagonAI
Jun 9
Introducing Duet Autopilot, the first verified self-improving AI agent for CX. It automates agent improvement by turning conversation signals into validated improvements ready for human review, helping agents get better with every conversation. 🧵
00:00
15K
Decagon reposted
MTS
@MTSlive
Jun 18
.@AshwinSreenivas, co-founder of @DecagonAI, on why cheap AI creates more demand for customer support: "This kind of goes back to the lump of labor fallacy. There's not a fixed amount of software to be done in the world, and if I can do it more efficiently, then I need less
00:00
4.6K
Decagon reposted
MTS
@MTSlive
Jun 18
Why can't you just wrap a frontier AI model in customer support and call it done? @AshwinSreenivas, co-founder of @DecagonAI: "Yes, there's some work you need to do for the model, but there's a lot of things that need to happen around that." "If I have a 99% accurate model
00:00
5.2K
Decagon reposted
Aishwarya Goel (Ash)
@aishwarya_08
Jun 16
We went deeper on this today. Keep tweeting 😀 @baseten 💚 @DecagonAI
Cyrus
@cyrusasg
Jun 1
I get asked a lot about what actually matters in the inference space. The conversation has shifted as OSS frameworks have closed much of the gap on raw latency, but workload-specific tuning remains an open problem. Increasingly, more differentiation lives in the product layer
510
Decagon
@DecagonAI
Jun 16
🤝
Together AI
@togethercompute
Jun 16
Article
How Decagon Engineered Sub-Second Voice AI with Together AI
The challenge with running voice agents Voice latency is audible. Decagon’s leadership frames voice as the most demanding surface because latency is immediately perceptible. Long pauses create awkward...
3.1K
Decagon reposted
Together AI
@togethercompute
Jun 16
.@DecagonAI cut voice agent cost per turn nearly 6x with Together AI. They moved from closed models to fine-tuned open models, while keeping latency low enough for real-time voice: → <400ms p95 model latency per turn → custom speculators and prompt caching → optimized
Together AI
@togethercompute
Jun 16
Article
How Decagon Engineered Sub-Second Voice AI with Together AI
The challenge with running voice agents Voice latency is audible. Decagon’s leadership frames voice as the most demanding surface because latency is immediately perceptible. Long pauses create awkward...
5.1K
Decagon reposted
Jesse Zhang
@thejessezhang
Jun 15
Article
How to compete in “crowded” AI markets
Generative AI, over the last few years, has presented one of the greatest opportunities in history to create new products and new companies to advance humanity. It’s no wonder that so many talented...
33K
Decagon
@DecagonAI
Jun 11
Two design teams shipping AI that people use every day. We're hosting a design night with @harvey during Config week. Register below!
713
Decagon
@DecagonAI
Jun 11
Harvey x Decagon Design Night · Luma
From luma.com
568
Decagon
@DecagonAI
Jun 11
We're proud to welcome @CorePowerYoga, the largest yoga studio brand in the US, to the Decagon customer family. Across 220+ studios in 23 states, CorePower makes yoga challenging, accessible, and worth coming back to for millions of students.
372
Decagon reposted
Ashwin Sreenivas
@AshwinSreenivas
Jun 10
One of the hardest parts of building self-improving agents is proving they are actually improving. That’s why, alongside Duet Autopilot, we built DuetBench: the first benchmark designed specifically for CX agents that learn and improve over time. To evaluate Duet Autopilot, we
532
Decagon
@DecagonAI
Jun 10
Yesterday, we introduced Duet Autopilot as the first self-improving agent for CX. It’s a big claim, which is why we decided to build a benchmark to back it up. DuetBench is the first benchmark designed specifically for CX agents that learn and improve over time.
871
Decagon
@DecagonAI
Jun 10
Replying to @DecagonAI
As self-improving systems become more common, verified evaluation will matter just as much as model capability.
147
Decagon
@DecagonAI
Jun 10
Full research and methodology:
DuetBench: An evaluation of self-improving customer service agents
From decagon.ai
146
Decagon reposted
Daniel C. Liem
@damndanielliem
Jun 9
Our AI at @DecagonAI is now doing more agent-building than we (humans) are. Duet wrote 81% of our test simulations, and made 54% of the edits to our customers' agents. AI products naturally evolve toward AI doing more of the work, and we're pushing that to the frontier
Decagon
@DecagonAI
Jun 9
Introducing Duet Autopilot, the first verified self-improving AI agent for CX. It automates agent improvement by turning conversation signals into validated improvements ready for human review, helping agents get better with every conversation. 🧵
00:00
4K
Decagon reposted
Kimberly Tan
@kimberlywtan
Jun 9
.@DecagonAI's ability to push the frontier of agent development is truly impressive. Duet Autopilot is the first verified self-improving AI agent for CX – learn more below!
Jesse Zhang
@thejessezhang
Jun 9
Today, we’re launching Duet Autopilot, the first verified self-improving AI agent for CX! It automatically analyzes conversations, identifies opportunities for improvement, validates updates, and surfaces them for human review, improving itself with each cycle. 👇
00:00
1.4K