An AI agent can answer every question correctly and still feel completely tone-deaf.
Today, we're introducing Sentiment in Tango, Simple's conversational awareness model. It makes Simple the first voice agent on the market to process sentiment directly from audio.
For more: usesimple.ai/blog/tango-sent…
After the call, every conversation becomes a record of how the customer felt. You can see where it turned at a glance, and identify patterns to improve your customer experience.
Tango keeps the beat, but now it reads the room, too. 💃🏻 For the full breakdown, with audio examples: usesimple.ai/blog/tango-sent…
Most end-of-turn models are trained and benchmarked on clean, high-bandwidth audio. Real customer calls are compressed 8kHz telephony. Tango is dual-channel and telephony-native from the ground up, built for the calls you actually get.
Tango is live across Simple's production traffic today, bringing more natural conversations to the calls our customers handle every day. Full breakdown and demos: usesimple.ai/blog/tango-voic…
On LiveKit's public benchmark, Tango is the fastest of 11 systems once you require 90%+ detection. It’s also the most accurate, missing only 15 of 400 true turn endings, versus 54 for Soniox and 197 for Deepgram Flux.
It also knows the difference between a real interruption and a quick "mm-hmm." The two look identical in a transcript, but call for opposite behavior.
The result: 64% fewer interruptions than our previous agent in production.
Most voice AI waits on a transcript before it decides anything. Tango reads end-of-turn directly from streaming audio.
In production, it decides to act a median 300ms before a standard voice-activity detector even registers a pause.
Voice AI's biggest challenge is timing. Today we’re excited to introduce Tango, Simple's conversational awareness model.
Tango is industry-leading on LiveKit's public end-of-turn benchmark, catching 96% of turn endings, responding in ~800ms, and barging in 5% of the time.
For more: usesimple.ai/blog/tango-voic…
With Tango, we're able to predict the end of a turn a few hundred milliseconds before it actually happens.
At a fixed interruption rate, Tango detects end of turn faster and more reliably than any other systems we tested (benchmarked using LiveKit's public EOT Bench).
Tango is already live for Simple’s customers, delivering lower latency, higher accuracy, and more intuitive conversations across industries. We’re ready to keep raising the bar.
Read more about Simple Tango at usesimple.ai/blog/tango-voic….
Hello Vegas! We’re in town this week and we bring gifts. 👀 🎁
The Simple Wishline is live! Look out for our trucks on the Strip. Whether you're here for Genesys Xperience, or just enjoying the summer. When you spot one, simply call the number and tell Wendy the Wishline Agent your gift.
We’re granting wishes all week. Some asked for tickets to the Sphere. Some, that fancy buffet. What do you wish for?
Eight days from contract sign to AI taking live customer calls. And that was only the beginning.
A year into the deployment at Omaha Steaks:
- 60% of all calls are completely handled by AI, 70%+ on sales
- Call abandonment down from 16% to 3%
- Seasonal hiring plans cut from 5,000 to 692
- 36% higher upsell rates than the seasonal hires it replaced
At Customer Contact Week Las Vegas, our CEO @catheryn_li sat down with Grant Young and Rob Bradshaw of Omaha Steaks to walk through the whole story, including the details vendor case studies usually skip: the phone tree assumption that turned out to be wrong, the broken add-to-cart button that Simple flagged in the middle of peak season, and what happened to the QA team after the AI started helping.
The full 45-minute conversation is now on YouTube: youtube.com/watch?v=ZCmXLwly…
