We're sunsetting Ginni AI.

This wasn't easy to write. But it's the right call.

On May 30th, Ginni AI will shut down.

I've been putting off writing this. Not because I didn't know what to say, but because saying it out loud makes it real. Closing a product you've poured two years of your life into, one you genuinely believed in, one that was working, is not easy. It doesn't feel like a clean, rational decision. It feels like letting go of something you love.

And in many ways, it is.

What we built, and what it meant

We started Ginni AI in late 2023 with a clear belief: AI in sales is one of the most underutilised opportunities in enterprise software. Every sales team has a handful of top performers whose instincts and patterns, if properly understood, could be replicated across the whole team. The data is there, hiding in every call, every conversation, every objection handled well or badly.

The vision was to unlock the potential of every salesperson. Not replace them, elevate them. Find the patterns that work, understand why they work, and give managers and reps the tools to scale those insights with data.

What kept us going was seeing it actually work. Watching customers find insights in their conversations they'd never noticed before. Seeing reps improve. Seeing managers finally have the data they needed to coach with confidence. That impact, real, tangible, human, was the fuel. It still is.

The opportunity hasn't gone away. If anything, it's grown. The best version of AI-assisted sales is still waiting to be built, and whoever builds it will have an enormous market waiting for them.

So why are we stopping?

This is the question I've sat with the most.

About 6 months ago, we started building something on the side. Tuner, an observability and reliability layer for Voice AI, began as an exploration and quickly became something we couldn't ignore. The traction was faster. The conversations with developers were electric. And somewhere along the way, almost without realising it, the team had fully shifted.

We got completely hooked on Voice AI.

Here's what I know about us as a team: we don't believe in splitting ourselves thin. We never have. Doing two things at once, at our stage, doesn't mean doing two things. It means doing neither of them properly. And Ginni AI deserved better than half our attention. So did Tuner.

There was also a bigger truth we couldn't ignore. Voice is becoming the new interface. The technology has reached a point where it's genuinely powerful, capable of handling complex, nuanced conversations at scale, and it's going to reshape entire industries. Frontline sales, customer support, cold outreach, the first line of human-to-human calls is already starting to shift. That's not a distant prediction. It's happening now. We could see clearly which side of that transition we wanted to be building on.

We chose to build the infrastructure, not be disrupted by it.

Letting go of Ginni meant letting go of something that had real momentum, real customers, and a mission we still believe in. That's the part that stings. It would have been easier to shut down something that wasn't working. This was working. But we genuinely believe we are onto something bigger with Tuner, something that deserves everything we have. And so we made the call.

Ginni AI didn't just precede Tuner. It built it.

Over two years, we went deep in conversation analysis. Deeper than we expected. We monitored over a million conversations across different languages, different environments, different accents and acoustic conditions. We learned what breaks audio calls. We learned how to navigate background noise, how to work across different speech-to-text and text-to-speech engines, how to build LLM-as-a-judge at scale in a way that's actually reliable in production.

We built QA and AI tooling to monitor human calls and in doing so, we developed an intuition for what makes voice communication fail, and what it takes to make it work consistently. That's not knowledge you can read about. You earn it by being in the trenches with real calls, real customers, real edge cases, at scale.

That experience is the foundation Tuner is built on. Every architectural decision, every insight about where voice pipelines break down, every instinct we have about what developers actually need when their voice agent fails in production, it came from Ginni.

We didn't pivot away from our expertise. We went deeper into it.

Where we are with Tuner

Tuner went live in February. In under three months: hundreds of developers from 33 countries on the platform, native integrations with Vapi, Retell, ElevenLabs, Pipecat, and LiveKit, and an active enterprise pipeline spanning contact centers, voice AI scaleups, financial services, healthcare, and more .

Tuner is the reliability layer for Voice AI, the developer-first platform teams use to run evals, monitor, debug, and analyse their voice agents in production. As Voice AI scales in production, Tuner becomes a critical infrastructure. That's the bet we're making, and everything we've seen so far tells us the timing is right.

To our customers

You trusted us with something important, your sales process, your team's conversations, your time. Seeing what you built with Ginni was what made this hard. You'll receive a separate email with everything you need to export your data before May 30th. If you need anything at all, reply directly and we'll take care of you personally.

I'm rooting for you. I hope you find a tool that carries this mission forward the way it deserves.

Thank you

To our customers, you gave us something no amount of research could: the real thing. A million conversations, in the wild, with all the messiness that comes with it. That was the gift, and we don't take it lightly.

To our investors, thank you for betting on us early, for staying with us through the pivot, and for believing in the team as much as the product. We're not done. We're just getting started.

To our team, past and present, none of this existed without you. Every hard call, every late night, every moment you cared. You didn't just build a product, you built the foundation everything at Tuner stands on. We're stepping into a bigger challenge, and we're doing it because of you.

To everyone who followed, supported, cheered us on, gave feedback, or simply took a chance on what we were building, thank you. You were part of this, and this is part of what comes next.

We're heading into Tuner with more conviction, more experience, and more momentum than we've ever had. We're excited, and we're ready to go bigger.

On to Tuner.

Mai, Co-founder