We’re at the edge of a shift. One that’s quietly—and radically—transforming how businesses show up for their customers.

That’s why we recently introduced what’s next for Ada at our Spring Product Launch , turning AI from tool to teammate. With new capabilities that help AI learn from every interaction, improve continuously, and deliver human-level conversations, Ada isn’t just helping teams scale support—we’re redefining what great customer experience can look like.

And the biggest leap forward? Voice .

Voice is notoriously difficult to automate. It's a costly, messy channel full of edge cases, but it can also be the most exciting one if you’re able to crack the code. We’re proud to announce that Ada’s generative voice AI is finally getting it right.

Alongside Ada’s product leadership, we were joined by Ismail Ostrilski, Head of Operations at Epos Now, who shared how his team is scaling AI across channels to deliver fast, consistent, and surprisingly human customer service. And yes—there were live demos. You can catch the full thing in the on-demand replay .

Here’s what we unveiled, and what it means for the future of AI in customer service .

AI that shows up where it counts

The future of AI support doesn’t live in the background—it leads from the front. That was the message from our Chief Product Officer, Mike Gozzo, who opened the event by reframing the role of AI in customer service.

This isn’t about task automation or deflection metrics anymore. It’s about building AI agents that are intelligent enough to become your brand’s most consistent, capable, and cost-effective team member, everywhere your customers are

This bold vision is already taking shape. Ada now powers consistent interactions across chat, email, and voice. And every time you coach or optimize your AI agent , the improvement cascades across channels, giving you a single layer of intelligence that gets smarter with every interaction.

voice, re-engineered for dynamic, human-like conversations

Voice support has long been a headache—expensive to staff, frustrating to scale, and easy to get wrong. Many businesses have deprioritized it entirely, hoping customers will default to chat or email instead. But as Shorit Ghosh, GM of Voice at Ada, made clear: voice isn’t going anywhere.

Ada’s second-generation voice AI doesn’t just hear customers, but truly listens, reasons, and responds like a human would.

Ada’s voice AI was re-engineered from the ground up to:

• Speak like a human
• Learn like a human
• Think like a human
• Collaborate like a human

built-in brilliance: ada’s newest AI tools for scale and precision

One of the biggest misconceptions about AI agents is that they’re set-it-and-forget-it tools. In reality, the best AI improves over time, and Ada’s latest product capabilities are designed to make that growth seamless. With new tools that help you coach, guide, and scale your agent's performance, you can now drive smarter resolutions with less effort.

These aren’t just feature drops. They’re force multipliers. Here’s how each one helps your AI agent go from good to exceptional.

Recommendations

Your AI agent now learns from your human team. When a question gets escalated, Ada can analyze how your agents resolved it—and automatically draft new help content to fill the gap. You can approve, tweak, and publish in a few clicks, turning human insight into scalable automation without starting from scratch.

Scorecards

Want to know how your AI is really doing? Scorecards let you measure performance based on your own QA standards—things like empathy, accuracy, and resolution quality. Instead of relying solely on CSAT, you can now track how well your agent is living up to your brand and service standards at scale.

Coaching

Just like your best agents get feedback, so can Ada. Leave coaching instructions directly on past interactions to fine-tune tone, add context, or adjust answers. The next time a similar question comes up, your AI agent will apply the improvement automatically.

Playbooks

Playbooks are simple instructions that teach your AI agent to follow your standard operating procedures, just like human agents do. Instead of scripting every path, you can use natural language to guide how your AI handles multi-step tasks, ensuring they follow your policies while still adapting to real-world scenarios. It’s a fast, flexible way to automate the complex, without compromising control.

Epos Now: scaling smarter, not harder

When your company powers over 80,000 small businesses globally, customer experience isn’t just an afterthought—it’s your foundation. That’s the case for Epos Now , an operating system for small and medium-sized businesses.

Their customer base spans baristas, retailers, and entrepreneurs—people with their hands full, both literally and figuratively. And as Head of Operations Ismail Ostrilski shared, supporting them across fragmented, siloed systems just wasn’t cutting it.

So they looked for a partner focused on more than just automation and could deliver smarter service through a single, customer service automation platform. Ada stood out.

They started with messaging. And within weeks, saw an immediate 30% lift in CSAT. But they didn’t stop there. They took a big leap into voice—and the results were just as clear:

• 15% increase in CSAT on voice since launch
• 40-50% containment on voice interactions
• 60-70% automated resolution across all channels
• 60,000 hours of support labor saved annually

And when a handoff is needed? The AI pre-populates cases, pulls in conversation context, and sets up human agents for faster resolutions.

For Epos Now, it’s not just about ticking boxes or closing tickets. It’s about understanding the customer, their business, and what they need next—even if they haven’t asked for it yet.

That kind of support doesn’t feel transactional. It feels personal. And when done right, it opens the door to better retention, smarter upsells, and long-term trust.

the AI agent era isn’t coming, it’s already here

Ada’s Spring Product Launch wasn’t just a moment—it was a milestone. We saw what’s possible when AI doesn’t just react to problems but anticipates needs. When it doesn’t just handle conversations but owns outcomes. When it doesn’t just scale, it scales with quality.

That’s the benchmark now. Not automation for its own sake, but automation that earns trust, reflects your brand, and moves with the intelligence of your best human agents.

And as the demos and results made clear: that benchmark isn’t just aspirational. It’s here. It’s working. And it’s only getting better.