Ada Interact 2025 wasn’t about predicting the future of AI customer service —it was about actively building it, together.

On September 16, AI & CX leaders stepped away from theoretical talk to roll up their sleeves, exchange practical insights, and tackle the toughest challenges head-on. With an agenda built for action and community-driven learning, Interact showcased not only what’s next in AI customer service , but what’s already working for industry leaders like Pinterest, monday.com , Zapier, Ancestry, and more.

It became clear: the best insights often came from the leader sitting next to you.

Through meaningful peer-to-peer conversations, collaborative breakout sessions, and intimate roundtables, attendees formed lasting connections and gained firsthand experience solving real-world problems.

If you couldn’t join us—or simply want to relive the experience—here’s your inside look at the boldest ideas, sharpest takeaways, and most impactful strategies from Ada Interact 2025.

AI-native from day one: The enterprise shift starts now

Ada Interact opened strong with a keynote from CEO Mike Murchison, emphasizing just how critical and challenging it is for enterprises to operationalize AI at scale.

But this challenge isn’t insurmountable. Mike highlighted groundbreaking results from Ada’s AI agents, now surpassing human performance in customer service.

To help more enterprises reach this tipping point, Mike introduced Ada ACX, Ada’s operating model that combines technology, proven playbooks, and expert guidance to deliver concierge customer experiences at scale.

The goal? Equip enterprises not just to pilot AI, but to systematically scale it.

Murchison framed the challenge: most enterprises stall out in pilots. CPTO Mike Gozzo followed up with the solution. Here’s how he broke it down.

transforming AI from experiment to enterprise reality

In his product keynote, Gozzo cut through the AI hype with a focus on what enterprises actually need: the tools to move beyond pilots and scale AI with confidence.

He introduced four major enhancements to the ACX platform, all aimed at solving the enterprise’s biggest AI challenges: reducing risk before deployment, scaling automation into core workflows, and giving teams more control to coach their AI agents to higher performance.

• Playbooks: Turning rigid scripts into dynamic workflows that handle complex, multi-step tasks. Instead of weeks of brittle scripting, leaders can now describe a process in plain language, upload a PDF, or coach the AI directly to resolve nuanced cases like refunds or compliance-sensitive requests without human intervention.

• Instant AI Testing: Replacing spreadsheets and manual QA with one-click simulations. Enterprises can validate integrations, catch edge cases, and debug with full transparency into the AI’s reasoning before a single customer sees it, eliminating risk and accelerating deployment.
• MCP Integrations: A new universal integration standard that lets Ada plug into the entire enterprise stack. From updating a CRM to triggering an incident workflow in Jira, MCP turns Ada from a standalone layer into an embedded teammate across systems.

• Developer Platform: Expanded APIs and interfaces that give technical teams the freedom to extend Ada with custom logic, channels, and message types—without needing deep ML expertise.

Together, these capabilities cement Ada’s shift from being a technology layer to becoming a trusted, integrated teammate inside the enterprise ecosystem.

Ada isn’t just announcing features in isolation. Enterprises are already putting these ideas into practice. Leaders from Pinterest, monday.com, Zapier, and Ancestry shared how they’re deploying AI at scale—and what others can learn from their journeys.

real-world enterprise AI: lessons from leaders

Some of the most compelling moments at Ada Interact 2025 came from enterprise leaders who’ve moved AI from experiment to full scale deployment. Their experiences provided powerful blueprints for navigating the complexities of deploying AI at scale.

Pinterest: scaling trust and safety for 500M+ users

Pinterest’s VP of Support Operations, Lisa Stoner, shared insights into how Pinterest meticulously built trust and safety into their AI customer service.

Recognizing that serving over 500 million users required exceptional diligence, Pinterest conducted a rigorous AI readiness audit that encompassed over 300 safety evaluations, prioritizing brand-specific safety concerns around viral trends and sensitive topics.

Pinterest’s exhaustive safety audit shows what’s possible when AI adoption is built on trust —a scalable deployment that protects brand integrity while unlocking faster rollout across teams and regions.

Key takeaway: Enterprise-grade AI isn’t just about advanced technology. It's about rigorous safety and compliance frameworks that build user trust and mitigate risk from the outset.

monday.com: AI as a cultural shift, not a technology bolt-on

Gal Ohayon, CX Product Manager, and Ziv Shechtman, CX Systems Ops Lead at monday.com illustrated how their global team adopted AI as a strategic, cultural transformation—not just another technology deployment.

By starting from customer pain points and aligning AI strategies around measurable outcomes, monday.com rapidly restructured teams, ran internal hackathons, and adopted a relentless "deploy, measure, refine" approach to embed AI into every customer interaction.

Key takeaway: Sustainable AI success requires shifting organizational culture, empowering cross-functional teams, and embedding AI within clearly defined processes—not simply adding new technology.

Ancestry and Zapier: aligning teams and processes for generative AI success

Joe Wang, Director of Customer & Community Experience at Ancestry, and Carissa Frear, Director of Global Support at Zapier provided practical guidance on aligning organizational readiness with AI initiatives. Both emphasized the critical importance of preparing frontline teams with the right skills, adapting operational processes to manage generative AI effectively, and transparently navigating cultural change to ensure successful adoption at scale.

Key takeaway: Effective generative AI deployment depends on people readiness, adaptive processes, and clear alignment across the entire organization, not just technology infrastructure alone.

While the mainstage delivered proven strategies and examples of deployments, attendees broke out in smaller groups to hammer out practically how to scale A. In workshops and roundtables, leaders got candid about the obstacles they’re facing and the practical strategies that are moving the needle.

real talk, actionable strategies: expert-led roundtables

The heart of Ada Interact wasn’t in the presentations or product announcements. It was in the small, strategic discussions where CX leaders and AI managers connected directly, tackling real challenges in real-time.

Across workshops and roundtables, four clear themes emerged:

• AI quality at scale: Leaders agreed that as AI takes on more conversations, the challenge shifts from containment to consistency. And that improving QA and coaching practices directly drives CSAT, loyalty, and revenue growth at scale. New approaches to QA, measurement, and coaching are becoming critical to maintaining standards across millions of interactions.
• From reactive to proactive: Teams are no longer satisfied with AI that only resolves tickets. The focus is shifting to AI agents that anticipate customer needs, surface insights, and drive engagement before issues arise.
• Making the business case: Securing executive buy-in is still a top hurdle. Leaders shared strategies for framing AI investments in terms of revenue impact, customer loyalty, and operational resilience—not just cost savings.
• The shift to generative: Customers who’ve moved from declarative to generative AI shared what it really took to make the leap: overcoming internal resistance, restructuring teams, and redefining the role of human agents once AI handles the majority of conversations.

breakout sessions: hands-on learning, immediate results

The breakout sessions brought CX leaders and AI managers into smaller groups to tackle the practical side of scaling AI. These weren’t lectures—they were working sessions where attendees compared approaches, shared obstacles, and traded strategies for what’s actually working in their organizations.

The discussions revealed what it really takes to operationalize AI-first customer service inside large enterprises:

• Control is king: AI managers shared how having the ability to guide and refine their agents in real time is critical for scaling without sacrificing quality. Playbooks, analytics, and coaching tools are becoming everyday levers of performance.
• Composability unlocks scale: Enterprises are no longer content with siloed AI. They’re looking to plug their agents directly into core systems—CRM, billing, logistics—so automation doesn’t just resolve questions but triggers real actions.
• Voice is back on the table: Far from legacy IVR, leaders are revisiting voice as a channel where AI can drive instant value—reducing wait times, cutting drop-off, and modernizing the phone experience customers still rely on.
• The path to generative: For many customers, the question isn’t if but how fast they can make the shift to generative AI. Breakout discussions unpacked the organizational, technical, and cultural shifts required to adopt generative AI responsibly.

your next step: turning insights into action

Ada Interact 2025 wasn’t just another conference—it was an accelerator, equipping CX leaders and AI managers with the strategies, tools, and connections needed to confidently scale AI across their enterprises.

From product announcements showing how Ada is raising the bar for enterprise AI, to candid lessons from leaders already deploying it, attendees left with frameworks and strategies they could put to work immediately.

Now it’s your turn. Ready to turn ambitious AI visions into measurable, enterprise-grade realities?