Building AI-powered puppy training that users trust

Jack explained that scalability was a major challenge for the certified B-Corp startup. Originally, Zigzag provided WhatsApp chat support with real dog trainers, but as the company grew, this wasn’t sustainable. They needed a way to offer science-backed, consistent training advice while still maintaining the personal touch that users valued​.

Zigzag’s AI assistant, Ziggy, is more than just a chatbot—it’s a personalized, interactive puppy coach trained by expert dog trainers. But building an AI feature that users actually engage with wasn’t as simple as flipping a switch.

"The biggest challenge was balancing personalization with consistency," Jack explained. "We needed to make sure AI could deliver training advice that was both scalable and scientifically accurate."

How Zigzag built trust in AI-powered training

To increase AI effectiveness, the team focused on three key areas:

• Expert-validated AI responses – The AI-generated advice was reviewed and refined by professional trainers to ensure accuracy.
• Clear distinction between AI and human trainers – Users know when they’re chatting with Ziggy versus a real trainer, which boosts confidence in the feature.
• Continuous user feedback – Every AI response is rated by users, allowing the team to refine and improve results over time.

"We see a 96% positive rating on our AI responses," Jack shared. "That’s because we track feedback at every step and continuously refine Ziggy’s responses."

“The key to AI adoption is trust. Users need to know they’re getting the right advice—every time.” — Jack Mitchell

Why AI chat features increase engagement

One surprising insight from Zigzag’s research? Many users were hesitant to ask questions in live chat with human trainers.

"Some people felt like their questions were ‘too silly’ to ask a real person," Jack said. "Even though they were paying for a subscription, they didn’t want to bother a trainer with minor concerns."

AI solved this problem by removing the fear of judgment. Now, users feel more comfortable asking quick, bite-sized questions—which keeps them engaged with the app.

The result?

• Higher retention rates – Users stay engaged with the app for longer.
• Increased daily interactions – More users ask questions.
• Stronger trust in the product – AI becomes a gateway to deeper engagement with human trainers.

“AI isn’t replacing human trainers—it’s making training more accessible.” — Jack Mitchell

How Zigzag measures AI’s ROI

One of the biggest challenges in AI adoption is proving its business impact. For Zigzag, tracking AI’s ROI meant running an A/B test to compare user behavior with and without the AI feature.

"We A/B tested the AI version against a version without it," Jack explained. "That allowed us to measure the direct impact on subscription rates and retention."

Key AI success metrics

Zigzag focused on three main KPIs to measure AI effectiveness:

• Reduction in trainer workload – AI helps scale human coaching efforts.
• Subscription conversion rates – AI must drive engagement and renewals.
• User feedback ratings – AI responses must remain high-quality and trusted.

By running controlled A/B testing, the team proved that AI not only reduced human workload but also increased customer lifetime value (LTV).

“If your AI feature doesn’t drive real business value, it’s just a novelty.” — Jack Mitchell

The power of generative AI in product development

Beyond training, Zigzag also uses AI internally to speed up product development.

"We’re a small team, so AI helps us move faster," Jack shared. "From user research to legal checks, AI streamlines processes so we can focus on building a better product."

How Zigzag leverages AI internally:

• User feedback analysis – AI sorts and categorizes qualitative data.
• Content creation – AI helps refine training materials.
• Legal compliance checks – AI ensures content meets regulatory standards.

For small teams, generative AI can be a game-changer, allowing them to iterate faster and bring new ideas to life without needing massive resources.

“AI doesn’t just enhance the product—it accelerates how we build it.”— Jack Mitchell

Key takeaways for businesses integrating AI

For companies looking to implement AI-powered features, Jack offers a few pieces of advice:

1. Start with a real user problem – AI should solve a specific challenge, not just be added for the sake of AI.
2. Keep a human in the loop – AI works best when paired with expert oversight to ensure accuracy and context.
3. Track AI’s ROI – Test, such as A/B,  and measure success using subscription rates, engagement metrics, and customer support calls.
4. Make AI feel personal – Users engage more when AI feels tailored to their experience.

"Every company is thinking about AI right now," Jack said. "But the ones that succeed will be the ones that focus on real user value—not just novelty."

The future of AI-powered puppy training

So what’s next for Zigzag’s AI journey? Jack and his team are continuously refining Ziggy based on user interactions. They’re also exploring new AI-powered features to make dog training even more intuitive and effective.

As AI continues to evolve, Zigzag is proving that AI-powered puppy training isn’t just a trend—it’s the future of pet care.

"We’re learning from every user interaction," Jack said. "The goal is to keep making AI better, smarter, and more useful for puppy parents everywhere."

Episode links: 

• Try ZigZag: Download the app
• Connect with Jack Mitchell on LinkedIn: Find him on LinkedIn
• A comprehensive guide to A/B testing: A comprehensive guide on conducting effective A/B tests to enhance user experience and conversion rates.
• How to use human insight to supercharge your A/B tests: This article discusses integrating qualitative insights to improve the success rate and effectiveness of A/B testing.
• Transforming customer experience with responsible AI strategy: This podcast episode explores how to align AI initiatives with business goals and ethical standards, ensuring they deliver genuine value.