Delta Galil runs a world-class multi-brand CX operation without turning support into a cost center: five DTC brands, one Gorgias help desk, 60%+ automation. Director of Customer Service, Ajit Patel, and Senior Manager, Nicole Parker, walk through the architecture, AI knowledge bases, and the metrics behind it, including what broke along the way and how they fixed it.

Key takeaways

  • CX is a revenue department, not a cost center. CS-driven demand went from 10% to 18% as AI Agent scaled, with phone converting at 20%+, chat at 15%+, and Shopping Assistant at 13–14%.
  • Treat AI like a new employee. Ajit seeded the first Guidance from their training deck, then Nicole refined them with their CSM, like coaching a new hire's tone.
  • Don’t always optimize for speed. At 60%+ automation, the team stopped tracking handle time so agents can spend real time with customers when it counts.
  • AI catches what humans miss. Gaia and Opportunities flagged a franchise-license address error on the site before it became a compliance problem.

Five brands, one inbox

‍

Delta Galil runs Bare Necessities, 7 For All Mankind, Splendid, Florence by Mills Fashion, and P.J. Salvage out of a single Gorgias instance. Each brand gets its own queues, channels, and routing, built on rules, teams, and omniview smart routing. This allows agents to see a clean, brand-specific view instead of one tangled inbox.

Adding a new brand is closer to plug-and-play than rebuild: once Shopify's connected and the rules are set, a new store takes about 10–15 minutes to stand up.

Treat AI like a new employee

‍

The team's mental model for AI is what unlocked it: don't expect it to know your business, train it like a new hire. Ajit seeded the first guidance by dropping their training deck into an AI tool, and then Nicole rebuilt it properly with their CSM, tuning the tone and edge cases the same way you'd coach a new agent.

The payoff is that AI absorbs the repetitive volume like order status, returns, tracking and even scaled cleanly through a multi-week warehouse move that would otherwise have buried the team in tickets.

The metric they stopped tracking

‍

At 60%+ automation, Ajit stopped looking at handle time. When AI resolves one-touch tickets, the goal for a human agent is no longer speed; it's care. Agents now have room to take on higher-touch initiatives with customers, like virtual bra fittings for Bare Necessities.

Nicole frames the before-and-after plainly: a few years ago agents were exhausted sending the same reply all day against hour-long phone queues. Now, AI handles quick tickets so the team can spend real time with the customer.

Meet Gaia

‍

Gaia is Gorgias's "AI for AI" — a coaching agent that helps configure, clean up, and analyze your knowledge base. It's how the team collapsed a messy, sprawling Guidance library into something lean enough for the AI to read and respond fast.

It's also a safety net. While beta-testing Opportunities, Gaia surfaced a site issue where the AI was giving out the wrong address for a franchise-licensed brand, a compliance catch the team fixed before it ever became a problem.

CX as a revenue department

‍

The case Delta Galil makes to leadership is in the numbers. As AI scaled, CS went from driving 10% of demand to 18%, with phone converting at 20%+, chat at 15%+, and AI Agent alone at 13–14%. Chat on-site converts at 15–20% versus the 4–5% a typical site sees.

And it's not revenue at the expense of experience: the team targets 90%+ CSAT (best-in-class) and tracks it alongside sales, because when CSAT dips, sales follow.