Steve welcomes Ibby Syed, Co-Founder of Cotera, for a conversation about the gap between AI hype and AI that’s actually doing autonomous work for ecommerce operations. Ibby brings a unique perspective as one of Peloton’s first data scientists, where he watched DTC brands scale rapidly while their teams drowned in manual work analyzing scattered customer data across disconnected tools. That experience showed him exactly where the traditional data science workflow breaks down—marketers ask for actionable insights they can implement in Klaviyo or Braze, but data scientists disappear for five to six months and return with complex model files nobody knows how to action.

The conversation centers on what makes autonomous agents fundamentally different from chatbots or query-based AI tools. Ibby breaks down the framework: instead of constantly prompting a tool or duct-taping together multiple platforms, you define a process in plain language (exactly like writing instructions to an intern), give the agent access to relevant tools (Instagram, Google, your customer data), and let it execute independently. He shares a compelling example—a D2C fitness brand wanted to identify which of their top 5% spenders could become brand ambassadors. The traditional approach would require a team of interns manually researching each customer on Instagram, checking their follower counts and fitness content, then reaching out individually—work that would take months. With Cotera’s autonomous agent, they defined the process once, gave it access to Instagram and Google, fed it their high-value customer list, and the agent completed months of research work in five minutes.

Ibby explains why the human role shifts rather than disappears in an AI-augmented operation. The repetitive work—analyzing support tickets, categorizing returns, segmenting customers based on purchase patterns—gets automated, but humans elevate to high-leverage strategic decisions, creative problem-solving, and building the memorable customer experiences that drive long-term loyalty. He uses the ATM analogy effectively: when ATMs were invented, people predicted massive job losses in banking, but what actually happened was humans got liberated from counting cash and moved into more valuable advisory and relationship roles. The banking sector employs more people today than before ATMs existed because the efficiency gains funded business growth and created new opportunities.

The conversation closes with a reality check on AGI (artificial general intelligence). Despite the hype suggesting AI will soon replace human strategic thinking entirely, Ibby spends his days wrestling with getting AI agents to execute basic tasks correctly. Current AI technology augments human capabilities powerfully for defined, repeatable workflows, but nuanced judgment, adaptive problem-solving, and strategic creativity remain firmly in human territory. This isn’t a story about AI replacing your team—it’s a blueprint for multiplying what your team can accomplish by automating the grind and elevating everyone to higher-value work.

👉  Define your AI agent workflows exactly like you’d delegate to a human team member. Write out the process in plain language as if you’re sending an email to an intern—step by step, with clear decision points and expected outcomes. The breakthrough with autonomous agents isn’t technical complexity; it’s being able to articulate your business logic clearly enough that an agent can execute it independently without constant supervision.

👉  Focus AI automation on high-frequency, low-complexity tasks first. Don’t start by trying to automate your most strategic decisions. Instead, identify the repetitive work your team does daily—analyzing support tickets for common issues, categorizing return reasons, segmenting customers based on purchase patterns—and build agents for those workflows. These are the tasks where saving 3-4 hours daily compounds into massive efficiency gains without requiring perfect accuracy on every decision.

👉  Give your AI agents the right tools, not just the right data. An agent analyzing customer churn needs access to your ESP for engagement data, your support platform for ticket history, and your Shopify store for purchase patterns. Think about what a human would need to do the job properly, then ensure your agents have those same integrations. Half-connected agents produce half-useful results, regardless of how sophisticated the underlying AI model is.

👉  Shift your team’s mindset from task execution to process design. The marketers who’ll thrive in AI-augmented operations aren’t the ones protecting their current workflows—they’re the ones who get comfortable defining processes, testing agent outputs, and iterating on automation. Start using these tools now, even imperfectly, because the skill isn’t in doing the work anymore; it’s in orchestrating the systems that do the work at scale.

👉  Measure AI ROI in time saved, then reinvest that time strategically. If your team saves 15-20 hours weekly by automating ticket analysis, returns categorization, and customer research, that’s not a license to reduce headcount—it’s fuel for growth. Redeploy those hours into strategic initiatives: launching new customer segments, testing creative campaign concepts, building deeper customer relationships. Companies that use AI savings to fund growth will outpace those that use it purely for cost reduction.

👉  Set realistic expectations about what current AI can and can’t do. Despite the hype, AGI isn’t around the corner. Current AI agents excel at executing defined, repeatable workflows but struggle with nuanced judgment calls, adaptive problem-solving, and strategic creativity. Don’t wait for perfect AI that replaces human thinking—deploy good-enough AI that handles repetitive work today, freeing your humans to do the strategic work that actually differentiates your brand.