The connective thread running through this week’s conversations is what it takes to earn a customer’s confidence when AI increasingly sits between the brand and the person on the other end. At Furniture.com, that means building a “decision layer” that helps overwhelmed shoppers choose rather than throwing more options at them. ReflexAI applies the same confidence problem to the sales floor, closing the gap between what reps learn in training and what actually happens in a live buyer conversation. Infobip extends it into the stands, turning one-way broadcasts into ongoing, personalized relationships with sports fans. Scrunch takes the argument to its logical endpoint, arguing that brands now have to be legible to the AI agents doing the searching in the first place. And this week’s bonus pick from The Artificial Intelligence Show zooms out to the bigger question underneath all of it: who controls the models, the data, and the transformation when AI becomes the layer every business runs on.
Furniture.com’s CMO Dan Bennett and SVP Alexandra Seaman on helping consumers navigate endless choice
Dan Bennett, CMO and founding team member, and Alexandra Seaman, co-founder and SVP of Furniture.com, join Greg Kihlström at the company’s New York headquarters to explain why they’ve stopped thinking of discovery as the customer’s problem and started solving for confidence instead. Their reframe is that furniture already offers a million ways to choose, so a “decision layer” that sits on top of trusted retailers, semantic search, and an AI assistant named Dottie does more good than another 200 sofas on an endless aisle. Much of the conversation dwells on the unglamorous foundation beneath all of it: roughly a year of machine-learning work to standardize wildly inconsistent retailer data before any AI could be built on top, and a two-year bet on an in-house editorial team publishing around 140 pieces a month that now drives strong visibility in LLM answers and converts at a higher rate than organic. Bennett also makes the case for measuring a “trust score” assembled from time on site, content engagement, return visits, and Dottie prompts rather than optimizing purely for speed to checkout.
ReflexAI CEO Sam Dorison on the gap between sales training and what happens in the real world
Sam Dorison, co-founder and CEO of ReflexAI, makes the case that a great marketing engine can be undermined in the final mile by a single sales conversation that goes off the rails, and that most reps have been handed info sheets and PowerPoints and wished good luck. His argument is that AI-powered simulation moves practice past static role-play into deliberate reps against realistic buyer personas, including the emotional register of a question asked optimistically versus pessimistically and the unexpected pivot when a prospect brings a new decision-maker or a fresh line of questioning into the room. Dorison is pointed about measurement and ROI, insisting that gains have to be quick and quantifiable, and he describes how knowing the right things to measure, like whether an 85-percent trainee is one skill away from 98 or several skills away from 88, is what compressed one team’s ramp and produced a thirty-plus-percent performance bump in under eight weeks. He also sketches where simulation goes next, from quarterly business reviews to manager feedback to candidate evaluation before a hire is ever made.
Infobip’s Ben Lewis on building deeper relationships with sports fans
Ben Lewis, VP of Marketing and Growth at Infobip, unpacks what sports franchises are really after when they invest in a connected omnichannel fan experience, and why research showed fans want continuous, off-season connection rather than a steady drip of merchandise pitches. Drawing on Infobip’s work with the TGR Haas F1 team and its AI-powered RaceMate companion, he describes how unifying fragmented ticketing, support, and merchandise data into a single platform is the prerequisite for personalization that actually feels empathetic instead of tone-deaf. Lewis argues for a fan-first, channel-second approach, and frames data collection as a value exchange that only holds if the brand keeps delivering relevance after the fan has handed over their information. His measurement lens moves past vanity metrics toward session depth, repeat engagement, ad-to-conversion rates, and, eventually, the lifetime value of a fan, with hyper-personalization and proactive AI conversations as the trajectory he expects next.
One Amazing Thing About Scrunch with Devin Stevens
Devin Stevens, Head of Solutions Engineering at Scrunch, walks through the company’s agent experience platform and the uncomfortable data point behind it: a growing share of website visitors are now bots and agents, and the rich, JavaScript-heavy experiences marketers spent decades building don’t work for them. He distinguishes among indexer bots, training bots, and the retrieval bots that arrive when someone asks a question and the model comes to see whether a brand belongs in the answer, then demonstrates how Scrunch creates an alternate, structured version of each page so a retrieval bot isn’t burning thousands of tokens fighting through navigation and noise. The payoff he describes is more mentions, more citations, and higher-intent AI referrals, all without ripping apart the existing site. Stevens is candid that much of the work is still educational for marketers steeped in traditional SEO, and that Scrunch leans on heavy research investment to keep pace with models that change how they read content almost daily.
Bonus Pick: [The AI Show Episode 224]: Fable 5 Is Back, Palantir CEO’s Explosive Interview, the Pillars of Business AI Transformation & OpenAI Offers 5% of Company to US Government — The Artificial Intelligence Show
Paul Roetzer, founder and CEO of SmarterX and Marketing AI Institute, and co-host Mike Kaput, SmarterX’s chief content officer, use this episode to work through the widening question of AI sovereignty: who actually controls the models, data, and competitive edge as businesses come to run on frontier systems. They break down the U.S. government lifting its export controls on Anthropic’s Fable 5 and Mythos 5, Palantir CEO Alex Karp’s combative argument that enterprises are handing their “alpha” to the labs, and Roetzer’s new eight-pillar framework for business AI transformation spanning vision, strategy, data, technology, governance, literacy, people, and performance. It pairs naturally with this week’s Agile Brand conversations because it names the stakes underneath them: the same shift toward an AI-mediated marketplace that Furniture.com and Scrunch are navigating is the one raising hard questions about ownership, trust, and who is prepared to compete in it.
Taken together, this week’s guests are describing the same terrain from different vantage points. Whether the goal is a confident furniture purchase, a sales rep who can handle the unexpected pivot, a fan who feels genuinely known, or a brand that shows up in an AI-generated answer, the winners are the ones building trusted, well-structured foundations beneath the AI layer rather than chasing the surface of it. That is the quieter, harder work that turns AI from a threat into an advantage. See you next week!
