LogRocket

Most AI products getting hype right now were built for people like us: the tech industry. But your customers don't care how the AI works — they just want their problems solved. Brian McMullin, SVP and Head of Product at Network Solutions, has spent 15 years building for small businesses at HubSpot, ezCater, Wayfair, and SamCart. In that time, he's yet to meet a business owner who couldn't wait to use AI. For most of them, an empty prompt box might as well be a terminal window. On today’s episode of LaunchPod, Brian breaks down what it actually takes to build AI products for the other 95% — the people who don’t care how the AI works but just want the outcome. TL;DR: 1️⃣ Why the answer is never a big empty text box, and why Brian’s team prioritizes shipping the finished artifact instead of selling AI as the feature 2️⃣ How proactive AI falls flat the moment it feels generic, and why the context you build on a specific customer is going to be your biggest differentiator at a time when everyone has access to the same models 3️⃣ Brian’s warning about baking inference into the core of your product before token pricing settles, and the two monetization models he's testing to find out what SMBs will actually tolerate 4️⃣ What SamCart taught Brian about churn: customers left because basic things were broken, not because features were missing If you're building AI into a product for customers who don't necessarily care about AI, this episode is a playbook for selling the outcome instead of the technology. Check out the full episode with Brian below 👇 📖 Read on Substack: https://lnkd.in/gFcQaEnH 📺 Watch on YouTube: https://lnkd.in/gFCYJSbi 🎧 Listen on Apple Podcasts: https://lnkd.in/g_v8BQkw 🎧 Listen on Spotify: https://lnkd.in/gkj8YWwq

AI agents are turning the web into one giant CMS. Every day, less of your traffic is human, but you're still paying to serve bots that never see your brand. Brandon Harris, VP of Product at Pantheon, has seen this shift from both sides. Before helping power a large part of the WordPress and Drupal world, he was the one sending the bots, leading a team at Wiser that scraped millions of pages a day for retail pricing intelligence. On the latest episode of LaunchPod, Brandon explains why AI agents aren't killing the web, and why the companies that adapt now are about to pull way ahead. TL;DR: 1️⃣ Why agents are just the new "mode of transportation" for information, and why human traffic remains as valuable as ever 2️⃣ Why site visits are the new impressions, and why you should start grading your traffic now, tracking which agents drive real conversions and which are pure cost, so you know what's worth paying to serve 3️⃣ How AI is turning the entire web into one giant CMS, and how to leave your brand's "fingerprints" on content that gets pulled when agents don’t visit your site 4️⃣ How to catch user drift before it quietly erodes trust in your product Check out the full episode with Brandon below 👇 📖 Read on Substack: https://lnkd.in/gwqUycVE 📺 Watch on YouTube: https://lnkd.in/gtVK3S-S 🎧 Listen on Apple Podcasts: https://lnkd.in/gBu43rCG 🎧 Listen on Spotify: https://lnkd.in/gUmE47XQ

Millions of recordings. One smart summary 💡 AI just changed my debugging routine 🎯 On one of our test apps, we capture full sessions by using LogRocket. Every API call. Every network error. Every user click. That's great for debugging chaos. But it also means millions of recordings pile up every month. Nobody has time to watch all of that. So today I connected Claude to our LogRocket data by using Claude Connectors. And used prompt engineering to automate the process. Now it delivers, once a week: 📌 Bottleneck reports 📌 Graph-based visualizations 📌 Issues sorted by severity 📌 Suggested solutions for each It's scheduled and running directly in production. My team just gets the report. Clean, simple, useful. Small setup. Big time saver. This is a small change with a big impact on how we handle QA. This is what AI-assisted QA should look like. #SoftwareQA #AITools #TestAutomation #EngineeringLife #AIinTesting

AI can build almost anything you ask it to, but it has no idea whether it should. That gap is what product sense is for. Kevin Sung, VP of Product Management at Life360, joined LaunchPod today to define the term every PM uses and very few can actually define, and to explain why AI is making it the most valuable skill in product. TL;DR 1️⃣ What product sense actually is: knowing a customer's pain well enough to have a real opinion about it, and being able to turn that opinion into business value 2️⃣ The Dropbox retention bet that was wrong and worked anyway, raising $24M in ARR thanks to churn reduction 3️⃣ Kevin's "moments of resentment" theory: your most loyal users aren't always happy — sometimes they’re stuck; and the small frustrations that pile up until a price increase or an outage finally gives them a reason to churn 4️⃣ How to get leadership to fund “unglamorous” quality work when the driver model says it can't possibly move the number If you've ever been told to justify a good idea with a financial model you didn't totally believe in, or you're trying to work out which part of the PM job AI genuinely can't do, this episode is for you. Kevin's secret to developing product sense is straightforward — but it's the thing a lot of teams have quietly stopped doing. Check out the full episode with Kevin below 👇 📖 Read on Substack: https://lnkd.in/gacyUen5 📺 Watch on YouTube: https://lnkd.in/gf-HS2gj 🎧 Listen on Apple Podcasts: https://lnkd.in/gPwirfAr 🎧 Listen on Spotify: https://lnkd.in/gWAz8BwE

Not all friction in your product is bad. Sometimes the step you're about to delete is the reason your users trust you. That's Kunal Thadani's take. He's a Senior Director of Product at Houzz, where he works on a genuinely hard consumer problem: helping people who have no idea where to start make expensive decisions about their homes. Before Houzz, he was Head of Product at The League, the dating app Match Group acquired. He also writes Insider Growth, a newsletter about how companies like OpenAI, Asana, and Uber actually grew. Kunal visited LaunchPod today to talk about which steps in a flow are worth keeping, and why the personal AI setups everyone is lauding don't hold up in a team setting. TL;DR 1️⃣ Kunal’s three buckets for friction that's doing real work, including steps that signal trust, delays you can turn into progress, and constraints that get people to commit 2️⃣ The two ways friction goes bad, including one most companies never notice 3️⃣ What Kunal’s team built instead of individual AI setups: A team OS with a shared context layer, one repo, and a bunch of non-engineers working directly in the terminal Check out the full episode with Kunal below 👇 📖 Read on Substack: https://lnkd.in/gGtkUwyR 📺 Watch on YouTube: https://lnkd.in/giN88vGF 🎧 Listen on Apple Podcasts: https://lnkd.in/gjpYyUiu 🎧 Listen on Spotify: https://lnkd.in/gQZeVQZV