Mike Galvin

Stop wasting months on perfect automations. Build, launch, and learn in under a week: You want fast feedback. You need practical skills. The secret? Low-code automation. Most people overcomplicate this. They get stuck picking the “best” tool. They worry about custom code versus low-code platforms. Then months pass. No clients. No progress. Today’s coaching call with my client hammered this home: Both had strong software backgrounds. Both wanted to build AI automations. Both debated custom builds versus platforms like n8n or Make. They had a long list of ideas—receipt processors, email classifiers, language assistants. Instead of starting, they compared tool features, worried about hosting, and got lost in tech details. Here’s what worked: - Pick one low-code platform (n8n is my go-to for devs). - Use Airtable as your flexible data layer. Learn the basics, skip the headaches of Google Sheets. - Build direct—work left to right, node by node. Don’t plan every step. Don’t drop dozens of placeholders. - Test every node as you go. Manual triggers keep things real. - Launch ASAP to real users (even if it’s basic). - Learn from actual feedback—not guesswork. Why take this path? Because the longer you stay stuck comparing tools, the longer you wait for critical input from users and clients. Every day spent tweaking is a day you aren’t earning or improving. You avoid the classic stall-out that kills momentum for most automation agencies and solopreneurs. Start now: - Pick ONE idea on your list today. - Choose n8n and Airtable. - Build a basic flow in two hours. Don’t overthink. You’ll have something live—usable, testable, and ready for honest critique. Want faster progress? Stop optimizing and start executing. Feedback beats feature lists.

4

Not every product failure is a tech problem. Some are taste problems. Quinn Favret and Hassaan Raza make a powerful case in this episode. Fast shipping is great. But if you don’t love what you’re building or who you’re building it for, it shows. This conversation is a reminder to create with conviction, not just code. Want more insights like this? Subscribe to The AI Report here: https://lnkd.in/gGi2d5fK

91

65 Comments

Nobody cares what you built. They care what it delivers. Founders love to talk about their stack: "We built an orchestration layer with real-time compliance." Cool. But… so what? That’s like saying you installed plumbing. Where’s the running water? Features don’t sell. Outcomes do. Most teams pitch like this: “We offer backend infrastructure for operations.” Translation: We’re generic, and you’ll have to figure out the value yourself. Now try this: “We cut your operating drag by 60%, automate compliance, and get your GTM live in 10 days.” See the difference? One makes you think. The other makes you buy. At Oddz, we got tired of shipping tools that sounded smart but landed flat. So we reframed everything around outcomes. Take GridOS: It’s a backend operating system, sure. But who cares? So we stopped talking about what it is. We started talking about what it does. What GridOS does: — Cuts manual ops by 70% — Automates your logistics, GTM, and compliance in days — Gets you from product to market without a 6-month engineering sprint — Powers 10+ ventures from Day 1 without spinning up new infra each time Suddenly, people lean in. Because now it solves their problem—not just your architecture puzzle. Here’s the shift: Stop describing the tool. Start describing the result. It’s not "internal systems"—it’s speed, certainty, margins. It’s not "custom integrations"—it’s launch in 3 days, not 90. It’s not "infra-as-a-service"—it’s $0 wasted on duplicated dev work. Founders don’t want a backend. They want to go live fast, stay compliant, and scale without chaos. GridOS does that. That’s what we sell. Not lines of code. Not abstract systems. Just clean outcomes with clean ROI. If you're still pitching the “what,” you're already invisible. Buyers, investors, users—they’re overloaded. They're not reading your spec sheet. They're scanning for results. So ask yourself this: If I delete my product name from the sentence… Does the value still stand? If not, you’re selling the wrong thing. “We built a modular backend system.” vs. “We help ventures go live in days—not quarters—and scale ops without adding headcount.” Outcomes sell. Features confuse. Make the leap. Stop shipping tools. Start delivering outcomes. Your users will thank you. Your growth will show it.

5

2 Comments

Founders obsess over tech debt. But design debt? It’s worse. Messy UX piles up, onboarding breaks, your support team drowns in tickets. By the time you fix it, growth is already stalled. Design debt compounds just like code debt. Ignore it, and you’ll pay twice.

64

1 Comment

𝗧𝗵𝗲 𝗦𝗮𝗮𝗦 𝗦𝗲𝗹𝗹-𝗢𝗳𝗳 𝗜𝘀𝗻’𝘁 𝗔𝗯𝗼𝘂𝘁 𝗩𝗮𝗹𝘂𝗮𝘁𝗶𝗼𝗻𝘀. 𝗜𝘁’𝘀 𝗔𝗯𝗼𝘂𝘁 𝗕𝗿𝗼𝗸𝗲𝗻 𝗕𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗠𝗼𝗱𝗲𝗹𝘀. 𝗣𝗿𝗼𝗽𝗲𝗿𝘁𝘆 𝗠𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 𝗦𝗵𝗼𝘂𝗹𝗱 𝗣𝗮𝘆 𝗔𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻. Over the past few weeks, we’ve watched a sharp sell-off across the SaaS sector. The market message is pretty clear: 𝗔𝗜 𝗶𝘀 𝘂𝗻𝗱𝗲𝗿𝗺𝗶𝗻𝗶𝗻𝗴 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗺𝗼𝗮𝘁𝘀 𝘁𝗵𝗮𝘁 𝘄𝗲𝗿𝗲 𝗯𝘂𝗶𝗹𝘁 𝗼𝗻 𝘀𝗲𝗹𝗹𝗶𝗻𝗴 𝘀𝗲𝗮𝘁𝘀, 𝗹𝗶𝗰𝗲𝗻𝘀𝗲𝘀, 𝗮𝗻𝗱 𝗮𝗰𝘁𝗶𝘃𝗶𝘁𝘆 — 𝗻𝗼𝘁 𝗿𝗲𝘀𝘂𝗹𝘁𝘀. This isn’t new. We’ve seen this movie before. Advertising once sold impressions. Then Google and Facebook rewired the industry. CPM (cost-per-impression) gave way to 𝗖𝗣𝗔 (cost-per-acquisition) and 𝗥𝗢𝗔𝗦 (return-on-ad-spend), and budgets flowed to platforms that could prove impact. Incumbents didn’t fail because ads disappeared — they failed because their pricing models stopped making sense. 𝗣𝗿𝗼𝗽𝗲𝗿𝘁𝘆 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 𝗶𝘀 𝗰𝗹𝗼𝘀𝗲𝗿 𝘁𝗼 𝘁𝗵𝗮𝘁 𝗺𝗼𝗺𝗲𝗻𝘁 𝘁𝗵𝗮𝗻 𝗺𝗼𝘀𝘁 𝗽𝗲𝗼𝗽𝗹𝗲 𝗿𝗲𝗮𝗹𝗶𝘇𝗲. For decades, PM has operated on a CPM-like model:  • Fees based on % 𝗼𝗳 𝗿𝗲𝗻𝘁 𝗼𝗿 𝘂𝗻𝗶𝘁𝘀 𝗺𝗮𝗻𝗮𝗴𝗲𝗱  • Value measured by activity (rent collected, tickets closed, reports delivered)  • Weak linkage between 𝗳𝗲𝗲𝘀 𝗽𝗮𝗶𝗱 and 𝗮𝘀𝘀𝗲𝘁 𝗽𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 In plain terms: owners pay for effort, not outcomes. That worked when execution itself was scarce. AI is changing that. Traditional PM providers now face structural — not cyclical — challenges:  • 𝗠𝗶𝘀𝗮𝗹𝗶𝗴𝗻𝗲𝗱 𝗶𝗻𝗰𝗲𝗻𝘁𝗶𝘃𝗲𝘀: fees stay flat whether NOI improves or deteriorates  • 𝗟𝗮𝗯𝗼𝗿-𝗯𝗼𝘂𝗻𝗱 𝗺𝗮𝗿𝗴𝗶𝗻𝘀: scale is limited by headcount, not insight  • Focus on managing process, not 𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗶𝗻𝗴 𝗿𝗲𝘀𝘂𝗹𝘁𝘀 AI breaks this equilibrium:  • Leasing becomes a 𝗰𝗼𝗻𝘃𝗲𝗿𝘀𝗶𝗼𝗻 𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝗽𝗿𝗼𝗯𝗹𝗲𝗺  • Maintenance becomes 𝗽𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝘃𝗲, not reactive  • Delinquencies and renewals become 𝗽𝗿𝗼𝗯𝗮𝗯𝗶𝗹𝗶𝘁𝘆-𝗱𝗿𝗶𝘃𝗲𝗻  • Pricing becomes 𝗰𝗼𝗻𝘁𝗶𝗻𝘂𝗼𝘂𝘀 𝗮𝗻𝗱 𝗺𝗮𝗿𝗸𝗲𝘁-𝗮𝘄𝗮𝗿𝗲 Once outcomes are measurable and optimizable, charging “X% of rent” starts to look like charging per ad impression in 2005. The next generation of winners won’t sell “property management.” They’ll sell performance:  • Fees tied to 𝗡𝗢𝗜 𝗹𝗶𝗳𝘁  • Upside-sharing on rent growth and occupancy  • Accountability for cost avoidance and bad-debt reduction The SaaS sell-off is a warning shot. 𝗣𝗿𝗼𝗽𝗲𝗿𝘁𝘆 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 𝘄𝗼𝗻’𝘁 𝗱𝗶𝘀𝗮𝗽𝗽𝗲𝗮𝗿. 𝗕𝘂𝘁 𝘁𝗵𝗲 𝗮𝗰𝘁𝗶𝘃𝗶𝘁𝘆-𝗯𝗮𝘀𝗲𝗱 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗺𝗼𝗱𝗲𝗹 𝘄𝗶𝗹𝗹. AI isn’t making PM cheaper. It’s making the old pricing indefensible.

5

January 2026 has changed a lot of tax rules. And if you're not paying attention, you're going to get hit with penalties you didn't see coming. Here's what's new: - Virginia might tax digital goods and services for the first time. Streaming. Software. Cloud hosting. All potentially taxable. - 1099 reporting thresholds jumped from $600 to $2,000 for tax year 2026. Good news if you're a small vendor. Messy if you're still using old processes. - Illinois introduced a punitive 15% tax rate for sales without proper destination documentation. Higher than any actual tax rate in the state. - USPS postmarks are now delayed by days. Mail something today; it might not get postmarked until next week. Miss a tax deadline because of it? That's on you. - The EU is ending its €150 customs duty exemption. Low-value imports? About to get expensive. Avalara APAC has published the full breakdown of what changed and what you need to do: https://ow.ly/7zWc50Ye9nS

10

Most growth-stage startups don’t need more feature engineers. They need engineers who: -Can walk into a degraded system -Diagnose root cause -Improve deployment velocity -Realign stakeholders and customers -And leave the system stronger than they found it I spoke to an engineer this week who does exactly that. That’s not a nice-to-have skill. That’s how you protect product velocity between 50 and 150 employees. If you’re scaling engineering, ask yourself: Are we hiring builders? Or are we hiring stabilizers? Different profiles. Different outcomes.

2

This has been David Gratton 's conviction - people with domain expertise - a 20 year lawyer, doctor, specialist, are best positioned to leverage AI ... or, "context is king." I have a counter-hypothesis - that really strong product and design teams have the empathy craft with anti-bias training and they can add subject matter experts early to incorporate context. Craft specialists who work closely with customers in a pairing relationship or customer council relationship have tighter loops than beta release users where can be a longer feedback cycle. My bet follows Andrew Ng's recent feedback on engineers in one of his teams moving so fast that they need more PMs to focus on what should be built. This is the specific edge where domain experts, and often engineers, can anchor too strongly on their own background and priors and drive hard at a solution, whereas product managers and designers can step more comfortably into the white space to capture a business opportunity that aligns user adoption, technical capability, and business value. AI definitely demands more domain experts - especially on the customer side, much earlier in the product development work, possibly at the very beginning. I'd add that a16z very likely has a sampling bias problem in the claim - as more technical expertise was needed for the last three years. We now have a wave of tools that massively lowers the entry barriers - Claude Code and OpenAI's Playground to build customGPTs for example.

7