When Software Becomes a Chat
How AI are companies are using the product-led-growth strategies that built big software companies
In 2009, a company built on free tools went public raising $112 million. Austin-based SolarWinds had cracked a code that would define an entire generation of software companies: give away utility tools to capture customers.
SolarWinds’ strategy was deceptively simple. Their marketing team built 75+ free utility programs—subnet calculators, TFTP servers, network monitoring tools—everything IT administrators needed for daily work. As IT professionals downloaded and used these tools, they entered SolarWinds’ ecosystem and mailing list.
The formula became the foundation of product-led growth: build a useful tool that solves a real problem, give it away for free, capture contact information, then nurture those leads into paying customers for your broader platform. Companies like HubSpot, Mailchimp, and Moz would perfect this playbook over the next decade.
I lived this strategy firsthand at Spiceworks, where we drove 6 million monthly users to our platform through a combination of content and free utility tools. The approach felt unstoppable—provide immediate value, build trust, convert users into customers.
But last week, I recreated that same subnet calculator—one of the tools that helped launch a multi-billion company—in Claude’s Artifacts in 30 seconds.
No download required. No email capture. No sales funnel. Just a working subnet calculator, custom-built for my specific needs, available instantly whenever I need it.
This is the story of how the product-led growth playbook might be changing.
The Utility Tool Empire
HubSpot’s Website Grader has quietly become one of the most successful lead generation tools in B2B marketing history. Since its launch, it has analyzed over 2 million websites, each analysis turning a curious visitor into a potential customer. But HubSpot’s grader isn’t unique—it’s part of a vast, largely invisible economy of utility tools that power the modern internet.
Search for “subnet calculator” and you’ll find dozens of options: SolarWinds’ Advanced Subnet Calculator, Calculator.net’s IP Subnet Calculator, Site24x7’s IPv4 calculator, and MxToolbox’s CIDR calculator. Each serves thousands of IT professionals daily, quietly building mailing lists and nurturing leads.
The developer tools landscape is even more crowded. Need to format JSON? Choose from Toptal’s formatter, JSONLint, or CuriousConcept’s validator. Testing regular expressions? Try regex101 or OpenReplay’s tester. Every category of utility tool—from IP calculators to performance testers to security scanners—hosts dozens of competing options, each one a potential entry point into a company’s sales funnel.
This isn’t the freemium model that Figma or Slack popularized, where you get limited features and upgrade for more. Utility-first PLG is different: companies give away fully functional, standalone tools that solve complete problems. No signup required, no feature limitations, no upgrade prompts. Just pure value, delivered instantly.
The psychology is brilliant. When HubSpot’s Website Grader tells you your site scores 67% and needs work, you’re not being sold to—you’re being helped. When SolarWinds’ subnet calculator saves you 10 minutes of manual IP math, you remember who solved your problem. Trust builds through repeated value delivery, not sales pitches.
SolarWinds perfected this multiplication effect by building 75+ tools, each targeting the same IT administrator audience from different angles. Network monitoring, server management, security scanning—every tool became another entry point to the same ecosystem. The compound effect was extraordinary: instead of one lead magnet, they had dozens, each reinforcing their authority in IT infrastructure.
This approach dominated because it solved the fundamental challenge of B2B software sales: how do you build relationships with buyers who don’t want to be sold to? The answer was elegant—don’t sell, help. Provide immediate value through tools that solve real problems, capture contact information naturally, then nurture those relationships over time.
But that balance is about to shift.
When AI Builds Your Tools
But why look for a tool or utility if AI can whip up a custom tool just for you and your team’s needs? This is exactly what happened when I asked Claude to generate a subnet calculator for me. Within 30 seconds I had a React app that was 3-D printed into existence. I could share this with my team if I wanted. I didn’t have to go anywhere, learn some new UI, or join yet another mailing list with annoying sales people.
This isn’t unique to Claude. ChatGPT’s Canvas provides a collaborative interface for building React apps and interactive tools that you can iterate on in real-time. Google’s Gemini Canvas integrates with Firebase Studio for instant app prototyping, complete with deployment capabilities. All three platforms now let you create, share, and collaborate on custom applications without leaving the chat interface.
If you’re in B2B marketing, you’ve already seen AI come for your content strategy. The long-form guides, templates, and FAQ content you’ve created is now answered directly by AI bots. Now AI is coming for your utility tools—those JSON formatters, regex testers, and subnet calculators that have been quietly building your mailing lists for years.
But will people realize they can do this directly within AI? Are people so used to searching for utilities, calculators, and spreadsheet templates that this isn’t going to be a route they take? Research from Nielsen Norman Group suggests that “many users still default to Google” even when AI tools could solve their problems more efficiently. The search habits that made utility-first PLG so successful might be exactly what protects these tools from AI disruption.
At least for now.
The Same Playbook, Zero Cost
But what if AI vendors take the same approach that’s worked for PLG companies? After all, SolarWinds realized that people need subnet calculators, spent money building subnet calculators, launched them via Google, and converted those users into customers. The formula was simple: find the customer’s job to be done, build tools for those jobs, and capture those users.
But AI companies have near-zero cost to build those tools. Customers build them themselves using the AI interface. So why not use the same strategy?
And that’s exactly what Anthropic is starting to do. Looking for a JSON formatter? You’ll likely run into a Google Ad like the ones below that Anthropic is running.
They’re running ads for brand strategists, copywriters, marketing roadmap designers, market research analysts, and perhaps many more. After all, they’re limited only by API calls to run the ads. Where SolarWinds needed development teams, QA testing, and deployment infrastructure for each of their 75 tools, all Anthropic needs is one API call to launch the ad campaign.
They can use the same strategy PLG companies perfected, with one big difference—their cost of development is zero. Just infinite utility tools at zero marginal cost.
And they want all users. They want sysadmins, website builders, developers. Everyone.
How Far Up the Tree?
We’ve seen this pattern before. A few years ago, I wrote about how Google was taking over informational searches with their one-page results. Instead of sending users to websites for basic facts, weather, or simple calculations, Google started answering these queries directly in the search results. They took the low hanging fruit first—the easy wins that required minimal context or nuance.
Now AI companies are doing the same thing with utility tools. They’re starting with the simple stuff: JSON formatters, subnet calculators, regex testers. The utilities that are straightforward to build and don’t require deep domain expertise or ongoing relationships.
This is how new technology always works. It takes the low hanging fruit first, forcing existing companies to climb higher up the value tree. Websites had to offer more sophisticated content once Google started answering basic questions directly. Now PLG companies will need to offer more complex, relationship-dependent value as AI commoditizes their simple utilities.
But how far up the tree does the value stop? That’s the question driving the AI cycle.The answer will determine not just the future of product-led growth, but the future of software itself.