How ColdIQ built a 2,500-tool GTM directory – 1,000+ pages in 60 days – on 60 automation workflows

A sharp idea that needed a machine, not a team

ColdIQ had a bold, clear idea: build the world's first GTM-only tool directory. Not another G2 clone. Not another bloated review site. A lean, structured, automated engine that helps go-to-market teams instantly find the right tools, their alternatives, their pricing, their integrations, and even their discounts. Something that doesn't just list tools but generates real insight – and scales without an army of researchers.

The idea was clear. The execution was a different story.

To make a niche directory genuinely useful, it needed thousands of tools each with clean structured data, nested categories and directories, dynamic product pages, and a whole layer of narrow pages – {tool} alternatives, {tool} integrations, {tool} discounts – for every tool in the database. It needed a way to auto-discover tools, check their pricing, scrape their pages, and assemble full product profiles. And it needed a system that could run hundreds of tasks at once without collapsing.

On top of all that, it had to happen fast. ColdIQ wanted the directory to grow aggressively in a couple of months, not a couple of years.

Building all of that by hand would have taken a team of ten a full year. Three ways to attack it, each with an obvious failure mode:

1. Hire a research team. Ten people manually researching, writing, categorizing, and maintaining 2,500 tools and their alternatives, integrations, and pricing. A year of work and a permanent payroll line – and it's stale the moment the market moves.
2. License generic directory data. White-label a G2-style dataset and skin it. That defeats the entire premise: you end up with the same bloated, un-trusted data everyone already has, not the lean GTM-specific engine that was the whole point.
3. Build a static site and fill it in over time. Fine for fifty tools. It collapses at 2,500 – every tool multiplying into alternatives, integrations, and discount pages – and it never stays current.

The fourth option – the one that worked – was to build a machine instead of a directory. More than 60 automation workflows that continuously discover tools, enrich them with structured data, generate the content, and publish the pages, running on a stack built to scale. This case study walks through how that machine was built, what it produced in two months, and where an automated directory hits its limits.

What we built

The engagement produced a self-running content machine: a GTM-only directory architecture, a stack chosen for scale, and 60+ automation workflows that discover, enrich, and publish tool data on their own. The pieces only matter together – the architecture defines what gets built, the stack makes it fast, and the workflows keep it growing without anyone touching it.

The directory architecture

Everything started with a structure that could scale and stay automated from day one. Four layers:

1. Directories – five main groups organizing the whole GTM landscape
2. Categories – grouping tools with similar use cases underneath each directory
3. Product pages – the home for each tool's data, pricing, integrations, and descriptions
4. Narrow pages – {tool} alternatives, {tool} integrations, and {tool} discounts generated for every tool

That last layer is where the scale comes from. Every tool in the database doesn't produce one page – it produces a product page plus an alternatives page, an integrations page, and a discounts page. The architecture was designed so that adding a tool automatically spawns its whole page family, rather than requiring anyone to build them one at a time.

The stack that made it "busyless"

To make the system genuinely hands-off, it runs on four pieces working together:

• Airtable – the operational database, where discovery and enrichment tasks are queued and tracked
• MongoDB – the primary database powering the live directory
• Paywall – the CMS the pages publish through
• AIrops – the content generation engine that writes descriptions and assembles pages

The combination gives the flexibility of a startup stack with the backbone of a bigger platform: Airtable for fast operational iteration, MongoDB to hold 2,500 tools without straining, a CMS to publish, and an AI content layer to fill the pages. None of the individual tools is exotic – the leverage is in how the workflows chain them together.

60+ workflows: the engine

This is the heart of the project. More than 60 custom workflows run the loop that discovers, enriches, and publishes tool data continuously. Two of them show how the whole system thinks.

Workflow #1 – automatically finding pricing. For each tool, the workflow searches the web for its pricing page using Tavily, then checks whether that page actually exists. If it doesn't, it falls back to looking for pricing inside the homepage. Once pricing is found, it extracts the smallest pricing tier, whether pricing is per user / per month / per seat, whether a trial is available, and a short pricing description – then assembles a full pricing table with all the plans. What normally takes a researcher 15 to 20 minutes per tool became a zero-touch step that runs across thousands of tools.

Workflow #2 – identifying alternatives. Using each tool's homepage, pricing page, and features page, the workflow matches it against the market through two APIs – DiscoLike and Ocean.io – both returning lists of similar GTM tools. Then it does the clever part: for each alternative, it checks whether that tool already exists in the database. If it doesn't, it adds the new tool automatically and generates the tasks to create both its product entry and its full product page. The directory grows itself – every alternative discovered becomes a candidate that the system researches and publishes without anyone searching for new tools by hand.

The same programmatic-seo-ai approach (linked in the sidebar) underpins the whole build: define the page templates once, then let the workflows populate them at scale from structured data.

From system to explosion

Once the machine was running, it didn't just add tools – it accelerated. In two months the directory created 1,000+ pages, added 2,500 tools to the databases, generated both short and long descriptions, assigned each tool to the right categories, placed them in the correct directories, and built the alternatives, integrations, and discount pages for each one. It even generated affiliate links – automatically.

The loop is the same every time a new tool appears: collect the data, classify it, generate the content, and publish the pages. No human intervention at any step. The directory stopped being something anyone maintained and became something that maintains and expands itself.

What I delivered

• The full directory architecture: five directory groups, categories, dynamic product pages, and {tool} alternatives / integrations / discounts pages
• 60+ custom automation workflows for tool discovery, pricing extraction, alternative matching, enrichment, and publishing
• The database setup – Airtable (operational) + MongoDB (primary) – feeding the Paywall CMS and the AIrops content engine
• The pricing workflow: Tavily search → existence check → homepage fallback → tier extraction → auto-built pricing tables
• The alternatives workflow: DiscoLike + Ocean.io matching → auto-add new tools → auto-generate product entries and pages
• Automatic affiliate-link generation baked into the publishing pipeline
• A self-expanding system that discovers, classifies, generates, and publishes new tool pages with no human in the loop

The numbers

• 1,000+ pages published automatically in 60 days – product, alternatives, integrations, and discount pages
• 2,500 GTM tools added to Airtable + MongoDB, enriched with descriptions, categories, pricing tables, and affiliate links
• 650 visits and 25 free trials in the first 30 days
• 60+ custom workflows running the full discover → classify → generate → publish loop

What changed

After two months, ColdIQ had something no one else in the GTM world had: a fully automated, always-growing directory built specifically for go-to-market teams. It now works as a discovery engine, an integration explorer, a competitor-analysis tool, a pricing hub, and an affiliate revenue channel – all at once, and all from the same underlying data.

The real win is that the directory expands on its own. If a tool isn't in the database yet, the system adds it. If it needs pricing, alternatives, descriptions, or integrations, the workflows generate them. A traditional team would have spent months just outlining a directory this size; instead it grows day by day, pulling more GTM searches and strengthening the brand with every tool it absorbs.

And the early signals show it's more than a content dump. 650 visits and 25 free trials in the first 30 days is real intent from a brand-new directory – GTM teams finding the tools they're searching for and converting into ColdIQ's funnel. This wasn't just building a directory. It was building a search engine for GTM tools, powered by data and automation.

Honest caveats

• Automated enrichment isn't flawless. Auto-scraped pricing and generated descriptions need periodic spot-checking – some tools have pricing structures the workflow simplifies, and a fully automated directory trades a little per-page polish for enormous scale. Ongoing QA is part of running it.
• The 650 visits and 25 trials are a strong 30-day start, but they're the beginning of the curve. A 2,500-page directory's SEO compounds over months as pages get indexed and rank – the first month is a leading indicator, not the ceiling.
• A self-expanding system needs guardrails. Auto-adding tools and generating affiliate links at scale means monitoring for duplicate or junk tools, thin pages, and quality drift. The machine runs itself, but someone still has to watch the dials.
• Programmatic directories carry SEO risk. Thin or near-duplicate auto-generated pages can draw algorithmic scrutiny. The structure and enrichment are built to avoid that, but it's a real risk that needs monitoring as the directory scales past its first few thousand pages.

Want to build your own growth machine?

Book a 30-min call. Bring the directory or dataset you've been putting off because it needed a team of ten. We'll map the workflow system that builds and grows it automatically – and you decide from there.