The Goal
Hyper-personalized landing pages in minutes, not weeks
Account-Based Marketing (ABM) is incredibly effective, but it faces a significant scaling bottleneck. To win high-value accounts, our sales team needs bespoke landing pages that speak directly to a target company’s unique business challenges and digital goals.
Historically, creating a single targeted landing page was a multi-week ordeal. It required a sales representative to research the prospect, a copywriter to draft tailored messaging, a designer to mock up the page, and a web developer to build and deploy it in Contentful. Our goal was to build a system that compressed this entire cycle down to a few minutes, putting page generation directly in the hands of the sales team while maintaining strict brand and architectural standards.
Strategy
Separation of concerns: research vs. generation
When building an automated system, we knew that human-in-the-loop validation was critical. We decided to split the workflow into two distinct phases to ensure the final landing page was both accurate and highly relevant:
- Information Synthesis: The sales representative prompts the system with the prospect's details. The AI then scans public sources—including recent financial statements, public filings, and interviews—to synthesize the company's current challenges and identify exactly how Amperity can help.
- Review & Refinement: The sales representative reviews, edits, and refines the synthesized research. This step ensures data accuracy and gives the representative complete control over the strategic positioning before triggering the automated page build.
Once the research is vetted, the second phase—automated generation—begins.
Execution
Orchestrating a multi-agent AI pipeline
To make the generation process as frictionless as possible, we built the front-end interface as a simple Slackbot. For the sales team, the technical complexity is entirely hidden: a representative simply sends a slash command containing the approved research and waits.
Under the hood, a sophisticated multi-agent pipeline takes over:
1. The Vercel Serverless Orchestrator
The Slackbot sends the payload to a Vercel serverless function, which orchestrates the AI agents and coordinates the API integrations.
2. Agent 1: The Brand Voice Copywriter
The first AI agent is trained extensively on our brand voice, messaging framework, and product positioning guidelines. It ingests the raw research summary and translates it into high-converting headlines, body copy, and value propositions tailored to the prospect.
3. Agent 2: The Front-End Layout Architect
This agent is trained on our Next.js component library and Contentful content models. It analyzes the generated copy and maps it to our existing library of responsive modules—such as hero banners, feature grids, and customer testimonials—selecting and ordering the best blocks to match the narrative flow.
4. The Contentful CMS Integration
Once the page structure is defined, the orchestrator calls the Contentful Management API to programmatically create the page entry, instantiate and link the nested modules, and populate the copy.
Impact
From weeks to clicks
Once the pipeline completes, the Slackbot alerts the sales representative with:
- A link to the draft entry in Contentful (allowing them to add any custom marketing assets).
- A staging URL to preview the rendered Next.js page.
- The production link that goes live once the page is published in Contentful.
By automating the copywriting, layout composition, and CMS entry creation, we compressed the time required to spin up a personalized ABM landing page from several weeks of cross-team coordination to under two minutes. This empowers the sales team to personalize campaigns at scale while allowing the web development team to remain focused on core platform engineering.