$ man how-to/abm-personalization-architecture

ABM is not a tool. It is an architectural decision. Instead of broadcasting to everyone and hoping the right people respond, you identify specific accounts, research them deeply, and build personalized experiences for each one. The traditional approach: marketing generates MQLs, sales qualifies them, most are garbage. The ABM approach: sales and marketing agree on target accounts upfront, build personalized content and outreach for each, and measure engagement at the account level. The architecture that supports this has four layers: signal detection (who to target), research and enrichment (what to say), personalized delivery (how to say it), and engagement tracking (did it work). Each layer is automated. Each layer feeds data to the next.

PATTERN

Signals tell you when to engage. Types of signals: Hiring signals: job postings that indicate budget and intent. "Hiring a RevOps Manager" means they are building infrastructure you might support. Funding signals: recent raises mean budget exists. Series A companies are building. Series B companies are scaling. Both need different things. Tech stack signals: companies adopting tools adjacent to yours are natural buyers. If they just adopted Salesforce and you sell Salesforce integrations, the timing is right. Engagement signals: website visits, content downloads, social interactions. PostHog tracks these with precision. A company visiting your site 5 times in a week without converting needs personalized outreach, not another nurture email. Cron jobs scan for signals daily. Job board scrapers, Exa web intelligence queries, PostHog engagement alerts. Fresh signals every morning.

PATTERN

Once you have a signal, you need context. RAG turns raw signals into actionable research. The pipeline: pull the company's website content (Exa), recent news (Exa), job postings (scraper), tech stack (BuiltWith or similar), firmographic data (Apollo), and any existing CRM history (Attio). Feed all of it into Claude context. Ask for a due diligence brief. The output: a structured research document with company overview, relevant pain points, buying signals, potential objections, and recommended approach. This brief feeds the personalized landing page builder and the outreach generation skill. Without RAG, outreach is generic. "I noticed you are growing and thought..." With RAG, outreach is specific. "Your 3 new RevOps hires and recent Salesforce adoption suggest you are building the data infrastructure to support your Series B growth targets. Here is how we helped a similar company cut their pipeline build time by 60%."

FORMULA

Every ABM asset creates connections: Landing page at /for/{slug} links to: relevant how-to guides, case studies, product pages, blog posts. How-to guides link to: related landing pages, other guides, knowledge terms. Blog posts link to: landing pages for companies mentioned, how-to guides for tools discussed, knowledge terms for concepts explained. Knowledge terms link to: how-to guides, blog posts, landing pages. Every new asset strengthens every existing asset. The graph compounds. 50 landing pages, 30 how-to guides, 20 blog posts, and 80 knowledge terms create a web of thousands of internal links. AI engines see comprehensive coverage. Search engines see topical authority. Visitors find relevant content on every page. This is not link building. This is content architecture. The links emerge naturally from the data relationships. The TypeScript data objects define the connections. The templates render the links. No manual linking required.