The Complete AI Marketing Automation Guide, Built with Claude + XERJ

5 stages, 20 automations, 40 prompts, 12 templates, 6 workflows and 5 checklists to take routine work off your marketing team: research, planning, approvals, publishing and reporting. Claude does the work, XERJ gives it every file to search and remembers your context

<aside> 🚀 FREE STRATEGY CALL Want to grow your product through viral reach? We help teams across iGaming, mobile apps, AI products, and SaaS build creator-led organic distribution generating tens of millions of views at CPMs starting from $0.03.

<aside> 🚀 Book a free strategy call

</aside>

On the call you'll receive:

Viral-reach gap analysis for your product
Creator-led formats already working in your niche
Estimated install / brand-lift potential at scale </aside>

WHAT THIS IS: a ready system for the work around content. Five stages in the order a marketing week runs: Research, Content Planning, Team Workflows, Publishing and Reporting. Each stage has 4 automations. Every automation has what it replaces, the input it needs, 2 prompts and the template it fills.

WHO IT'S FOR: CMOs, heads of marketing, content leads and small marketing teams at eCom brands and AI products who spend more hours on research, status updates and reports than on the content itself.

HOW TO USE IT: set up the data folder and XERJ once (15 minutes). Then start with the stage that eats the most hours in your week. Run the prompts (P), fill the templates (T), chain them with the workflows (W) and check the result with the checklists (C).

WHY CLAUDE + XERJ

Claude is great at the work itself: sorting, summarizing, drafting, checking and reporting. It gets weak when you paste 2,000 reviews, a quarter of support tickets, 40 briefs and 12 weekly exports into one chat: the context fills up and the answers get vague. And every new chat starts from zero, so you explain your brand again.

XERJ is a free, open-source search engine and memory for AI agents. It runs locally, indexes a folder of files (docs, PDFs, CSV, transcripts, exports) and lets Claude search that folder instead of loading everything. Claude pulls only what each task needs, answers with the source file, and recalls your saved context in every new session.

Set it up once (Claude Code):

curl -fsSL https://xerj.org/get | sh
claude mcp add --scope user xerj -- xerj mcp
xerj autoindex ./marketing-data

Put everything in one marketing-data folder, one subfolder per source:

Folder	What goes in
/research	Customer interviews, surveys, competitor pages, market reports
/reviews	Product, store and marketplace reviews
/support	Support tickets and chat exports
/content	Published posts, scripts, emails, landing pages and their numbers
/briefs	Approved briefs, drafts and source material
/ops	Calendar, task lists, approval log, team roles
/metrics	Weekly exports: reach, clicks, signups, orders, revenue by channel

Re-run xerj autoindex ./marketing-data after you add files. Most teams add one weekly export on Monday and re-index once.

THE RULES

Real data only. Claude writes [NEEDS INPUT: what] instead of guessing numbers, quotes or results.
Cite the source. Every insight from XERJ comes with the file it came from.
A human approves. Claude prepares, sorts, drafts and checks. A person approves anything that goes live or goes to a client.
One owner per task. Every task Claude creates has one owner and one deadline.

Paste this at the start of every Claude session: