The Lead, Growth Marketing post describes the goal exactly: an AI agent should "identify the right Redis solution, review the relevant documentation, and create an account in a single session". Redis Iris now leads the homepage navigation, and the docs already have about 60 Iris and context-engine pages in markdown. But redis.io/llms.txt, the map Redis serves to agents (also at /agents/), never mentions Iris, Context Retriever or the Redis MCP server, links only one of those 60 pages, and has no link to /try-free/. An agent following the map today can learn every Redis command but cannot find Redis's newest product or the step that creates an account.
First action
Add the Iris, MCP and sign-up links to llms.txt today.
How to measure it
Iris sign-ups from agent and assistant sessions
- Now
- –
- Target
- 2x
- By
- 60 days
Steps
List the path
For each product: overview, quickstart, create-service doc, pricing, sign-up.
Lead, Growth Marketing 1 day
Write the section
"Agents and context" in llms.txt with markdown links.
Developer Relations 1 day
Add a chooser
Need to product, in plain lines.
Product Marketing 1 day
Link the account step
/try-free/?rcplan=iris with a note on the free plan.
Lead, Growth Marketing 1 hour
Publish
Same file at /llms.txt and /agents/.
Web Engineering 1 hour
Test with agents
Ask three coding agents to build an agent with memory on Redis.
Developer Relations 2 days
Tag sessions
Mark sign-ups whose referrer is an assistant or agent.
Lead, Growth Marketing 1 week
Review monthly
Add every new product page and quickstart.
Lead, Growth Marketing monthly
Why it works
Add an "Agents and context" section to llms.txt: Redis Iris, Context Retriever, Agent Memory and LangCache, each with its overview, quickstart and create-service doc in markdown (the .md versions already exist), the Redis MCP server install page, the pricing page, and /try-free/?rcplan=iris as the account step. Add a short "Choosing a Redis product" entry that maps needs (cache, vector search, agent memory, context retrieval) to products. Then ask the main coding agents to build a small agent with memory on Redis and see whether they reach the Iris docs and the sign-up.
Agents follow what the map gives them; adding the newest product and the sign-up turns the existing docs into the single-session path the role is hired to build. It is a text-file change that takes a day, and it can be measured by watching agent sessions reach /try-free/.
Ready to use
Checklistllms.txt "Agents and context" section
[ ] Redis Iris overview [ ] Context Retriever: concepts, create service [ ] Agent Memory: overview, Python/TypeScript/REST quickstarts [ ] LangCache: concepts, create service [ ] Redis MCP server install [ ] Pricing [ ] /try-free/?rcplan=iris as the account step [ ] "Choosing a Redis product" lines [ ] Same file at /llms.txt and /agents/
Risks and how to handle them
File grows too long
Keep a short top section and link a full version.
Agents create accounts without a person
Keep the human confirmation step in sign-up.
Evidence
- HiringLead, Growth Marketing (United States, posted September 30, 2026): "the audience discovering and evaluating Redis is no longer exclusively human. AI agents can identify a need, decide which solution fits, review technical documentation, and create a new account within one continuous session." The role owns "the website, SEO, answer engine optimization, generative engine optimization, conversion-rate optimization, and the measurement system", must "Build a low-friction journey in which an agent can identify the right Redis solution, review the relevant documentation, and create an account in a single session", and directs "contractors and an agency partner".Source
- Your websitellms.txt check (October 22, 2026): redis.io/llms.txt (also served at /agents/) lists 81 links under headings from Core Docs to Tutorials. It never mentions Redis Iris, Context Retriever or the Redis MCP server, and does not link the free sign-up page (/try-free/); of the roughly 60 docs pages under /operate/iris/ and /develop/ai/context-engine/, it links one (LangCache concepts). Those docs already exist in markdown, e.g. /docs/latest/operate/iris/agent-memory/index.html.md.Source
- Your websiteHomepage: title "Redis - Real-time data for agents & apps"; the navigation leads with "Redis Iris - Real-time context for agents", then LangCache, Context Retriever, Agent Memory, Data Integration and Flex; a banner reads "86.5% accuracy. ~$0.07 per session. See how Redis stacks up in the Agent Memory Benchmark"; a sign-up link is /try-free/?rcplan=iris; links ask ChatGPT, Claude and Perplexity "How can I use Redis Iris as a real-time context engine for AI apps?".Source
- Your websiteRedis Iris page headings: "Your agents should be getting smarter", "Introducing Redis Iris", "Address any data with Context Retriever", "Agentic AI gets smarter with Agent Memory", "The State of Context Engineering 2026" and "Take the context maturity self assessment". Docs include Agent Memory quickstarts for Python, TypeScript and REST, and Google ADK integration guides.Source
- Your websitePricing page: Free "From $0 /hour"; Essentials "From $0.007 /hour", shared deployment 250 MB-100 GB; Pro "First $200 free", "From $0.014 /hour", "Minimum $200 /month", dedicated, active-active, up to 99.999% uptime; sold direct and through partner marketplaces.Source