Turn Your Coaching Framework Into a Between-Session AI

TL;DR: Your coaching framework can become an AI companion in under two hours — upload your methodology documents, write instructions that define scope and voice, add conversation starters for common client needs, and share via direct link. The agent supports clients between sessions while you focus on the work only you can do.

Most coaches have already done the hard part. The framework exists. The exercises are documented. The client handbook is written. The FAQ has been answered so many times it's practically memorized.

That's the knowledge base for an AI agent right there — upload your framework documents to Alysium, configure behavioral instructions, and the agent answers questions in your voice from your content.

What most coaches haven't done is make any of that accessible without their own presence. It lives in Google Drive folders and email drafts and in their head — useful only when they're in the room.

This guide is about changing that. Here's how to take the methodology you've spent years building and turn it into an AI coaching companion that works around the clock.

The distinction worth making upfront: this is not about replacing how you coach. It's about making your framework accessible between sessions, to more clients, at any hour. Coaches who've done this consistently report that the live session quality improves — clients arrive more prepared, having already worked through the conceptual layer with the AI, leaving session time for the high-context, high-nuance work only you can do.

Step 1: Map What Your Framework Actually Contains

Before you upload anything, spend 20 minutes mapping what you have. A coaching framework typically contains several distinct content types, each serving a different purpose in the agent:

• Core methodology overview — what the framework is and how it works: the backbone of the agent's knowledge base
• Exercises and worksheets — the tools clients use to apply the methodology: high-value for between-session access
• Client FAQ — the questions clients ask most often, with your standard answers
• Session preparation guides — what to bring to sessions, how to reflect on progress, what to expect
• Resource recommendations — books, tools, or practices you point clients toward

Each category contributes differently to the agent. The methodology overview anchors everything. The exercises and FAQ are what clients will actually search for between sessions. Map what you have before uploading — it'll save time and produce a better agent.

The mapping exercise often reveals that more of the framework is already documented than coaches realize — it's just scattered. The coaching deck for onboarding new clients. The welcome packet. The email sequences. The worksheets. These are all framework documentation in different formats. Inventory what already exists before deciding what to write from scratch. In most cases, 60–70% of the knowledge base can be built from existing materials with light editing, not new creation.

Step 2: Prepare Your Documents

Raw documents often need minor preparation before they're agent-ready. The most common issue: content that made sense as a linear read doesn't produce good retrieval results when chunked.

A few quick improvements that help: add a clear heading to each section. Break long explanatory paragraphs into shorter, topic-focused ones. Rewrite any section that assumes the reader has just read the previous section — agents don't retrieve sequentially, they retrieve contextually.

You don't need to rewrite everything. Target the 20% of content that covers the most common client questions — that's what the agent will retrieve most. The rest can be uploaded as-is and improved later based on real usage data.

The most effective format for an AI knowledge base is not your most polished document — it's your most detailed one. The overview deck you show prospective clients is designed to intrigue, not explain. The facilitator guide you use internally is designed to explain, not impress. Upload the facilitator guide. The AI retrieves from specifics; it performs poorly on high-level summaries because they contain the right words but not the right answers.

Step 3: Upload to Alysium's Knowledge Base

Create your Alysium account and set up a new agent. Then upload your prepared documents — Alysium supports 11 file types including PDF, Word, and plain text. You can also paste content directly, which is useful for FAQ material that lives in email drafts or message threads.

Upload in priority order: methodology overview first, FAQ second, exercises third. This mirrors the order of retrieval frequency for most coaching use cases — and if you run out of time, the most valuable content is already indexed.

Documents process in the background with a live status indicator. Most files index within a minute or two. Check the status before moving to the next step.

One tip that consistently produces better results: for your FAQ document, structure it as explicit question-and-answer pairs rather than continuous prose. Write the question as a heading, then answer it in 2–4 sentences beneath. This format mirrors how clients will search, and the agent retrieves individual Q&A pairs much more precisely than paragraphs that weave multiple topics together.

Expected outcome: Your core framework content is uploaded and indexed.

Upload order matters less than upload completeness. Start with your highest-priority documents — the framework overview and the FAQ — since those typically cover the highest-volume questions. Then add supplementary materials: worksheets, case examples, troubleshooting guides. You can upload incrementally, so there's no need to wait until everything is ready to start testing. A core knowledge base of 3–4 documents often gets 80% of the quality of a comprehensive one.

Step 4: Write the Behavioral Instructions

This is the most important step — and the one that determines whether your agent sounds like you or sounds like generic AI.

Write instructions that cover four things:

Identity: "You are the AI coaching companion for [Your Name]'s [Framework Name]. You support clients of [Your Name] in applying the framework between sessions."

Tone: Write your actual communication style here. Specific language, not vague directions. "Warm and direct. Use plain language. Short paragraphs. No jargon unless it's part of the framework itself."

Scope: "You answer questions about the framework, the exercises, and how to apply them. For questions requiring personalized guidance specific to a client's situation, you acknowledge what they've shared and direct them to bring it to their next session with [Your Name]."

Fallback: "If you cannot find a clear answer in your knowledge base, say so honestly and suggest the client reach out to [Your Name] directly."

Spend 30–45 minutes on this. Good instructions are what separate a generic AI from a coaching companion that genuinely sounds like you.

Expected outcome: Instructions that capture your voice and define the agent's scope.

Step 5: Add Conversation Starters for Common Client Needs

Think about the 5 questions your clients ask most often between sessions. Not the complex personalized ones — those need sessions. The ones that have consistent, valuable answers in your framework materials.

These become your conversation starters. They do two things: they tell clients what the agent is good at, and they pull every first interaction toward a win.

Examples for a life coaching practice:

• 🔑 "Walk me through the 3-step reflection exercise"
• 📋 "What should I prepare for my next session?"
• 💡 "I'm stuck on a decision — where do I start?"
• 📚 "What does the framework say about [common challenge]?"
• 🌱 "How do I apply what we covered in session to this week?"

Expected outcome: 3–5 conversation starters that represent your agent's strongest Q&A.

Between-session client questions tend to cluster around three moments: right after a session ("I want to apply what we covered today"), mid-module ("I'm stuck on this exercise"), and before the next session ("I want to prepare for tomorrow's call"). Design conversation starters for each moment. Starters that name the moment — "I just finished our session and want to apply what we discussed" — convert better than generic ones because they meet clients where they actually are.

Step 6: Test as a Client Would

Before sharing with real clients, run 15–20 test conversations as if you were a new client who'd just finished their first session.

Ask questions a real client would ask — including some that fall outside the framework's scope. Verify that the agent handles both kinds correctly: good answers for in-scope questions, honest redirects for out-of-scope ones.

Pay attention to tone above everything else. Does it sound like you? Does it feel like a resource that represents your practice well? If the tone is off, the fix is almost always in the instructions — not the knowledge base.

Expected outcome: You're confident you'd share this with a real client.

Test from the mindset of a client at day 7 of your program, mid-struggle. Not "how does the framework work" (a question from day 1) — but "I'm stuck on step 3 and I've tried X, what do I do?" (a question from someone who's been working with it). The agent should handle the second type as well as the first. If it doesn't, the knowledge base is missing your troubleshooting and obstacle-handling content — the material that distinguishes coaches from information.

Step 7: Share With Active Clients

Alysium generates a direct collaboration link that you can share with clients via email, Slack, your client portal, or any onboarding sequence. Recipients click the link and get an immediate session — no account required.

Frame it clearly when you share: "This is your between-session resource, trained on my framework and materials. Use it when you have questions and I'm not available." That framing sets expectations and positions the agent as a feature of working with you, not a replacement for it.

Most clients appreciate having 24/7 access to framework support. The key is the framing: this is a premium, not a workaround.

A few words on the experience you're creating: the clients who use a between-session AI companion most effectively are the ones who understand it's a tool for their development, not a substitute for the relationship. They use it to go deeper on things you covered together, to practice applying frameworks before the next session, to get unstuck without waiting. That's different from using it as a passive information library — and the difference shows up in their outcomes.

You can shape this by being explicit in your onboarding about how to use the agent: "Think of it like a study companion who's read everything we work with — ask it things you'd ask me if you could reach me right now."

Ready to build your coaching companion? Start free on Alysium — your methodology documents are already the foundation.

The introduction matters as much as the tool. Clients who understand why you built the companion and what it's for — "so you can access framework support between sessions without waiting for my availability" — adopt it faster than clients who receive a link with minimal context. A brief personal message ("I built this for you — it knows our whole framework and is available 24/7") creates more engagement than a generic announcement. Consider rolling it out to 2–3 existing clients first and asking for their honest feedback before broad launch.