WMS DIGITAL | AI + EXNESS PARTNER API The client journey pack | 3 October 2026, Accra START WITH ONE PERSON Someone enquires while you are asleep. Capture their interest, answer the question, establish the next step and remember what happened. The Partner API lets you check attribution and available account milestones. AI interprets the question and prepares useful language. Your workflow decides who may see data. FIVE WORKFLOWS YOU CAN REUSE 01 Capture and onboarding: Telegram consent, six fields, human identity review, Exness attribution/status and community join requests. 02 Qualification: read the private queue, distinguish self-declared experience from account milestones, and choose a suitable next action. 03 Status-aware follow-up: recheck an authorised person's attribution and available status, then draft the message for that step. 04 Education: choose experience and topic; AI drafts a short lesson and quiz. 05 Re-engagement: identify a recorded conversation gap, respect opt-outs and prepare a permission-based check-in. 00 Demo: six fictional conversations. Start here before any real client. QUICKEST REHEARSAL A. Open the workshop Journey Lab. Change the scenario and inspect the decision. B. Read templates/messages.txt and copy one approved example. C. Import workflows/00-demo.json into n8n. It runs manually only. D. The full service is included under service/. Follow service/SETUP.txt to deploy your own Worker and private database, then connect your model. E. Replace YOUR_ONBOARDING_WORKER.workers.dev on the HTTP node. Select an HTTP Header Auth credential: Authorization = Bearer YOUR_SERVICE_TOKEN. F. Set scenario to welcome, qualify, onboarding, engage, return or stop. Set draft:false if you want to test decisions without a model key. G. Execute. Inspect the actual node output. The stop case produces no draft. WHAT EACH FILE NEEDS 01: your Telegram bot + deployed service + private Exness gateway. Use one Telegram trigger per bot token. Never activate two owners of the webhook. 02: deployed service + Header Auth + people already onboarded through your bot. 03: the same plus your Exness gateway and a verified consenting client record. 04: deployed service + direct Meta API key + Header Auth. No client account needed. 05: deployed service + Header Auth + eligible records. Empty output is valid. 00: deployed service + Header Auth. Exness credentials are not used by demo cases. The companion private gateway and Form/CRM workflows are in service/workflows/. AI PROVIDERS The working generation adapter calls direct Meta Muse Spark 1.3 Contributor. Its input is a fixed template, topic and experience level. It does not receive names, emails, chat history, Telegram IDs, API account data or credentials. Prompts in templates/prompts.txt also work as manual inputs in your chosen model. Other providers in the main service support intent routing; the new generation method is direct Meta only. Do not claim other generation adapters were tested. WHAT IS LIVE VS A TEMPLATE The WMS bot's original onboarding/API connection is live. /journey CASE is a partner-only live AI drafting demonstration with clearly fictional records. The downloadable n8n workflows are inactive and contain no credentials. The private queue/check/draft routes work on your deployed service. The queue uses at most 500 bot profiles. A check refreshes one authorised record at a time. Every new follow-up workflow ends in a draft or review queue. There are no customer-send nodes and no scheduled customer broadcasts in this pack. The existing bot replies to incoming messages and performs requested handoffs. TURN A DRAFT INTO FOLLOW-UP Review the request and contact permission, latest reply, opt-out and support status before sending. Use your existing approved contact channel. If you later add a send node, recheck these conditions immediately before sending and store an idempotency key plus last-sent time. Limit frequency and stop on reply. Keep review on while rehearsing. Do not activate a loop against your whole list. Onboarding consent by itself is not a recurring marketing subscription. A /human request holds the support flag. After resolving it, the partner can use /resolved TELEGRAM_ID. This clears the flag without messaging the person. WHAT THE PARTNER API ADDS - The correct existing partner registration link. - An attribution check after independent identity verification. - Recorded KYC, first-deposit and first-trade flags for that authorised client. - Private partner reports, which remain outside the model's prompt. A new verified check supersedes a saved stage. Missing or >24-hour-old cached status holds status-based follow-up until a fresh check. Reports can still lag upstream even after a fresh request. The client report does not prove current trading activity, explain a KYC failure, or certify CPA qualification. HOW THIS RELATES TO EARNINGS DA: useful acquisition and onboarding can help appropriate prospects progress to referrals that meet Exness qualification rules. Measure qualifying referrals and approved rewards against acquisition costs. A registration is not a payout. IB: useful education, support and ongoing client relationships can support a business earning commission from eligible client trades. Measure observed commission, client support quality and acquisition/service costs together. These are commercial hypotheses to test, not promises of revenue improvement. No workflow pressures people to deposit or trade to generate commission. PRIVATE DATA AND REHEARSAL Never paste real client reports into a public model or workshop projector. Never put credentials into exported JSON. Use n8n credentials and Worker secrets. Keep execution-data saving off. Use fictional records in the room. The sample CSV uses fictional people. It is a worksheet, not a live client import. The form/phone capture, bot identity review and API status paths require their own rehearsal. A successful sample AI draft does not prove a real client journey. See rehearsal.txt for exactly what was checked in this release.