AffiliateWP stores affiliates in wp_affiliate_wp_affiliates, referrals in wp_affiliate_wp_referrals, payouts in their own table, and commission rules as plugin options and per-product meta. Each affiliate is also linked to a WordPress user, so the logged-in context maps directly to one affiliate record.

SleekAI's data-source wizard reads from those tables and option records. The chatbot's system message can include the current affiliate's status, rate, lifetime earnings, unpaid balance, last referral date, and program terms as named variables. The bot can then answer real questions: how much an affiliate has earned this month, when the next payout runs, which commission rate applies to a specific product, and how to share their referral link with a custom UTM. None of that is invented because the variables come from the affiliate's actual record.

Display conditions scope the bot to logged-in affiliates on the affiliate dashboard, while a public marketing bot can answer prospective-affiliate questions on the program landing page with no access to other people's data. Every conversation is logged inside WordPress with the affiliate ID, page URL, model, and token usage attached.

AffiliateWP partners ask the same questions over and over: what is my balance, when do I get paid, what is my commission rate, why did a specific referral not credit. The answers all live in the AffiliateWP tables, but partners do not always know which dashboard column maps to which question, and a misread can spark a support ticket within minutes. A generic chatbot trained on the open web cannot help because it has no access to a specific affiliate's record, and a guessed number is worse than no answer because affiliates rely on accuracy to trust the program.

AffiliateWP already stores everything the assistant needs: per-affiliate records linked to WordPress users, referral and payout history, and global plus per-product commission rules. SleekAI's data-source wizard maps that data into the chatbot's system message at request time, scoped to the current logged-in affiliate, so every answer reflects their real account. Display conditions keep the affiliate-aware bot on the dashboard only, multibot lets a public bot answer prospective-affiliate questions on the landing page, and an admin bot can summarize program-wide performance without exposing per-partner data publicly.

Conversation logs feed back into both program documentation and the system prompt over time. The result is a program that scales: affiliates get accurate self-service, program managers get audit-friendly logs, and support hours stay focused on the questions that genuinely need a human.