Yes. Add your current share certificate, share savings, money market, auto, and HELOC rates as a page or ACF fields, and SleekAI reads from there rather than guessing. The bot consistently frames the quote as 'current as of this week, subject to change' and points the member to the application page where the live rate is locked in at funding. Rate edits happen in WordPress so the rate team can update without touching the bot.

Add your field-of-membership rules as a page (SEG list, geographic eligibility, family eligibility, association partners) and the bot walks the criteria with a visitor. Clean eligible cases get pointed at the online membership application. Edge cases (a SEG that's no longer current, an unusual family relationship, a multistate situation) get routed to a member service representative who can confirm in your core system.

No. SleekAI's system prompt is configured to explain products, fees, and processes without recommending specific accounts, projecting earnings on IRAs or HSAs, or making personalised tax or investment suggestions. When a member asks an advice-shaped question (which certificate ladder is best for retirement, should I refinance), the bot frames the answer as general product information and routes them to a member service representative or your credit union's financial advisor partner.

Yes, and this is critical for compliance. The bot is explicitly instructed never to confirm balances, recent transactions, or any data tied to a specific account on the public site, since the public site can't authenticate the member. Account-specific questions get routed to the authenticated online banking portal, phone, or branch. The bot's job on the marketing site is product and process, not account servicing.

Yes. Use multibot plus display conditions to run a membership-focused bot on the membership and eligibility pages, a lending-focused bot on auto, mortgage, and HELOC pages, and a business-focused bot on the business banking section. Each bot has its own system prompt, knowledge sources, and routing logic, so the conversation matches the product the member is actually looking at on the page.

GPT-4o and Claude Sonnet both handle financial product language well and follow eligibility logic reliably. Smaller or older models tend to drift on field-of-membership rules and to invent rates if a page is incomplete. Bring your own key for whichever provider your compliance team has approved, and switch the model per bot if you want to run a stronger model on lending and a cheaper one on general FAQ.

Every conversation is logged in your WordPress database with the model name, token usage, and page URL. Most credit unions export transcripts on a recurring schedule via webhook into the same archival workflow that already handles email and secure-message communications. Retention is fully under your control, and the logs are searchable from the WordPress admin if compliance needs to look up a specific conversation later.

Yes. If your branch list lives as a custom post type or pages with location ACF fields, the bot can ask for the member's ZIP, return the two or three closest branches with hours, and offer to drop a pin in a map widget if you have one. It can also surface branch-specific staffing notes (Saturday hours, drive-thru, notary on-site) so a member knows what to expect before they come in.