Yes. If your fees live on a services page, in ACF fields, or in postmeta, SleekAI pulls that data into the system prompt as live context. The bot then quotes the same numbers your written proposals do. When you update a fee on the page, the next conversation reflects it without any retraining or redeploy step. For ranged pricing, the bot is configured to quote the range and then ask the qualifying questions that move it toward a specific number.

Encode your minimums in the system prompt: typical budget floor, sweet-spot project size, and the industries where you actually have proof. When a prospect clearly doesn't fit, the bot says so politely, links to a relevant blog post or template you sell, and skips the calendar entirely. A discovery call that ends in "we can't afford you" wastes 45 minutes. A chat that ends the same way wastes nothing.

Yes. Add case studies as a custom post type with an industry taxonomy in WordPress. SleekAI uses the taxonomy as a filter, so a fintech inquiry surfaces fintech work, a healthcare inquiry surfaces healthcare work. For long case study libraries, the OpenAI Files vector store handles retrieval so the bot fetches the most relevant two or three rather than dumping the entire portfolio into context.

Not necessarily. Many designers prefer to keep public pages directional ("projects from $X") and let the bot give tighter ranges in chat once it has the brief. The system prompt can be configured to quote a public-friendly range first, then sharpen the number once scope and timeline are clear. Conversation logs stay private in your WordPress install, so what the bot says in chat is not indexed for SEO.

BYO API key means you pick: OpenAI, Anthropic, Google, or OpenRouter. For most designer portfolios a mid-tier model is enough since the bot is mostly retrieving from your site content and asking qualifying questions, not writing essays. Logs include the model name and token usage per conversation, so the cost-per-lead picture is visible from day one rather than something to estimate.

The bot doesn't replace Calendly or SavvyCal, it hands off to them. When the qualifying answers fit, the bot drops the booking link and pre-fills the email and project type so the visitor doesn't repeat themselves. Most designers use a Zapier or webhook trigger to push the chat transcript into Notion or Airtable so the scoping call has full context before the camera turns on.

Yes, that's what Multibot is for. Run a brand identity bot on the identity pages with one system prompt, a web design bot on the web pages with another, and a generalist bot on the homepage that triages between them. Each bot keeps its own conversation logs tagged with the bot ID, so post-hoc analysis of which voice converted better is a database query, not a manual transcript review.

Every conversation is logged in WordPress with the model name, token count, and the page URL. You can review them in the admin, export to CSV for tax records, or push them into your CRM via webhook. Nothing routes through a hosted Sleek service. The data path is browser to your WordPress, then WordPress to your chosen AI provider with your API key, then back.