AI Knowledge Bases
Grounded AI assistants built from your actual content, tested against wrong answers, capped with a token budget, and monitored in production.
AI assistants
Not a chat toy with a personality. A support agent that answers from your documents, declines what it can't source, and hands humans the conversations that need one.
TL;DR
A grounded business chatbot answers only from your content, retrieved at question time. Ours costs $4,500 fixed for the base build (docs/FAQs/policies, wrong-answer test suite, token caps, human escalation) plus mandatory production monitoring from $497/month. Generic SaaS widgets cost less and know nothing about your business; that's the actual trade.
Three real options. SaaS widgets deploy fast and answer from scripts or nothing. In-house RAG builds take months and a team. A custom grounded build on a proven boilerplate lands between: your content, production discipline, weeks not months. The deciding question is what a wrong answer costs you; if the answer is money or trust, scripts and improvisation are out.
Answers with sources. The bot cites which document it answered from, declines questions your content doesn't cover, and escalates to a human with a clean summary. The wrong-answer test suite runs before launch and keeps running in production. The full mechanics live on the knowledge-base page.
Tokens, re-indexing, and monitoring. Tokens are a real per-conversation cost that scales with traffic; re-indexing keeps answers current as your content changes; monitoring catches drift before customers do. All three live in the mandatory attach, from $497/month, on the Website Management page, priced and published because surprise costs are how AI projects die.
Grounded AI assistants built from your actual content, tested against wrong answers, capped with a token budget, and monitored in production.
An AI-readiness assessment before anything gets built.
Predictable monthly retainers that keep your website updated, secure, backed up, and monitored, on any platform, including custom PHP/CodeIgniter applications.
Going deeper
Choose grounding first, widget second: decide what content the bot may answer from, then pick the build that enforces it. Ours embeds as a lightweight widget on any platform and answers only from your indexed content.
$4,500 fixed for the base build, from $497/month for the mandatory monitoring. The boilerplate is why we can hold that price against a fully custom scope.
Yes, and that's precisely where grounding matters most: it quotes your published prices and policies verbatim from the indexed source, and the test suite includes traps for exactly these.
It says so and escalates: the visitor gets a human handoff, and your team gets the transcript plus a summary. A bot that guesses instead of escalating is a liability, not a feature.
Yes. The widget is a small script embed; the intelligence runs serverless. Platform-agnostic is the house rule.
