Free vs Pro
The question this page exists to answer is the one a careful buyer actually has: if all the good stuff is free, what’s the catch?
The principle
Trust is free. Leverage is Pro. Everything that decides whether an AI assistant can be believed — where its answers come from, whether it cites them, whether it admits what it doesn’t know, what happens to a visitor’s data, and what it refuses to do — is in the free plugin. What you pay for is what the conversations can do for your business afterwards.
There is no catch, and the reason is not generosity. A compliance-first product that puts its compliance behind a paywall is a product with a hole in the middle of its own argument. We would rather have the argument intact.
Everything free, in one paragraph
Everything that makes the assistant trustworthy is in the free plugin, and stays there. Citations on every answer. An honest “I don’t know” when your content doesn’t hold the answer, and guardrails that stop an invented link or an unsupported claim being served at all. Moderation on every visitor message, with no off-switch on any plan. Consent receipts, retention policy, browser-storage controls, auto-clear, and GDPR export and erase — every privacy control, not a subset. Any provider, any model, on your own key. A health dashboard that tells you what is answering and what is failing. Nothing here is a demo of something else.
The five things Pro adds
Each of these is a separate capability with its own screens and its own work behind it — not a bigger number on something you already have. That distinction is the whole design of the boundary, and it is why the list is short.
1. Your catalogue and the wider web as sources
Free grounds answers in your pages, posts, documents and JSON. Pro adds WooCommerce products and external URLs, so the assistant can answer from your catalogue and from pages that live outside your WordPress site. Who it’s for: shops, and anyone whose real answers live somewhere their site doesn’t publish.
2. Business Intelligence
One screen for what the conversations are telling you: grounding and deflection analytics, how the answers are distributing across quality, and the knowledge gaps worth filling next. Who it’s for: anyone who wants the assistant to improve on evidence rather than on hunches. Free keeps the health facts and the grounding-outcomes card — real counts of what actually happened, never estimates of what might have.
3. Idea capture
Product feedback, feature requests and demonstrated unmet demand, pulled out of real conversations into a ledger you can triage. Who it’s for: product and service businesses whose customers keep asking for the thing they don’t sell yet. The free plugin never captures this, and that is deliberate: it does not store visitor-derived material the site owner hasn’t asked for.
4. Sales-lead detection
Buying intent flagged as it happens and emailed to your team with the transcript. Who it’s for: anyone for whom a missed enquiry costs more than the licence.
5. Human-handoff escalations
The workflow around a handoff: triggers, customer notifications, email transcripts and rate limiting. Who it’s for: teams with someone to hand off to. Note the split — on every plan, when the assistant can’t answer it says so and points the visitor at your site’s usual contact routes. What Pro adds is the machinery that offers a handoff, gets that conversation to a person and keeps a record of it — and the assistant only offers a handoff where that machinery is switched on.
Two control surfaces ride along with them: a grounding-depth dial, and moderation tuning — the strike limit and the wording of moderation replies, so you can name the crisis helplines for your own region.
What we deliberately don’t gate, and why
This is the part nobody else on the shelf will write down, so here it is in full.
- Moderation has no off-switch, on any plan. Every visitor message is checked for abuse before it reaches your provider — because your provider’s terms make you responsible for what your visitors send, and because a brand should never be made to argue with abuse. Charging for that would mean selling the unprotected version as the default one.
- Privacy controls are never paywalled. Every privacy and compliance control is free — and we don’t move them behind payment. A compliance control you cannot reach is a compliance control you do not have, and a paywalled one is worse than absent, because it looks like you have it.
- Provider choice is yours, because the keys are yours. Every provider and every model in the catalogue works on every plan. You are paying your own AI bill; being told which of your own accounts you may use would be a strange thing to buy. (Some models need a paid account with the provider — that is between you and them, and nothing to do with your AnswerRails plan.)
- Citations and honest refusals are the product, not a feature of it. An assistant that will invent an answer when the paying version wouldn’t is not a cheaper product. It is a different, worse one.
The honest limits — what Pro doesn’t do either
Paying does not change what the assistant is. It still answers from the content you give it and says so when that content falls short; it still refuses off-topic and unsafe requests; it still cannot know anything your site hasn’t published. Pro widens what counts as your content and tells you more about the conversations — it does not loosen the guardrails, and there is no setting anywhere that does.
