Topic
AI customer support
AI customer support answers customer questions automatically from a business’s own published content, rather than from a decision tree or a model’s general knowledge. Done properly it retrieves the relevant passages from your website, answers from those, cites the page, and hands anything account-specific to a person.
How it works
- 1
Your published pages are crawled and split into passages.
- 2
Each passage is indexed by meaning, not by keyword, so a question phrased differently still finds it.
- 3
A visitor question retrieves the closest passages from that index.
- 4
The model answers from those passages only, and links the page it used.
- 5
When retrieval comes back weak, the honest answer is 'I don't know' plus a route to a person.
The technique has a name — retrieval-augmented generation — and it is why updating an answer means editing a page, not retraining a model.
How to judge one
It answers from your content
Not from what a model believes about your industry. If it cannot show you which page an answer came from, you cannot verify it — and neither can your customer.
It admits when it doesn't know
A system that always produces an answer is a system that sometimes produces a wrong one confidently. Refusal is a feature.
It routes to a human quickly
Anything account-specific should reach a person in one step. Hiding the escalation route to protect a deflection number is how support software earns its reputation.
It tells you what it missed
The questions it could not answer are the most valuable output. That list is what your website should say next.
By industry
The questions differ enormously by business. These break down what visitors actually ask and which of your pages hold the answers.
SaaS companies
AI customer support for SaaS companies
SaaS support is dominated by questions your documentation already answers: what a plan includes, whether a feature exists, how an integration works. RubyRep reads your docs, pricing and changelog and answers those on the site, so trial users get unblocked in seconds instead of waiting on a ticket.
online stores
AI customer support for ecommerce stores
Most ecommerce support is four questions wearing different hats: where is it, when will it arrive, can I return it, and will it fit. RubyRep reads your product pages and policies and answers them at the moment of hesitation, which is the moment the sale is won or lost.
estate agencies and brokerages
AI customer support for real estate
Property enquiries arrive at night and on weekends, and the first agent to reply usually wins the viewing. RubyRep answers questions about listings, fees and the buying or renting process from your own site, then hands warm enquiries to an agent with the contact options you set.
clinics and practices
AI customer support for healthcare practices
Most calls to a practice are logistics, not medicine: opening hours, what a visit costs, which insurance is accepted, what to bring. RubyRep answers those from your website so reception can focus on patients in front of them — and it stays firmly out of clinical advice.
schools, colleges and course providers
AI customer support for schools and course providers
Admissions seasons concentrate a year of questions into a few weeks, and they are the same questions every year: entry requirements, deadlines, fees, term dates. RubyRep answers them from your own prospectus and admissions pages, at whatever hour a parent or applicant is actually researching.
agencies and studios
AI customer support for agencies
Agencies field two kinds of question: what do you do, and what does it cost. Both are answerable from the site, and both currently interrupt someone senior. RubyRep answers them from your services and case study pages, and doubles as something you can deploy on client sites.
Common questions
- What is AI customer support?
- Software that answers customer questions automatically, using a language model grounded in a specific business's own content. The grounding is the important part: a general model knows the internet, not your refund window.
- What can it actually answer?
- Anything your website already explains — shipping, returns, pricing, hours, availability, how a product works, what is included. It cannot answer anything about a specific customer's account or order, because it has no access to those systems.
- How is it different from a rules-based chatbot?
- A rules-based bot follows a decision tree someone drew, and breaks the moment a visitor phrases something unexpectedly. A retrieval-based one searches your content by meaning, so it handles phrasings nobody anticipated.
- Will it make things up?
- It can, and that is the central risk. The defence is retrieval plus a confidence threshold: answer only from passages actually found in your content, and say 'I don't know' when the search comes back weak rather than filling the gap.
- Does it replace a support team?
- No. It absorbs the repetitive layer — the same twenty questions asked in a hundred ways — so people spend their time on the conversations that genuinely need judgement or account access.
- How long does it take to set up?
- With a crawl-based system, minutes: it reads your existing pages rather than asking you to write a knowledge base. The work is not setup, it is reviewing what it could not answer and filling those gaps.
See what yours would answer.
Enter your website. RubyRep reads it and you can question it about your own business in about a minute.