
Posts
Posts


What a conversational careers website is, how it differs from a basic chatbot, what good looks like, and a 12-question checklist to use in every demo. It applies to any product, including ours.
A conversational careers website adds a natural-language layer to your careers website. A candidate types or says what they want, for example 'weekday jobs near Leeds with no nights', or asks a question, for example 'what is the first month like?'. The system finds matching jobs, answers from content you have approved, and suggests the next step. The jobs come from your applicant tracking system, and the answers should come from your own pages.
| Area | Basic chatbot | Conversational navigation |
|---|---|---|
| Where answers come from | A fixed script or a list of FAQs | Your live jobs and approved careers content, matched to the question |
| Understands the question | Keywords or menu choices | Natural language and intent, with follow-up questions |
| Jobs | Often a link to the job search | Finds and explains roles that fit what the candidate said |
| Employer content | Rarely linked | Can point to stories, benefits and what you give and ask |
| When it does not know | Loops or says "I did not understand" | Says so plainly and offers a person or the right page |
| What you learn | Chat counts | What candidates ask, and where answers or content are missing |
These are typical differences, not guarantees. The only reliable test is a live demonstration with your own jobs and content.
Tell people at the first message that they are talking to an AI system. Under Article 50 of the EU AI Act, providers of AI systems that interact directly with people must make sure people are informed, unless it is obvious. The obligation applies from 2 August 2026. Take legal advice for the markets you hire in.
Restrict answers to your live jobs and content you have approved, and show the source.
A wrong answer costs more trust than no answer. Offer a person or the right page.
Quote published facts with a link. Route the rest to a person.
Search, filters, job pages and apply should work for everyone, whether or not they use the assistant.
Use these questions in every demonstration. For each, ask to see it working, not described.
Good: only your live jobs and the content you approve, with the source shown. Ask to see what happens when the content is missing.
Good: it says so, then offers a person, an email or the right page. It does not guess.
Good: a clear message at the first interaction. EU rules on this apply from 2 August 2026. Take legal advice for your markets.
Good: you can set the voice, review conversations and correct answers, and you know who is responsible.
Good: a live feed from your ATS. Closed jobs vanish at once and are never offered.
Good: answers can draw on your employer content and link to it, so candidates learn what you give and what you ask.
Good: published facts only, with a link, and a human route for the rest. No promises.
Good: works by keyboard, with a screen reader, at high zoom, with clear focus and large enough buttons. Test it yourself. See our accessibility guide.
Good: loads after the main content and you can show a before and after test on a phone.
Good: a short privacy notice, no more personal data than needed, clear retention, and you own the data. Take legal advice.
Good: what candidates ask, searches with no answer, content gaps and which conversations lead to a job view or application start, exportable.
Good: search, filters, job pages and apply work without the assistant, and candidates can ignore it.
Questions asked, grouped by topic, and the share with no good answer.
Conversations that lead to a job view, an application start and a job alert sign-up.
Content gaps: questions that show you are missing a page or a story.
Page speed and accessibility results on the live site.
We have not included benchmark figures. If you want to compare against others, ask any supplier for evidence from live customers, and for how it was measured.
Happydance describes its conversational navigation as an assistant on the careers website that answers questions in natural language, pulls live roles from the ATS integration and draws answers from FAQs, policy pages and employer value proposition content. Our aim is to surface more employee stories and the Give and Get, so candidates learn what the job is like before they apply.
A careers website where candidates can ask questions in their own words and get answers drawn from your live jobs and approved content, with guidance to the right job or page.
A basic chatbot follows a script or a fixed list of answers. Conversational navigation interprets the question, searches your jobs and content and says so when it cannot answer.
It depends on the build and your content. Do not rely on a promised uplift. Ask suppliers for evidence from live customers and measure job views and application starts yourself.
Yes. Limit answers to approved sources, show where answers come from, make it say when it does not know, and review conversations.
They can be, but a chat window can create problems with keyboard use, screen readers and focus. Test any product yourself against WCAG.
It can. Ask for a before and after Lighthouse test on a phone and load the assistant after the main content.
It depends on the product. Ask what is stored, where, for how long and who owns it, and take legal advice on privacy.
Happydance conversational navigation answers candidate questions from your live jobs and approved content. Talk to us for a demonstration.