AI chatbot, FAQ page or enquiry form?

Step 1: define the enquiry problem first

This is a narrower decision than whether your whole business should automate. If you are still deciding which process deserves attention first, read the broader guide, “AI automation for small business: what to automate first” at /insights/ai-automation-for-small-business-what-to-automate-first. Here, the question is more specific: when someone lands on your website, should they read an FAQ page, fill in an enquiry form, or talk to an AI chatbot?

Business.gov.au advises small businesses to identify the problem they want AI to solve before choosing a tool, start small, test tools, monitor them and keep human oversight. That matters because a chatbot is not automatically better than a clear page or a well-designed form. The right answer depends on where enquiries are breaking down: customers cannot find basic answers, they do not give you enough detail, or they need guided help before a human follows up.

Use real enquiry data where you can: phone notes, emails, social media messages, form submissions and questions your staff answer repeatedly. Our available Search Console observations for this site are still sparse and mostly relate to broader terms such as “ai automation for business”, so this guide is based on practical service fit and official AI guidance rather than claimed keyword demand.

Sources: [2]

Step 2: run the three-question test

For this guide, we will use three fictional customer questions: “Do you service my suburb?”, “Can I get this done this week?”, and “Roughly what will it cost?”. Replace these with your own common questions before making a decision. The point is not that every business gets these exact questions; it is that most website enquiries include a fit question, a timing question and a cost question.

Sort each question into one of three types. A public answer is safe to publish without knowing who the customer is, such as your general service area. A details-needed answer depends on the customer’s situation, such as job size, location, access or urgency. A human-judgement answer needs review by you or your team, such as a quote, exception, complaint, sensitive issue or anything with legal, health or financial risk.

This sorting step prevents overbuilding. If all three answers are simple and stable, an FAQ page may be enough. If the customer’s details drive the answer, start with a better enquiry form. If customers need a guided conversation, and you can maintain approved answers and escalation rules, a chatbot may be worth scoping.

Sources: [2] [5]

Step 3: choose an FAQ page for stable public answers

An FAQ page works best when the answer is stable, public and easy to review. For the fictional question “Do you service my suburb?”, the FAQ can list suburbs, regions or a plain-English service area. For “Can I get this done this week?”, it can explain usual lead times without pretending to know your current calendar. For “Roughly what will it cost?”, it can explain what affects cost, what information is needed for a quote, or whether you provide estimates before inspecting the job.

The strength of an FAQ page is control. You approve the wording, customers can scan it, and your staff can point people to the same answer. It also helps keep expectations consistent because everyone sees the same public information. The limitation is that customers must read and self-select. If they ask something you have not covered, the FAQ cannot respond; it can only direct them to contact you.

Maintenance effort is modest but real. Review the page whenever service areas, prices, lead times, opening hours, cancellation rules or policies change. Add a quarterly review reminder so old answers do not quietly become wrong. The FAQ should have an owner in the business, because “the website” cannot decide whether an answer is still accurate.

Step 4: choose an enquiry form when you need better details

An enquiry form is useful when the main problem is not instant answering, but getting enough information for a human to respond properly. For “Do you service my suburb?”, make suburb or postcode a required field. For “Can I get this done this week?”, ask for the preferred date, urgency and flexibility. For “Roughly what will it cost?”, ask for job details, measurements, budget range or a photo upload if that is genuinely useful for your work.

The strength of a form is structure. It can reduce back-and-forth, help qualify the enquiry and route work to the right person. It is often the best first improvement when you receive messages like “How much?” with no location, timeframe or job details. The limitation is that a form does not answer instantly. If it is too long, asks for unnecessary personal information or uses vague labels, it may create friction instead of clarity.

Maintenance effort sits in the middle. Check that notifications reach the right person, spam controls are working, required fields still make sense and your promised response time is realistic. Measure form completion rate, usable enquiry rate, average response time and the most common missing details. If people keep calling after submitting the form, the next fix may be clearer wording, not a chatbot.

Step 5: choose an AI chatbot only when guided help solves the bottleneck

An AI chatbot may be useful when customers ask varied questions, need a guided path, or are unsure which service fits. For “Do you service my suburb?”, the chatbot should answer from approved service-area information only. For “Can I get this done this week?”, it can explain your normal booking process and collect preferred times. For “Roughly what will it cost?”, it can give approved guidance about what affects pricing, then hand off for a quote where details matter.

The strength of a chatbot is conversation. It can ask follow-up questions, clarify the customer’s need and hand off a warmer enquiry to your team. Its limits are important. It needs approved source information, testing, monitoring and escalation rules. It should not guess about prices, policies, sensitive matters or anything that needs professional judgement. Do not judge it by whether it feels modern; judge it by whether it solves a real enquiry bottleneck without creating confusion or risk.

Taking bookings also needs care. A chatbot alone cannot confirm live availability unless it is connected to a compatible booking system with the right rules and permissions. Without that integration, it can collect preferred times or explain the process, but a human or booking platform must confirm the appointment. Cyber.gov.au also highlights risks such as unreliable outputs, prompt manipulation and privacy issues, so customer-facing AI needs sensible boundaries and monitoring.

Sources: [1] [2] [4]

Checklist for choosing and maintaining the enquiry path

Use this checklist after you have listed your real customer questions. It keeps the decision practical and gives someone clear ownership after launch.

• Write down your top 10 customer questions from calls, emails, messages and form submissions.
• Mark each question as “public answer”, “needs details” or “needs human judgement”.
• Choose an FAQ page for stable answers customers can safely read without providing personal information.
• Choose an enquiry form when your team needs structured details before responding.
• Choose a chatbot only when guided conversation solves a real problem and you can maintain approved answers.
• Name one person who owns website enquiry accuracy, including FAQ wording, form fields and chatbot source information.
• Set review triggers for price changes, service-area changes, policy changes, staffing changes and booking-process changes.
• For unknown questions, make the channel hand off clearly: FAQ says contact us, form captures context, chatbot says it does not know and escalates.
• Limit personal data collection, check where enquiry data is stored, and review vendor terms before using AI tools.
• Tell customers when they are dealing with AI, and keep payment details out of chat by handing payment to a suitable payment provider or human process.
• Measure response time, qualified enquiry rate, repeated unanswered questions, follow-up effort and incorrect or unclear answers.
• Simplify, pause or change the system if it creates privacy risk, wrong expectations or extra work for the team.

Sources: [1] [2]

Fictional worked example and next action

Fictional example: Coastal Air Repairs is a small air-conditioning business servicing several suburbs. Its three common website questions are our test questions: service area, availability and rough cost. On an FAQ page, the service-area answer lists covered suburbs, the timing answer explains usual lead times, and the cost answer explains that final pricing depends on the unit, fault and access. In an enquiry form, the same questions become suburb or postcode, preferred date and urgency, plus unit details and photos. In a chatbot, the same questions become a guided conversation using approved service-area, booking and pricing guidance, with handoff for quotes and anything uncertain.

After the first month, Coastal Air Repairs notices that many enquiries still lack model numbers and access details. The next improvement is not automatically a chatbot; it is a better form with clearer fields and optional photo upload. Later, if customers still need help choosing between repair, replacement and maintenance, a chatbot may be worth scoping with approved answers, safe boundaries and clear escalation to the owner or admin team.

Human ownership is the constant across all three options. Business.gov.au says businesses remain responsible for what AI tools do and should decide who is accountable. In practice, one person should approve website answers, review common unknown questions, update stale information and decide when a customer needs human judgement. The technology can support the enquiry path, but it should not become the owner of your customer promise.

Your next action is simple: collect your top 10 questions, sort each one into public answer, needs details or needs human judgement, then build the simplest safe option first. If you want help mapping that enquiry path, grounding answers in approved business information, improving forms or scoping a chatbot with sensible handoff rules, Business Growth Clinic’s AI & automation service at /services/ai-automation is the relevant service to explore.

Sources: [2] [6]