Why chatbots have a bad reputation
Almost everyone has met the old kind. You type a question, it offers four buttons, none of which match what you asked. You type it again. It offers the same four buttons. Eventually you find the "talk to a human" option, if there is one.
Those were rule-based chatbots: fixed decision trees, designed in advance. They work perfectly for the paths someone anticipated and fail completely for everything else — which, in practice, is most real questions.
An AI chatbot interprets what was actually asked. It handles unexpected phrasing, mixed Hindi and English, three questions in one message, and the sort of half-sentence people actually type on a phone. That is a genuine difference, not a marketing distinction.
The design decision that matters most
Where the answers come from determines whether a chatbot is an asset or a liability.
Ungrounded: the chatbot answers from the language model's general knowledge. It will be fluent, confident and occasionally wrong about your business — inventing a policy, a price or a service you do not offer. This is the single most common failure and it is entirely avoidable.
Grounded: the chatbot retrieves relevant passages from your own material — service pages, price lists, policies, brochures — and answers from them. When nothing relevant exists, it says so and offers a person.
Any chatbot representing your business should be grounded. If a provider cannot explain how theirs is, that is the question to keep asking.
What a chatbot realistically does for a business
Stated in terms you can check afterwards.
- Captures enquiries outside working hours. The most consistent benefit. A large share of website visits happen in the evening, and a visitor who gets an answer at 10pm is a visitor who has not gone to a competitor by morning.
- Answers the questions that cause people to leave. Do you serve my area, roughly what does it cost, are you available on these dates, what is included. One line each, and the visitor stays.
- Improves enquiry quality. Qualifying questions asked in the conversation mean your team receives context rather than a bare phone number.
- Frees your team from repetition. The same twenty questions, handled consistently.
- Reveals what your website fails to say. Conversation logs are one of the most useful things you will read in the first month. Every repeated question is a gap on a page.
What it does not do: create demand, fix an unclear offer, or replace a sales conversation. It removes friction from people who were already interested.
What makes a chatbot annoying
Almost all of it is configuration rather than technology, which means it is avoidable.
- Interrupting immediately. A pop-up two seconds after arrival, before the visitor has read anything.
- Hiding the human. No visible route to a person is the single most resented pattern.
- Pretending to be a person. Customers work it out, and the ones who feel misled complain publicly.
- Being confidently wrong. A wrong price quoted by your own website is worse than no price.
- Implying someone is available at 2am. Say honestly when a person will respond.
- Demanding contact details before answering anything. Answer the question first; ask afterwards.
The knowledge base is the project
People assume building a chatbot is a technical exercise. In practice, the build is rarely the constraint — preparing the material is.
What has to happen:
- Collect the real material: current price lists, package inclusions, policies, service details, terms, and the answers your team already types out every week.
- Resolve the contradictions. Most businesses have an old price list somewhere, a superseded policy, a brochure with last year's timings. These have to be sorted out before the chatbot inherits them.
- Write down what was never written down. What happens if a booking is cancelled. Whether the price includes taxes. What the minimum order is. Which areas you will travel to.
- Decide what stays private. Internal margins, negotiation limits, staff information.
Step three is where businesses discover that the gaps were already costing them enquiries — visitors had been leaving with the same unanswered question long before anyone considered a chatbot. Filling those gaps improves the website and the way AI systems describe you as much as it improves the chatbot.
What it costs
Three separate numbers.
Build: one-off, driven mainly by how much knowledge preparation is required. A small business with clear, current content is a modest project. A large catalogue with many policies is not.
Running: AI model usage and hosting, scaling with conversation volume. Should be billed to accounts in your own name so you can see the usage and leave if you want to.
Upkeep: small but real. When prices, packages or policies change, the source material needs updating. A chatbot quoting last season's rates is actively harmful, and this is the most common way a good chatbot becomes a bad one.
Should your business have one?
A reasonable test:
Probably worth it if: you get a meaningful number of website visitors or WhatsApp messages each week; the same questions come up constantly; a significant share of enquiries arrive outside working hours; your information is documented or could be; and enquiries currently go unanswered on busy days.
Probably not yet if: your traffic is very low, in which case fixing visibility comes first; your enquiries are all genuinely unique; your information is not written down anywhere, in which case that is the project; or you already respond to everything within minutes, in which case a chatbot adds little.
There is also a middle path worth considering. Before commissioning a chatbot, publish clear answers to the ten questions you are asked most, on the relevant pages. Some businesses find that this alone resolves most of the problem — and if it does not, you have just produced the knowledge base the chatbot needed anyway.
Chatbot or agent?
A chatbot converses: it answers and captures. An AI agent also acts in your systems — creating records, checking availability, triggering workflows.
Most businesses should start with a chatbot. It is simpler, cheaper, and carries much less risk, because nothing changes in your systems if it gets something wrong. Once the conversation quality is proven over a few months of real use, adding actions is a smaller step than starting there.
Starting with an agent because it sounds more advanced is a good way to discover both your content gaps and your integration problems at the same time, under live conditions.
Considering a chatbot for your website?
PresenticAI builds grounded chatbots that answer from your own material, capture enquiries properly and hand over cleanly — and will tell you if your enquiry volume does not justify one.