“A chatbot? That is for big companies. I cannot afford one.”
I hear this a lot. And I get it: it sounds like one more tech gimmick that looks great at conferences and then gathers dust.
Except the data says otherwise.
Companies that roll out chatbots save around 30% of their customer service costs. Worldwide that adds up to 8 billion dollars a year. Klarna, the payments company, announced that within a month its AI assistant had taken over two thirds of all customer chats, doing the work of 700 full-time agents. The result? An estimated 40 million dollars in profit improvement in 2024.
Source: Klarna AI Assistant Announcement, 2024
The good news: you do not have to be Klarna to benefit.
What a chatbot actually does (and what it does not)
Let us start with the basics. A modern AI chatbot is not one of those irritating pop-ups from 2015 that kept replying “Sorry, I did not understand. Please rephrase your question.”
Today’s chatbots:
- Understand natural language (you do not have to write like a robot)
- Learn from your own materials (price list, terms, FAQ)
- Answer around the clock, within seconds
- Can qualify leads and collect contact details
- Hand difficult cases over to a human
What they still do not do well:
- Empathy in tense situations (a complaint from a furious customer)
- Unusual problems that need a creative solution
- Negotiating and building deep relationships
75% of customers believe chatbots cannot handle complicated issues. They are right. That is why the best setup is a hybrid: the chatbot takes care of the 80% of questions that repeat, and passes the hard cases to a person.
Source: Plivo, AI Customer Service Statistics 2025
Who it makes sense for (and who it does not)
A chatbot pays for itself quickly if:
1. You get a lot of repetitive questions
“How much is shipping?” “Do you ship abroad?” “What are your opening hours?” “How do I return a product?”
If you or your team answer the same questions 20 or 30 times a day, a chatbot can take over 80% of that work.
Research shows that chatbots can answer 79% of routine questions with no human involved.
Source: Sobot, Chatbot Statistics 2025
2. You lose clients outside working hours
64% of users name 24/7 availability as the biggest advantage of chatbots. Your potential client is browsing your offer at 11 pm, has a question about lead time, gets no answer, and goes to a competitor who has a chatbot.
Source: Blogging Wizard, Chatbot Statistics 2025
3. You run an online store
In e-commerce, chatbots reduce abandoned carts by 20 to 30%. A customer asks “can I pay with my usual method?” at midnight, the chatbot answers on the spot, and the customer completes the purchase.
Source: DemandSage, Chatbot Statistics 2026
4. You want to collect leads while the office is closed
A chatbot can ask for an e-mail address and a phone number in exchange for answering a question. The lead lands in your CRM at 3 am, and in the morning your salesperson is the first to call.
When a chatbot is a bad idea
Not every business needs one:
- You get 5 enquiries a day: there is nothing to automate
- Every case is different and complex: strategic consulting, for example
- Your clients are seniors who prefer the phone: 61% of people aged 65 and over feel negative about chatbots
- You have no content to feed the chatbot: without a solid FAQ and knowledge base it will be useless
Source: Katana, Customers Prefer Human Over AI Chatbot
What it costs (realistically)
There are two scenarios:
Ready-made SaaS tools (Tidio, Intercom, Crisp)
A monthly subscription, quick setup, template answers and integrations with popular platforms. For a small business with simple needs this is often enough.
A dedicated AI chatbot (built on Voiceflow, for example)
There is no off-the-shelf price list here. The cost of the build depends on scope: the number of integrations, the knowledge sources and the channels the chatbot has to work in. On top of that comes a monthly running cost (the AI model and hosting). In return you get a chatbot that really knows your business: trained on your materials, connected to your systems and answering exactly the way you want.
And the return on investment?
Studies report an average return of $3.50 for every $1 invested in chatbots. For companies with a high volume of enquiries it can reach 8 times the investment.
Source: Fullview, AI Customer Service Statistics 2025
Let us run an example:
- You receive 30 enquiries a day through a form or e-mail
- Each answer takes 5 minutes on average
- That is 2.5 hours a day, 50 hours a month
- A chatbot takes over 80% of them, which gives you back 40 hours every month
Multiply those 40 hours by what an hour of your time is worth and compare it with the monthly running cost. In most cases the return is obvious.
What customers really think about chatbots
I have good news and bad news.
The bad news: 49% of customers still prefer talking to a human.
The good news: 62% of customers would rather use a chatbot than wait in a queue for a human. And 74% will choose a chatbot for simple questions.
Source: Tidio Research, quoted by Fullview
The key is setting the right expectations:
- Do not pretend the chatbot is a person (customers can tell)
- Offer an easy way to reach a human for difficult matters
- Make sure the chatbot really helps instead of repeating “sorry, I do not understand”
One more thing: 96% of customers believe that companies using chatbots care about customer service. A well-built chatbot makes your business look modern and professional.
Source: Embryo, 20 Stats on Chatbots
How to roll it out in practice
If you have read this far and you are thinking “fine, maybe this makes sense”, here is a plan:
Step 1: Collect your FAQ
Go through the last 100 e-mails and messages from customers. Write down the 20 to 30 most common questions. This will be the chatbot’s knowledge base.
Step 2: Choose a tool
For small businesses with simple needs: Tidio or Crisp (they connect to WordPress and WooCommerce in 15 minutes).
For businesses that need something more advanced: a dedicated AI chatbot trained on your materials.
Step 3: Do not overdo it at the start
Begin with the 10 to 15 most common questions. A chatbot that answers 15 questions brilliantly beats one that answers 50 badly.
Step 4: Monitor and improve
A chatbot shows you exactly what customers ask about. It is a goldmine of knowledge about what is missing on your website and where your offer has gaps.
Summary: gimmick or savings?
An AI chatbot today is not a gimmick. It is a tool that:
- Answers 80% of repetitive questions with no human involved
- Works 24/7, with no holidays or sick leave
- Serves many customers at once
- Collects leads even while you sleep
- Returns 3 to 8 times the investment within a few months
Is it for everyone? No. If you get 5 enquiries a day, keep answering by hand.
But if you are drowning in repetitive questions, losing clients outside working hours, or your team spends hours copying and pasting the same answers, a chatbot may be one of the better investments you make this year.
Wondering whether a chatbot makes sense for your business? Book a short call: I will look at your enquiries and tell you honestly whether it is worth it.

