In short: A traditional chatbot follows a decision tree or reacts to keywords. An AI chat agent understands questions phrased in the visitor's own words, answers from your knowledge base and can get things done, such as booking an appointment or opening a ticket. For a handful of fixed questions with fixed answers, a rule-based bot is enough. As soon as visitors ask freely or something needs to be done, the AI chat agent is the better choice.
We rewrote this post in October 2026. The previous version promised that an AI chat agent doubles your conversion rate, quoted a three times higher purchase probability and claimed that answers from a knowledge base are free of hallucinations. We had no reliable source for any of these claims, so they are gone. The number of languages was also wrong: EchoCall supports more than 30 languages. New in this version are a comparison table, a worked example with its assumptions stated, and an honest answer to the question of when you do not need an AI agent at all.
What is the difference between an AI agent and a chatbot?
The word chatbot covers very different systems today. For a decision about your website, it helps to separate three levels.
The rule-based chatbot. It works with a decision tree or a list of keywords. The US Consumer Financial Protection Bureau describes such bots like this: they use decision tree logic or a database of keywords to trigger preset, limited responses [1]. Typical examples are buttons such as "Opening hours", "Delivery" and "Returns". Someone wrote every answer in advance. If a visitor asks "Can I exchange the jacket at your Cologne shop instead?", the bot looks for a keyword. If it finds none, you get the familiar "Sorry, I didn't understand that".
The AI chatbot. It uses a language model and writes its own answers. It copes with typos, paraphrases and follow-up questions. But it only talks. It cannot look up anything that is not in its material, and it cannot trigger anything.
The AI chat agent. It also understands free language, but it has tools as well. OpenAI defines agents as systems that independently accomplish tasks on a user's behalf. Applications that integrate a language model but do not let it control the workflow are explicitly excluded, and the guide names simple chatbots as an example [3]. IBM draws a similar line: non-agentic chatbots lack tools, memory and reasoning, and they need fresh input from the user for every step [4].
On a website chat, this means the agent answers the exchange question from your returns policy, looks up the order in your shop if asked, and hands over to a person when the case gets tricky. If you want to go deeper into tools, planning and autonomy, read AI Agents vs. Chatbots: What's the Difference?. The definition itself is covered in What Is an AI Agent? Definition, How It Works and Examples.
AI chat agent vs. chatbot: the comparison in one table
| Feature | Rule-based chatbot | AI chat agent |
|---|---|---|
| Input | buttons, keywords, fixed phrases | free language with typos and follow-up questions |
| Answers | texts written in advance | written from your knowledge base |
| Maintenance | every new case is a new branch in the tree | keep documents and web pages up to date |
| Actions | links and forms | tools: book an appointment, open a ticket, look up an order |
| Typical failure | dead end without an answer | a wrong answer that sounds convincing |
| Predictability | complete, every answer is known | high with a good knowledge base, but never complete |
| Several languages | every language is written separately | one agent can answer in several languages |
| Cost model | effort goes mainly into building and maintaining | usually usage-based, per conversation |
The most important row is the typical failure. A rule-based bot fails visibly: the visitor notices at once that they are stuck. An AI chat agent fails quietly: it answers fluently even when the answer is wrong. Both failures can be managed, but with different means. More on that below.
When is a rule-based chatbot enough?
Not every website needs an AI agent. In its guide, OpenAI explicitly advises checking whether a use case clearly meets the criteria before building an agent. Otherwise, a deterministic solution may suffice [3]. Anthropic recommends finding the simplest solution possible and only increasing complexity when needed [5].
A click-through bot is usually enough when
- your visitors keep asking the same five to ten questions,
- the answers rarely change,
- the bot mainly needs to point people to the right page or a form,
- every answer has to be approved word for word, for example where wording is strictly regulated,
- nobody expects a real answer outside business hours.
An AI chat agent pays off when
- visitors ask in their own words and the variations no longer fit into a tree,
- the answer depends on data, such as the order status or free appointment slots,
- your content already exists in documents and on your website and should not be maintained a second time as a dialogue tree,
- you serve visitors in several languages,
- enquiries arrive in the evening and at weekends that would otherwise wait until Monday.
The same ground rule applies to both: there has to be a way out to a person. The CFPB describes so-called doom loops. Customers end up in continuous loops of automated answers without an offramp to a member of staff. This often happens when the issue falls outside the bot's limited capabilities [1]. A bot without handover annoys exactly the customers who need help most urgently.
The risks of an AI chat agent and how to limit them
A language model works with probabilities. The taskforce of the European data protection authorities described this in 2024 using ChatGPT as the example: due to the probabilistic nature of the system, the model may produce biased or made-up outputs. At the same time, end users are likely to take the outputs as factually accurate [6].
A case from Canada shows how expensive a wrong bot answer can get. Air Canada's chatbot told a customer he could book at full price and claim back the difference to a reduced fare within 90 days. That was not true. On 14 February 2024, the Civil Resolution Tribunal of British Columbia ordered the airline to pay the fare difference of 812 Canadian dollars, plus costs and interest. Air Canada had argued that the chatbot was a separate legal entity responsible for its own actions. The tribunal called this remarkable and rejected it. The airline is responsible for the information on its website [7].
The lesson applies to every bot, rule-based or AI: you are liable for what it says. Four measures limit the risk.
- A clean knowledge base. Retrieval-augmented generation, RAG for short, connects a language model to a searchable document collection [2]. The agent then answers from your material instead of general model knowledge. This lowers the risk of made-up answers but does not rule it out. The agent repeats an outdated price from an old PDF just as convincingly as a current one.
- Clear instructions. Define what the agent must not promise: discounts, goodwill gestures, legal assessments.
- Handover for tricky cases. OpenAI names two typical triggers for handing over to a person: repeated failure and high-risk actions such as cancellations or payments [3].
- Read the conversations. In the first weeks, someone should review the transcripts regularly and close gaps in the knowledge base.
How much does an AI chat agent cost compared with a chatbot?
With rule-based bots, the cost lies mainly in building and maintaining the dialogue trees. Every new question, every price change and every additional language means manual work. Market prices vary widely, and we know of no reliable average figure.
AI chat agents are usually billed by usage. At EchoCall you book a monthly package of chat conversations [9]:
| Chat conversations per month | Price (net) | Works out per conversation |
|---|---|---|
| 100 | €40 | €0.40 |
| 300 | €120 | €0.40 |
| 1,000 | €400 | €0.40 |
| 5,000 and more | €1,000 | €0.20 at 5,000 |
To try it out there is a trial: 30 voice minutes and 30 chat conversations for 14 days, and the credit card is verified with €1.
Since October 2026, visitors can also speak instead of type in the EchoCall widget. A notice card appears before the microphone starts. The minutes come from the voice allowance, which starts at €75 for 250 minutes, or €0.30 per minute.
Worked example: when the agent pays for itself
The following values are assumptions, not measurements. Use your own figures.
- Your website has 300 chat conversations a month, so the €120 package.
- The agent resolves half of them on its own. Your team takes over the other half in the Live Inbox.
- A member of staff would need four minutes on average for such a conversation.
| Item | Value |
|---|---|
| Conversations the agent resolves on its own | 150 |
| Staff time saved | 150 × 4 minutes = 600 minutes = 10 hours |
| Package price | €120 |
| Cost per conversation resolved by the agent | €120 ÷ 150 = €0.80 |
| Hourly cost at which the package breaks even | €120 ÷ 10 hours = €12 |
If an hour of staff time costs you more than €12, the time saved in this example covers the package price. If the agent resolves only 20 per cent of conversations, with all other assumptions unchanged, the saving drops to 4 hours and the threshold rises to €30 per hour. Not included are your time for setup and maintenance, and the value of enquiries answered in the evening instead of on the next working day.
How to switch from a click-through bot to an AI chat agent
- Collect the questions. Export what your old bot failed to understand, and read the emails and chat transcripts of the past few weeks. The 20 most frequent requests are your starting list.
- Fill the knowledge base. At EchoCall you upload PDF, TXT, Markdown or CSV files, add single URLs with a re-read interval, have the whole website crawled with a refresh interval, or write free text. Each agent has its own knowledge base.
- Start with one action. According to OpenAI, customers typically achieve greater success with an incremental approach. The guide recommends starting with a single agent and adding tools step by step [3]. Typical first actions are an appointment in Google Calendar, Outlook or Cal.com, a ticket in Zendesk or Freshdesk, or a look at an order in Shopify or Shopware 6.
- Decide on handover. Your team takes over the conversations the agent should not handle in the Live Inbox.
- Install it. The widget goes onto any website with one script snippet, and there is a plugin for WordPress. A domain whitelist defines which domains it may run on.
- Check the AI notice. Under Article 50(1) of the AI Act, providers must design AI systems that interact directly with people so that those people are informed they are dealing with an AI [8]. The EchoCall chat widget shows this notice automatically, in German, English and French.
- Test it. On the demo page you enter your website address. In about 20 seconds an agent is built from your public pages. You can write to it or call it in the browser, without signing up.
Frequently asked questions
What is the difference between an AI agent and a chatbot?
A traditional chatbot plays back pre-written answers, triggered by buttons or keywords. An AI agent understands free language, looks for the answer in a knowledge base and can use tools, for example to book an appointment or open a ticket. It decides for itself which step makes sense next, and hands over to a person when it gets stuck.
What is the difference between a chatbot and an AI chatbot?
An AI chatbot uses a language model and writes its own answers instead of playing back fixed texts. That is why it copes with paraphrases and typos. It only becomes an agent once it can act as well, meaning it has tools and controls the flow of a conversation itself.
How much does an AI chat agent cost compared with a chatbot?
Rule-based bots mainly cost staff time for building and maintenance. AI chat agents are usually billed per conversation. At EchoCall, 100 chat conversations a month cost €40, 300 cost €120, 1,000 cost €400 and 5,000 or more cost €1,000, all net. To try it out there are 30 voice minutes and 30 chat conversations for 14 days, and the credit card is verified with €1.
When is a rule-based chatbot enough?
When your visitors ask a few questions that are always the same, the answers rarely change and the bot mainly needs to point to pages or forms. Where every wording has to be approved in advance, a decision tree is also the safer choice. In every case, there must be a way out to a person.
Can an AI chat agent give wrong answers?
Yes. Language models can produce plausible but wrong answers. A well-maintained knowledge base, clear limits in the instructions and handover to your team for tricky topics lower the risk. It cannot be ruled out completely. Prices, deadlines and commitments should therefore only come from up-to-date material.
Can you talk to an AI chat agent as well?
With EchoCall, yes, since October 2026. The website widget has a microphone, and a notice card appears first. The minutes come from the voice allowance, which starts at €75 for 250 minutes.
Sources
- Consumer Financial Protection Bureau, Chatbots in consumer finance (6 June 2023): consumerfinance.gov
- Lewis et al., Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks (2020): arxiv.org
- OpenAI, A practical guide to building agents: cdn.openai.com
- IBM, What Are AI Agents?: ibm.com
- Anthropic, Building effective agents (19 December 2024): anthropic.com
- European Data Protection Board, Report of the work undertaken by the ChatGPT Taskforce (May 2024): edpb.europa.eu
- Moffatt v. Air Canada, 2024 BCCRT 149, case summary: wikipedia.org
- Regulation (EU) 2024/1689 (AI Act), Article 50: eur-lex.europa.eu
- EchoCall pricing: echocall.de/en/pricing
All links were checked on 6 October 2026. This post is not legal advice.
