Optional analytics

With your permission, we use Google Analytics and Vercel Analytics to count visits and improve the website. We do not use this data for advertising or send form details. Cookie policy

Skip to content
Back to blog
Hiring4 min read

AI chatbots for B2B: cost, scope and implementation

Three levels of business chatbot, what drives the cost of each and how to compare automation with human support using the same scope.

ByAsier Mugica· Founder of Ace Digital· Updated 2026-09-18

"How much is a chatbot?" is roughly as answerable as "how much is a vehicle?". The word covers a decision tree that answers five questions and an agent that reads your documentation, checks your calendar and books the meeting. They cost very different amounts and they solve very different problems. This is how to tell which one you actually need before anyone quotes you.

The three levels, and what separates them

The table separates capability and cost drivers. It is not Ace Digital pricing or a market benchmark.

LevelWhat it doesMain cost drivers
1. Scripted FAQAnswers a fixed set of questions from a decision tree.Journeys, channels, languages and content updates
2. Retrieval chatbot (RAG)Searches approved documentation and answers varied wording.Document quality, access rules, evaluation and model usage
3. AgentAnswers and takes approved actions in connected systems.Integrations, permissions, error handling and audit logs

When a retrieval chatbot is a sensible starting point

A scripted FAQ works when the questions and routes are few and stable. A retrieval chatbot may fit when users ask the same underlying questions in varied language and the approved answer already exists in maintained documentation.

  • It can use documentation you already own, but conflicting or outdated material must be resolved first.
  • It can refuse or hand over when confidence is low if that behaviour is designed and tested.
  • It can expose the source used for an answer, which supports review but does not prove that the answer is correct.

What it looks like in practice

An illustration, not a case study. A guest messages a small hotel at midnight asking whether there is a room next weekend, what it costs and whether dogs are allowed. The chatbot answers the policy question from the hotel's own documentation, checks real availability and price, replies in the guest's language with a direct booking link, and passes anything it should not decide on its own to reception with a summary. That is levels 2 and 3 working together. Ace Reception 24 covers the answering part for independent hotels: it replies from a knowledge base the hotel approves and sends the direct booking link; checking live availability means connecting the booking system, which is quoted separately.

Ace Reception 24 has a published price of USD 1,990 setup. The first month is included; USD 290 per month starts on the day-30 report, with no minimum term. Connecting a booking system is quoted separately. This is the price of that defined hotel offer, not a general chatbot benchmark.

The comparison against hiring

The honest version of this comparison has to include the things a person does that a chatbot does not, so here it is both ways.

  • A chatbot has implementation, usage, maintenance and review costs. Model all four at the expected conversation volume.
  • A person provides judgement, empathy and flexibility beyond the conversations counted in the model.
  • Handover does not remove human work. Estimate the share of conversations a person will review or resolve.
  • Compare twelve-month cost and service coverage using observed enquiry volume. At low volume, a better FAQ or contact process may be enough.

How the implementation actually goes

  • Discovery: gather maintained documentation and a representative sample of real questions. Decide what the bot must not answer on its own.
  • Build: connect only the required channels and design refusal and handover before adding optional actions.
  • Evaluation: test against historical questions, edge cases and conflicting sources, with written acceptance criteria.
  • Pilot: monitor conversations, corrections, handovers and user feedback before expanding the audience or scope.

What to insist on before you sign

  • The guardrails in writing. Which topics it refuses, and what it says instead.
  • Full access to the configuration and the sources, so you are never locked in.
  • A named handoff path. A bot with no escape hatch is a customer service problem waiting to happen.
  • A billing model tied to a written scope, with change control and ongoing costs stated separately.