PIPELINEGuide
AI knowledge base: ongoing costs and operation
The project price is only part of the calculation. What actually arises in effort and costs after going live — and which item is most often forgotten.
After going live, AI knowledge base typically requires a few hours of maintenance per quarter, plus hosting, licenses, and third-party usage fees. The most commonly forgotten item isn't the money, but the responsibility: systems become outdated because no one is assigned, not because maintenance is demanding.
01 — Operation
What ongoing operation means
After the settling-in phase, the effort is low, but not zero. For AI knowledge base, that's typically a few hours per quarter: additions when something changes in your offering, a look at the conversation logs, occasional corrections.
More important than the scope is the responsibility. Systems don't become outdated because maintenance is demanding, but because no one is assigned. One person with a recurring calendar appointment solves this problem completely.
What additionally arises on an ongoing basis depends on the setup — hosting, licenses and third-party usage-based fees belong in the same calculation as the implementation. We show this separately in the quote instead of hiding it in the project price.
02 — Factors
What drives the price
- Scope and number of integrations. Every additional system that needs to participate costs effort — regardless of how simple the core is.
- State of your starting point. Clean data, existing content and documented processes significantly reduce effort. Chaos is the most expensive cost factor.
- How much new material needs to be created. What you can contribute doesn't need to be invented by anyone. This is the biggest lever you have in your own hands.
- Requirements for design and customization. A solid standard system costs a fraction of a custom solution — and is sufficient more often than vendors admit.
- Operating model. One-time project costs or ongoing support: calculated over three years, this often gives a different picture than the quote comparison.
03 — Contract
What needs to be clarified in writing
- Ownership of data and access. They belong to the client. If a provider won't clearly confirm this, that's the most important information from the conversation.
- Handover upon termination. In what format, within what timeframe, at what cost.
- No minimum term without consideration. Lock-in is justifiable if something is delivered in return — otherwise not.
- What a later change costs. The total cost isn't determined by the project price, but by the hourly rate in year three.
- Documentation as a deliverable. Not as a promise, but as a line item in the proposal.
04 — Questions
Frequently asked questions
Our documents are a mess. Will it still work?
To a certain extent, yes — the system handles different formats and structures. What it cannot do is resolve conflicting information. If three versions of a price list are in circulation, you need to clarify which one applies first.
What if we have multiple languages in our document base?
No problem — the search understands questions and documents regardless of language and finds the right answer even when the question and source are written in different languages.
Does this have to run in the cloud?
No. For sensitive areas, we use models that run entirely on your own hardware. That costs more to set up and slightly reduces answer quality, but no document leaves the building.
How long does setup take with a large document base?
It depends less on volume than on condition: well-structured repositories with few duplicates are ready in a few weeks, grown repositories with many old versions need a cleanup round first.
What is RAG?
The abbreviation stands for retrieval-augmented generation. Instead of having the language model answer from memory, your documents are searched first, and only the passages found are used to generate the answer. That's why the answer has a source.
Will everyone see everything?
No. The permissions from your existing systems are carried over. If someone isn't allowed to open a folder today, they won't get answers from it either.
Can we roll this out to more departments step by step?
Yes, that's the usual approach. Start with one department, learn from their questions and results, then connect the next one — rather than releasing everything at once.
05 — Next
Related pages
- AI Knowledge Base — Overview
- AI knowledge base — costs
- AI knowledge base — measuring success
- AI knowledge base — how it works
- AI knowledge base — pros and cons
- AI knowledge base — provider comparison
- AI Knowledge Base — Requirements
- AI knowledge base — where to start
- AI automation
- Document processing with AI
- AI chatbot
- All guides
Next step
Let's talk briefly about whether this makes sense for you.
A twenty-minute conversation is usually enough to clarify whether a project makes sense. If not, we'll say so — that's cheaper for both sides than an offer nobody needs.