Blog · Published · Updated
Company Brain AI: Ready Platform or Custom System?
Short answer
A ready platform wins when knowledge lives in one ecosystem, for example Microsoft 365 or Confluence, and the questions concern that content. Your own knowledge base wins with many sources, strict permissions or data that may not leave the company. For 120 people over 24 months, licences at $21 cost €53,976, and your own system from €3,000 with the model costs €4,748, before maintenance, which we price per project.
How we workDetails ↓
- 01Call30 minutes with an engineer, free
- 02PrototypeA working system on your process and data, free
- 03ImplementationTarget and fixed price in the contract; 50% at signing
- 04Acceptance50% once the agreed target is met
- net per application, fixed price before contract
- from €3,000
- from signing the contract to acceptance
- 4–10 weeks
- free prototype on your documents
- €0
We belong to the OpenAI and Anthropic partner programmes and learn through their official courses. What this means ↓
Author: Syntalith · 3 min read
Statistics
AI knowledge platforms: statistics 2026
€49,228
difference over 24 months for 120 people: licences at $21 a month against your own knowledge base with the model, without maintenance, which we price per project. For 25 people the difference is €7,881. Scenario.
5 May 2026
According to Microsoft (2026), federated Copilot connectors, which fetch data over MCP at the moment of the question, have been generally available since 5 May 2026.
25
According to Atlassian (2026), the Standard plan of Jira and Confluence includes 25 Rovo credits per user a month, Premium 70 and Enterprise 150.
42%
According to the European Commission (2026), 42% of EU SMEs and mid-caps already use AI or algorithmic management, mostly in a limited and experimental way.
Source: European Commission, Flash Eurobarometer 578 (2026) ↗
49%
According to Anthropic (2024), combining contextual embeddings with BM25 search cut failed retrievals by 49%, and by 67% with reranking.
Figures checked: 22 October 2026
A map of ready platforms in 2026
| Tool | For whom | Pricing model | When it wins |
|---|---|---|---|
| Microsoft 365 Copilot | companies on Microsoft 365 | $21 per user a month on the Business plan | knowledge in SharePoint, Teams and Outlook |
| Google Gemini Enterprise | companies on Google Workspace | per-user editions published by Google | documents, email and drives in Google |
| Atlassian Rovo | teams in Jira and Confluence | in paid Cloud plans, 25 to 150 credits per person | product knowledge in Confluence, tickets in Jira |
| Onyx | companies with their own IT | open source, MIT licence, you pay for the server and upkeep | data may not leave your infrastructure |
| Your own knowledge base | many sources, strict permissions | from €3,000 for the build, upkeep, model | sources from different systems and an answer with a quotation |
Sources of the descriptions: Microsoft, Google and the Gemini Enterprise editions, Atlassian, Onyx.
Since 5 May 2026, Microsoft has offered federated connectors that fetch data from other systems over MCP at the moment of the question, with the user's permissions and a link to the source. That narrows the advantage of your own systems for popular SaaS apps. The advantage stays where the source is an internal system, a folder on a server or a database with access rules no ready connector knows.
Knowledge in one ecosystem, questions about that content
that ecosystem's ready platform
a per-user licence
Data may not leave the company, you have your own IT
Onyx or your own system on the company's server
server and upkeep
Many sources, internal systems, strict permissions
your own knowledge base with quotations from the source
from €3,000
What an AI platform or a custom knowledge base costs on your numbers
- Platform: 120 × $21 × 24 months
- $60,480 (€53,976)
- Your own system: build from €3,000
- Model: 120 × 40 questions × 24 months, Claude Sonnet 5.5
- €1,748
- Your own system in total
- €4,748
A difference of €49,228 over 24 months; for 25 people €7,881
The number of people, the questions and the horizon are the scenario's assumptions, Maintenance is priced per project, so the scenario leaves it out. The Copilot Business rate with an annual commitment is Microsoft's, with a promotion of $18 in the first year until 31 December 2026. Model: 6,000 input tokens and 500 output tokens per question at Anthropic's rates; dollar amounts converted at the ECB rate of 1.1205 USD per euro on 22 October 2026. Copilot also does things beyond answering from documents, so the comparison concerns that function only.
Two items are easy to miss on the build side: tidying the documents before the start, and the time of the person who approves changes to the sources. On the platform side it is easy to miss that the number of users grows and the rate rarely falls. According to the European Commission (2026), 42% of EU SMEs and mid-caps already use AI, mostly in a limited and experimental way, so seats that nobody uses are a real risk.
When your own knowledge base wins
- An answer needs sources from three or more systems, internal ones included
- HR, finance or the board hold documents others may not see, even in part
- Data has to stay in your infrastructure or in a chosen region
- The answer has to quote a specific version of a procedure and its approval date
- The team asks questions in its own words, and keyword search fails
You can measure retrieval quality before you buy. According to Anthropic (2024), combining contextual embeddings with BM25 search cut failed retrievals by 49%, and by 67% with reranking (Contextual Retrieval). Run the same test on the platform and on a prototype of your own system: 50 questions the team has already asked, and the source that should appear.
What to ask a supplier for before you sign
Ask for a test on your own questions, in place of a presentation. Ask what happens to an answer after an employee loses access to a folder, how long the index remembers a deleted document, and whether the answer shows the version of the source. We ask ourselves the same questions when we build, as the post about a knowledge base after permissions change describes.
We build your own knowledge base as part of our AI apps, from €3,000 net. If the test shows that a ready platform is enough, we will say so in the free analysis, and a Copilot rollout for the team helps you put it in place.
Worked example · Scenario
How we count the payback: answers from company documents
We run the same count for your documents in the free analysis.
€3,723
a month for finding answers to 800 questions in documents
Payback: 1 mo.
Our assumptions are 8 minutes to find an answer in documents or ask a colleague, and checking in 15% of the old time. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat, for companies with 10 or more employees. Model cost: Claude Sonnet 5.5, about 6,000 + 500 tokens per answer from documents, at the ECB rate of 1.1205 USD per euro on 22 October 2026.
See the full calculation →
