Blog

Coming soon: Scienta connected to your AI tools via MCP

Written by Reinder Repko | Nov 16, 2026, 10:12:31 AM


That answer is right there in Scienta, but your AI assistant doesn't know it yet. That's about to change, because we're working on an MCP integration for Scienta. Soon, you'll be able to connect Scienta to your AI assistant, such as ChatGPT, Microsoft Copilot, Claude, or Gemini. The AI assistant can then consult the knowledge stored in Scienta. If you also use that assistant in other tools, such as Teams or Slack, you'll have Scienta's knowledge at hand there too. Keep in mind that this always works through an AI assistant that supports MCP – Scienta is not directly connected to every application you use. 

In this article, we'll take a look ahead. You'll learn what MCP is, what you'll be able to do with it, how to prepare for connecting your tools, and how security is handled.

In short

  • The MCP integration for Scienta is currently in development. The integration isn't available yet, but you can start preparing for it now.
  • MCP (Model Context Protocol) is an open standard that allows AI assistants to connect to other applications – and soon, to Scienta as well.
  • You'll soon be able to ask questions about policies, procedures, and work instructions directly from ChatGPT, Microsoft Copilot, Claude, or Gemini – without having to open Scienta itself. You'll receive an answer based on your organization's policies, along with a link to the source article.
  • The first version of the MCP integration focuses on reading articles. After that, we'll work on expanding its capabilities.
  • If you connect more systems, such as SharePoint, your AI assistant will combine information from multiple sources. You'll have to connect each system to your AI assistant separately, though.
  • The AI assistant only sees what you're authorized to read in Scienta. Your organization decides whether to enable MCP, and an administrator determines which groups get access.
  • Be aware of where your data is processed: with MCP, this takes place at the AI assistant provider's location, and the terms and conditions of both Scienta and that provider apply.

What exactly is MCP?

MCP stands for Model Context Protocol. It's an open standard that gives AI assistants a standardized way to communicate with other applications. Think of it as a universal plug: any application that supports MCP can be connected to any AI assistant that supports MCP. Without requiring a separate, custom integration for every combination.

For Scienta, this means your AI assistant will soon be able to consult your knowledge base whenever a question arises. Take the question: “What are our policies on working from home?” Without the integration, ChatGPT provides a general answer based on what's typical. With the MCP integration, the AI assistant looks up your organization's work-from-home policy in Scienta and bases the answer on that policy. It includes a link to the source article, so you can always verify where the information comes from.

How does this relate to Scienta AI Search?

With Scienta AI Search, you can already ask questions about the information in your Scienta environment. The main difference lies in where you ask those questions. AI Search works inside Scienta, while the MCP integration will soon bring Scienta to your connected AI assistant(s).

  Scienta AI Search MCP integration
Where do you ask your question? In Scienta, via the search bar or widget In your own AI assistant
What kind of information? Information from Scienta Scienta and possibly other linked sources
Which articles? Articles you're allowed to read Articles you're allowed to read
Where does processing take place? Within the secure Scienta environment At the AI assistant provider's location
What are the terms and conditions? Scienta's terms Those of Scienta and the AI provider

So it's not a matter of one or the other. AI Search is ideal for those who are already using Scienta. The MCP integration is the solution for employees who spend their day using an AI assistant or want to combine information from Scienta with other sources.

What will this mean for you in practice?

The biggest benefit: employees will soon be able to find the right answer right where they're already working. No extra tab, no searching through folders. Below are a few examples of what that will look like.

Instant answers to questions about your organization

An employee asks, “What steps do I need to follow to report an incident?” The AI assistant looks up the procedure in Scienta, summarizes the steps, and links to the article. Quick help, and the official source is always just one click away.

Knowledge in Teams or Slack

Do you use an AI assistant in Teams or Slack that supports MCP? Soon, you'll be able to retrieve information from Scienta right there, such as how to submit travel expense claims. The knowledge comes to the conversation, rather than the other way around.

Pointing the way to other systems

Does a Scienta article contain a link to a form on SharePoint? Soon, the AI assistant will be able to direct you to that location. The form itself doesn't need to be in Scienta. Scienta remains the place that explains how things work and where you need to go.

Analyzing and improving documentation

You'll soon be able to ask the AI assistant to compare procedures with updated legislation and identify potential areas for improvement. This saves policymakers and quality managers a lot of research work. However, it's important that you always evaluate the suggestions yourself and, if necessary, check the source article when making important decisions.

Reading is just the beginning

The first version of the MCP integration focuses on reading articles. Your AI assistant can look up information from Scienta, summarize it, and combine it with information from other sources. In this first version, you cannot yet edit or add anything in Scienta.

But it doesn't stop there. We're working to further expand the integration so that, in the future, you'll be able to do more with Scienta directly from your AI assistant.

Connect everything or just the most important tools?

The real power of MCP will soon lie in its ability to combine information. If your AI assistant has access to SharePoint, for example, in addition to Scienta, it can bring together information from both sources into a single response. Think of a procedure from Scienta, supplemented with the correct form or an up-to-date overview from another system.

That doesn't mean you'll have to connect everything all at once. You'll connect each system separately to your AI assistant – not to Scienta itself. The AI assistant is the hub that brings all sources together. This gives you the flexibility to make informed choices. Here are a few questions to start thinking about now:

  • Where do most questions come from? Start with the sources that employees consult most often. For many organizations, Scienta – with its policies, procedures, and work instructions – is the logical starting point.
  • Which combinations actually deliver results? Only integrate an additional system if it clearly improves the quality of answers – for example, because forms or documents are stored there.
  • Is the information reliable and up to date? An AI assistant is only as good as the sources it consults. Well-managed, up-to-date content yields better answers.
  • Which data can you share with the AI provider? The more you connect, the more important it becomes to decide, on a source-by-source basis, whether sharing that data with the AI provider fits within your policies.

Those who answer these questions now will be well-prepared when the integration becomes available. That way, you'll gradually build an assistant that truly understands how your organization works, without losing sight of the big picture.

How is security handled?

Sharing knowledge with an AI assistant sounds great, but of course you don't want confidential information to end up in the wrong hands. That's why security for the MCP integration in Scienta hinges on three things. 

Your organization is in control

The MCP integration won't be enabled by default. Your organization decides whether to activate MCP. If it does, an administrator grants access on a per-group basis. This allows you, for example, to start with a pilot group before rolling out the integration more widely.

Permissions remain the same

The AI assistant will only be able to access information you have read access to in Scienta. This follows the same principle as Scienta AI Search. If you're not allowed to open an article in Scienta, you won't be able to view it through your AI assistant either. The permissions you've already carefully set up remain fully in effect.

Know where your data is processed

This is the most important difference between Scienta AI Search and the MCP integration. With AI Search, processing stays within the secure Scienta environment. With the MCP integration, processing is handled by the provider of your AI assistant, such as OpenAI, Microsoft, Anthropic, or Google. This means the terms and conditions of both Scienta and that AI provider apply.

So it's worth checking what you've agreed to with the provider of your AI assistant. Do you have a business license? And what do the terms and conditions say about the storage and use of data? Involve your IT, security, or privacy officer in this assessment. That way, you can be sure that the integration, once it becomes available, will align with your organization's policies.

Bringing your knowledge to where the work happens

With the MCP integration, Scienta will soon be more than a place where you look things up. Your knowledge will be available in the AI assistant your employees already use, whether that's in their browser, in Teams, or in Slack. They'll get answers based on your own policies and procedures, with a link to the source included. And the more systems you connect, the more complete the answers your assistant can give.

At the same time, you stay in control. You decide whether to enable MCP, who gets access, and which sources to connect. What someone is allowed to read in Scienta still determines what the assistant can show them. The only thing you need to arrange in advance is a clear agreement with your AI provider.

The first version focuses on reading articles, and we'll build on that. Keep an eye on the WoodWing blog, where you'll find out how to get started as soon as the MCP integration is available. Naturally, we'll also keep you posted through our newsletter.