Optinex is now being further developed to act as an AI agent within the customer’s existing environment.
“Extremely complex functionality is presented in a very simple way to the user,” says Johan Eriksson, CTO at ForestX.
Optinex is ForestX’s planning and optimization solution for sawmills and wood-processing companies. The development team, led by CTO Johan Eriksson, is now taking the solution one step further by adding an AI agent.
A key part of the development is the Model Context Protocol (MCP). MCP makes it possible to connect AI agents to Optinex functionality and data in a structured way. Another central component is Skills: functions and capabilities from Optinex that can be published directly to the customer’s AI agent.
The idea is that the same Optinex backend can be used through multiple interfaces. Customers can work directly in the Optinex web app or access Optinex functionality through the AI agent they use in their daily work.
“It is not a single AI agent that solves the entire workflow. It is the interaction between MCP servers from different vendors that makes this possible. Optinex contributes its specialist expertise in planning and optimization,” says Johan Eriksson.
“In the future, this can be seen as a new user interface. We can let users work in natural language,” says Johan Eriksson, continuing:
“Extremely complex functionality is presented in a very simple way to the user. We let the AI agent help users access information that can sometimes be difficult to interpret. This is no longer just about retrieving data. We are taking it one step further and can, for example, proceed to create an order.”
How does Optinex’s AI functionality work in practice?
Imagine you need to answer the question: When can I promise delivery of 42×120×4200 CLS to Washington, USA?
Ask Optinex’s AI agent and receive an immediate answer based on current capacity and lead time.
The agent can then:
- interpret the product, quantity and destination,
- check production capacity and lead time through Optinex,
- calculate and confirm an exact delivery date.
In this example, the user can then ask their AI assistant to create the order.
This means that the AI agent does more than simply retrieve information. It can assist the user through several steps in the process – from interpreting a question and analyzing available information to providing a concrete basis for decision-making and, when requested by the user, taking action.
Finding alternative products
Another use case is finding the best product alternatives.
If the requested product is not in stock, the AI agent can search for alternative products that can be delivered within the desired timeframe.
For example, the user could ask:
“What are the best alternatives to 42×120×4200 CLS that can be delivered within two weeks?”
The agent will then:
- identify the requested specification,
- search current inventory and available production options,
- look for equivalent or nearby dimensions and grades,
- apply the customer’s specific rules,
- rank the alternatives based on availability and lead time.
This means the agent can help find a solution even when the exact product is unavailable.
From data to decisions
The technical architecture is designed to support multiple customers and AI environments. Optinex MCP servers and services run in a multi-tenant backend, while each customer’s data remains isolated in its own database. All interactions are authenticated and communication is encrypted.
This allows ForestX to publish Optinex functionality once and make it available across different AI environments.
“Optinex should be able to become part of the customer’s existing AI environment. It shouldn’t matter whether the user is working in Optinex, SAP Joule or another AI agent – Optinex should be available wherever it is needed,” says Johan Eriksson, CTO at ForestX.
