Can you say today whether the plan will hold up for your next production run?
Example
Anja Wörner, COO at Brenntal Antriebstechnik (example; company name fictitious), must determine four weeks before the start of production for the next model series whether the deadline can be met, and she compiles the status data from three systems and two spreadsheets. Could she answer that question as early as Monday morning?
For mechanical engineering and the manufacturing industry: AI becomes part of your daily routine when the workflow takes precedence over the tool. That way, you’ll know sooner whether your plan will work. To achieve this, we’re introducing an operations system: workflows, rules, and agents on your existing tools—no new software required. Agents collect data and perform calculations; your people make the decisions.
Example
Adoption and Challenges, According to the VDMA 2026
AI is particularly well-suited in mechanical engineering in situations where people gather information from multiple sources before making a decision. According to a February 2026 VDMA survey, 31 percent of the companies that responded to the AI question (223 out of 248) report using AI, while 37 percent are involved in pilot projects. The same survey cites IT security (60 percent) as the top obstacle to digital transformation, followed by these three:
Change Management, 47 percent. A pilot project is initiated by a dedicated individual, using their data, to address their specific task. For the team to incorporate it into their daily work, they need the necessary support and a designated owner for the process.
Implementation speed: 44 percent. A pilot project based on copies is already outdated on the first day. AI must read from where the work already exists.
Lack of human resources, 42 percent. Those who compile status updates from multiple sources don’t have time for what only humans can do: make decisions. An agent—an AI assistant with a specific assignment—can handle the data collection.
Getting Started in Practice
This is where people today compile bills of materials, change status, and delivery commitments from multiple sources. An agent can gather and prepare the information, while your staff reviews it and makes decisions.
Why Pilots Rarely Feel at Home in Everyday Life
Here's how it works for us
A Common Pitfall
In engineering, no system is commissioned until the foundation, power supply, and quality plan are in place. With AI, on the other hand, the machine is often ordered first, and the question of who will operate it on a day-to-day basis comes later.
Our Approach
At Lighthouse, organization comes first: a structured filing system, a registry of all projects, and clear rules. Only then do agents cooperate.
Each agent has a specific mission, specific rights, and a protocol. It gathers information, performs calculations, and prepares options.
Nico Röpnack gives the go-ahead before anything is released to the public. Anything that stands out during this process is incorporated back into the rules, and each round gets better.
An example from the order processing workflow
From the initial request to final approval: Agents collect, calculate, and verify data using your existing tools. The process only continues after approval.
Provider Information
In November 2026, Anthropic introduced a plugin for manufacturing that includes predefined workflows for tasks such as quotes, inspection discrepancies, and malfunctions. According to Anthropic, every workflow ends with a human review. What your company needs to add: the workflow, the rules, and a team to implement them. Source: Anthropic, Claude for Manufacturing
Smartsheet Customer
A manufacturer of compressed air treatment systems and a Smartsheet customer of Lighthouse. Read the case study to learn how the team experiences collaboration.
Customers in the mechanical engineering and industrial sectors
These companies know us from Smartsheet projects: consolidating data, improving processes, and managing projects. This is the foundation of our work with agents: Since March 2026, Smartsheet has been making its data available via an MCP server to AI assistants like Claude (according to Smartsheet).

Frequently asked questions
Answers to questions asked by executives and COOs in the mechanical engineering industry.
Documents
The "aha" moment came when I realized that Smartsheet lets us achieve transparency regarding projects, tasks, responsibilities, and deadlines much more quickly. This helps us tremendously, especially when we’re running multiple projects at the same time.
Resource planning with Smartsheet at Knocks Fluid-Technik in Selm, as told from the project manager's perspective.
Read the customer story.0 out of 5 stars from 37 Google reviews for Lighthouse Consultings. Read all reviews on Google. As of November 2026
Your contact person
Nico Röpnack has a background in engineering: over 20 years in the industry, ranging from start-up planning to leading global engineering teams.
Not quite ready for a conversation yet? Lighthouse Live at the Academy: free live webinars via Zoom, each featuring half an hour of hands-on practice.
Next step
On Monday morning, Anja Wörner wants to know if the production ramp-up is on track. During the initial consultation, we'll determine which process comes first for you: one goal, one area, then the next.
Select an appointment directly, free of charge and without obligation.
Get started for free with n8n, the open-source automation platform.
Briefly describe your situation. We will get back to you within 24 hours.