Challenge

Variant creation is quite manual. Engineers lose time through repetitive tasks and redevelop many things that have actually already been developed. The result: long lead times and high costs.

Why is that?

Product knowledge lives in people’s heads and in tools—but it isn’t captured in a consistent, reusable way.

No comprehensive model

The link between customer requirements, the product, and the production process isn’t captured in a single, machine-readable model—so AI can’t use it.
As a result, new inquiries and variants—and the associated documents like routings/work plans and BOMs—can’t be generated automatically.

Heterogeneous systems

Variant engineering typically happens across a fragmented tool landscape—not in a single standard system.
Sales, engineering, and industrial engineering work in separate, disconnected tools from different vendors. It’s not unusual to see 150+ tools in one organization.

Limited automation

Information and documents are reviewed manually.
Data and parameters are copied between systems by hand.
Teams spend hours searching for similar components, bills of materials, and routing/work plans.

Why is that?

Product knowledge lives in people’s heads and in tools—but it isn’t captured in a consistent, reusable way.

No comprehensive model

The link between customer requirements, the product, and the production process isn’t captured in a single, machine-readable model—so AI can’t use it.
As a result, new inquiries and variants—and the associated documents like routings/work plans and BOMs—can’t be generated automatically.

Heterogeneous systems

Variant engineering typically happens across a fragmented tool landscape—not in a single standard system.
Sales, engineering, and industrial engineering work in separate, disconnected tools from different vendors. It’s not unusual to see 150+ tools in one organization.

Limited automation

Information and documents are reviewed manually.
Data and parameters are copied between systems by hand.
Teams spend hours searching for similar components, bills of materials, and routing/work plans.

Why is that?

Product knowledge lives in people’s heads and in tools—but it isn’t captured in a consistent, reusable way.

No comprehensive model

The link between customer requirements, the product, and the production process isn’t captured in a single, machine-readable model—so AI can’t use it.
As a result, new inquiries and variants—and the associated documents like routings/work plans and BOMs—can’t be generated automatically.

Heterogeneous systems

Variant engineering typically happens across a fragmented tool landscape—not in a single standard system.
Sales, engineering, and industrial engineering work in separate, disconnected tools from different vendors. It’s not unusual to see 150+ tools in one organization.

Limited automation

Information and documents are reviewed manually.
Data and parameters are copied between systems by hand.
Teams spend hours searching for similar components, bills of materials, and routing/work plans.

What’s causing this?

The idea of a seamless variant engineering process—and of systematically deriving product variants—isn’t new. Industry, software vendors, and research institutions have been working on it since the 1980s.

Yet in practice, many approaches have failed, or worked only in isolated cases. Why?

What is the Cause?

The idea of creating a seamless variant creation process and systematically deriving product variants is not new. Industry, software providers, and research institutions have been working diligently, since the 1980s to be precise, to achieve this exact goal.

In practice, many of these approaches have failed or only succeeded in isolated cases. Why?

What is the Cause?

In practice, many of these approaches have failed or only succeeded in isolated cases. Why?

Rigid rule sets

High upfront modeling effort, constant manual upkeep, and fast decay as products evolve—often without integration into day-to-day operational workflows.

Rigid rule sets

High upfront modeling effort, constant manual upkeep, and fast decay as products evolve—often without integration into day-to-day operational workflows.

Rigid rule sets

High upfront modeling effort, constant manual upkeep, and fast decay as products evolve—often without integration into day-to-day operational workflows.

Fragmented tool landscape

Hundreds of tools, siloed data models, and broken end-to-end continuity.

Fragmented tool landscape

Hundreds of tools, siloed data models, and broken end-to-end continuity.

Fragmented tool landscape

Hundreds of tools, siloed data models, and broken end-to-end continuity.

Custom-built integrations

Costly, long-running IT projects create brittle solutions that are hard to adapt.

Custom-built integrations

Costly, long-running IT projects create brittle solutions that are hard to adapt.

Custom-built integrations

Costly, long-running IT projects create brittle solutions that are hard to adapt.

What’s different today?

New technologies make a new approach possible.

Knowledge

Modern AI can capture and connect product knowledge across requirements, product data, and process steps. This creates a structured, machine-readable foundation for automation.

Systems

Instead of building one-off interfaces, you can connect existing systems through standard connectors and adapters. That reduces integration effort and keeps your tool landscape flexible.

Automation

Information can be analyzed, shared, and executed automatically.
Automate step by step—without a risky big-bang rollout.

The approach

By combining workflows, knowledge graphs, and AI, we enable scalable, sustainable automation of variant engineering.

01

Interfaces

A library of standard connectors for major engineering and planning tools makes it easy to integrate fragmented IT landscapes. This reduces the need for costly custom development and avoids lock-in to a closed, single-vendor platform.

02

Workflow automation

Workflows define what information is needed, how it is transformed, and where decisions stay with people. You can automate step by step—without coding. For complex tasks, AI agents can support or execute individual steps.

03

Knowledge graph

For AI to work reliably, you need a shared, machine-readable knowledge layer. We capture and link relationships between requirements, products, and processes across your systems—creating a knowledge graph that is understandable for people and usable by AI.

04

AI services

On top of this foundation, AI services can identify similar products and assemblies, recommend suitable manufacturing resources for a component, and generate assembly precedence graphs from historical work plans.

Industries

For companies building highly variant mechanical and mechatronic products.

Medical Devices
Industrial Machinery
Automotive
Energy
Defense
Aerospace

Where it delivers value

Where automating variant creation delivers major efficiency gains.

Sales

Example:

RFQ

Engineering

Example:

Variant configuration

Industrial Engineering

Example:

Routing creation

Where it delivers value

Where automating variant creation delivers major efficiency gains.

Sales

Example:

RFQ

Engineering

Example:

Variant configuration

Industrial Engineering

Example:

Routing creation

Where it delivers value

Where automating variant creation delivers major efficiency gains.

Sales

Example:

RFQ

Engineering

Example:

Variant configuration

Industrial Engineering

Example:

Routing creation

What teams typically save

What we see in practice:
Systematically deriving variants can unlock major efficiency gains across the entire development process.

Our industrial experience shows:
By systematically deriving variants, significant efficiency gains can be achieved throughout the entire development process.

The figures below summarize typical results we’ve observed across companies and industries.

The following key figures demonstrate our insights gained from companies in various industries with this approach.

-10%

Personnel costs

+15-30%

Revenue

-30%

Lead time

A platform for engineers

Model tasks and workflows in an easy-to-use editor—and automate them step by step. Use standard connectors for PLM, CAD, ERP, and Excel, and add capabilities like CAD similarity search, LLMs, or deterministic rule sets via drag and drop. Start small and scale automation across variant engineering—without a risky big-bang rollout.

Where can you get the biggest gains in variant engineering?

Start a conversation