
Which AI Type Are You?
AI doesn’t just optimize work. It can transform how engineering organizations create value.
Today, hardly any industry can ignore the question of how to use Artificial Intelligence (AI) effectively.
But when we talk about AI in engineering, we don’t all mean the same thing.
For some, AI is a tool that improves documents. For others, it sorts requirements, takes over boring process-related tasks, or generates entire software solutions. And some expect AI to independently develop new ideas in the future.
That’s why I distinguish six AI types for engineering organizations.
1. The AI Optimizer – Better, Not Different
Value creation continues to follow the established process. AI improves the final output or supports individual steps along the way.
Typical applications include:
- Optimizing documents and reports
- Summarizing meeting minutes
- Creating presentations and management reports
The benefit: an easy entry point with visible results.
The limitation: the underlying engineering process remains unchanged. If approvals, coordination, or decisions take too long, a better report won’t make the process any faster.
2. The AI Sorter – Bringing Order to Complexity
The number of requirements, changes, and information sources continues to grow. The challenge is to identify relevant information quickly and make relationships visible.
AI can pre-sort requirements, identify changes, cluster related topics, and structure stakeholder interests.
The benefits:
- Less manual analysis
- Faster identification of changes
- Better preparation for decisions
- Earlier detection of conflicts and dependencies
The limitation: sorting is not decision-making. Priorities and trade-offs still need to be resolved.
3. The AI Connector – Making the Connections
In process-compliant engineering, requirements, architecture, implementation, tests, and approvals need to be traceably connected.
These relationships are essential for quality assurance, troubleshooting, change impact analysis, and verification. At the same time, maintaining them is often perceived as boring documentation work by engineers.
AI can help establish relationships between engineering artifacts, identify missing links, maintain traceability, and track changes across multiple artifacts.
It can also support configuration and change management.
The benefit: less manual maintenance, fewer gaps, and a more reliable basis for engineering decisions.
One important condition: AI must not only create links but also make their technical correctness and origin verifiable.
4. The AI Monitor – Securing Quality Earlier
Engineering continues to follow established processes, while AI increasingly takes over tasks related to checking and verifying results.
Examples include:
- Deriving test cases from requirements
- Identifying test gaps and inconsistencies
- Analyzing test results
- Summarizing failure patterns
- Suggesting additional tests for modified functionality
This can bring verification activities further upstream and integrate them more closely into the engineering process.
The critical point: AI-generated tests are not automatically good tests. Their quality must be independently verified. Otherwise, we risk automating nothing more than the confirmation of our own mistakes.
5. The AI Creator – From Supporting Development to Doing the Development
Here, the division of work changes fundamentally.
The engineer specifies the expected behavior, constraints, and quality requirements. AI increasingly takes over the independent creation of designs, architectural building blocks, implementations, and tests.
The engineer shifts focus from manually creating artifacts to specifying, evaluating, and verifying the results.
The potential: greater engineering output, shorter lead times, and faster implementation of changes.
The challenge: the more AI takes over, the more important clear specifications, suitable system architectures, independent verification, and clearly defined responsibilities become.
Writing less code does not mean taking less responsibility for the outcome.
6. The AI Innovator – Discovering New Possibilities
This type takes things a step further.
AI does not simply implement given instructions. It develops its own proposals, identifies untapped opportunities, and challenges established solutions.
For example, it could:
- Propose alternative architectures
- Develop new approaches to known problems
- Identify trade-offs between cost, performance, and quality
- Identify opportunities to improve existing engineering processes
Humans evaluate, prioritize, and decide which ideas to pursue.
The potential value is no longer limited to efficiency. It extends to new solutions and additional value creation.
Of course, the same principle applies here: a plausible suggestion is not yet proven progress.
Which AI Type Delivers the Greatest Business Value?
The answer does not depend on which AI type sounds the most advanced.
What matters is where the biggest bottleneck lies in your engineering organization.
Are we losing time on documentation? Change analysis? Missing traceability? Verification? Or does it take too long to turn an idea into a reliable product?
And above all: What measurable improvement are we trying to achieve?
- Time-to-market: Can engineering decisions and results be achieved faster?
- Productivity: Can we deliver more usable engineering output with the same resources?
- Quality: Can we detect defects earlier and reduce rework costs?
- Predictability: Can we make outcomes more predictable and dependencies more transparent?
- Innovation: Can we develop solutions that we would not have considered before?
AI should never become an end in itself. What matters is whether it improves how the engineering organization creates value.
The six types do not have to be introduced sequentially. An organization may already be using AI to generate code while still missing significant opportunities in requirements analysis or traceability.
One final thought: Every AI type requires suitable engineering processes.
If we want to unlock more of AI’s potential, introducing the technology alone is not enough. We also need to adapt the way we develop.
My question to you:
Which AI type is currently most prevalent in your engineering organization – and where do you see the greatest untapped potential?
#ArtificialIntelligence #SoftwareDevelopment #EngineeringManagement #DevelopmentProcessImprovement #Automotive
