Automotive engineering gap: Why AI must move beyond the hype

Automotive manufacturers are cutting engineering roles, reducing model portfolios and rethinking production capacity.

At the same time, they are investing in software, artificial intelligence and new product-development methods.
‍

This is not simply another cost-reduction cycle.
‍

It is a response to a structural problem: traditional automotive engineering is no longer moving at the speed of the market.
‍

Established manufacturers still employ some of the world’s most experienced engineers. They own decades of validated designs, advanced R&D infrastructure and extensive product knowledge.

‍

The problem is not talent. It is the time required to transform that knowledge into a validated, production-ready vehicle.

‍

Automotive manufacturers are under pressure to move faster

‍

Across the industry, major automotive groups are reorganising engineering teams, simplifying product portfolios and concentrating investments on software, electrification and AI.

‍

These decisions reflect growing pressure on margins, increasingly complex vehicles and competition from manufacturers operating with significantly shorter development cycles.

‍

Some newer automotive players can now bring a vehicle from concept to launch in roughly half the time traditionally required by established manufacturers.

‍

That gap changes more than the launch date.

‍

A faster manufacturer introduces new technologies sooner, gathers customer feedback earlier and begins improving the next vehicle generation while slower competitors are still validating the current one.

‍

Speed becomes a learning advantage. And that advantage compounds with every product program.

‍

The development benchmark has changed

‍

Automotive development was traditionally organized around long, sequential phases.

‍

Requirements were defined first. Architectures followed. Detailed design, simulation, validation and industrialisation came later. Information moved from one department to another through formal handovers and scheduled reviews.

‍

That model becomes increasingly difficult to sustain when vehicles combine mechanical systems, electronics, batteries, software and rapidly evolving customer features.

‍

Today, manufacturers must explore more configurations, coordinate more engineering domains and validate more interactions—while reducing overall development time.

‍

The obvious response is to invest in AI. But investment alone does not create faster engineering.

‍

Why more AI will not automatically solve the problem

‍

Automotive companies already possess enormous amounts of engineering information:
‍

• Native CAD models and technical drawings;
• PLM records and product configurations;
• System requirements and specifications;
• Bills of materials;
• Simulation and test results;
• Design and verification rules;
• Knowledge from previous vehicle programs.

‍

The difficulty is that these assets remain fragmented.

‍

A requirement may be stored in one platform while the geometry that satisfies it exists in another. Product structures are controlled in PLM, but verification rules may remain in spreadsheets, internal standards or the experience of senior engineers.

‍

When a design changes, engineers must determine manually which requirements, components, documents and validation results are affected.

‍

A generic AI assistant can retrieve a document or summarise a specification.

‍

It cannot reliably determine whether a component meets a new technical requirement, whether a 3D design respects company-specific rules or whether a geometry change has created inconsistencies across CAD, BOM and PLM data.

‍

Automotive engineering needs AI that understands the product itself—not only the documents describing it.

‍

What is Engineering Intelligence?

‍

Engineering Intelligence connects engineering data with the technical knowledge required to interpret and validate it.

‍

It brings together:

‍

• CAD geometry and assemblies;
• PLM data and product structures;
• Requirements and specifications;
• BOMs and component information;
• Engineering rules and validation logic;
• Previous technical decisions and test results.

‍

Instead of treating these assets as separate files, Engineering Intelligence creates a connected technical context.

‍

This allows AI applications to address questions engineers actually need to answer:

‍

• Which requirements influenced this architecture?
• Which existing components satisfy the new constraints?
• Has a comparable design already been validated?
• Which company rules apply to this configuration?
• What will be affected if a requirement or geometry changes?
• Are the CAD model, BOM and PLM configuration consistent?

‍

This is the missing layer between the engineering information manufacturers already own and the development speed they are trying to achieve.

‍

How Dessia helps automotive manufacturers move faster

‍

Most automotive manufacturers don't have a knowledge problem. They have a speed problem — and the knowledge that could solve it is already sitting inside their own CAD files, PLM systems, BOMs, and validation archives, locked away from the design engineers who need it most. Dessia unlocks it. No ripping out CAD, PLM, or simulation tools, no multi-year IT overhaul — just the connective layer that turns scattered engineering knowledge into speed, starting now.

‍

The unlock is CAD-aware AI: it doesn't just see geometry, it understands why that geometry exists — the design intent, the requirements it satisfies, the constraints it respects, the validation it's already passed. Two parts can look identical and mean two completely different things. Dessia tells engineers which is which, instantly, ending the guesswork that turns design reuse into rework.

‍

The payoff shows up everywhere it matters:

‍

• Faster programs. Catch design and compliance issues before physical validation, not after — the single biggest lever for compressing a development cycle.
• Less rework. Reuse validated designs with confidence instead of re-engineering from zero on every program.
• Knowledge that survives restructuring. As engineering headcount shrinks across the industry, expert knowledge stops walking out the door with retiring or reassigned engineers.
• Verification on autopilot. 2D drawings and 3D models get checked against business rules automatically, freeing engineers to actually design.
• Fewer surprises downstream. Inconsistencies across CAD, BOM, and PLM data surface before production, not during it.
• More options, same team. Explore more design alternatives without adding a single headcount.

‍

This is how engineering scale finally turns into engineering speed — how the R&D billions already being spent stop funding slowdowns and start funding a real competitive edge.

‍

The real transformation is happening inside engineering

‍

The restructuring now visible across the automotive industry is a symptom of a deeper shift.

‍

Scale alone is no longer enough.

‍

Large engineering teams, extensive software stacks and decades of product data do not automatically produce faster development. Their value depends on how easily knowledge can move between systems, teams and vehicle programs.

‍

Reducing costs may improve short-term performance.

‍

It does not shorten a validation cycle, preserve an expert decision or connect a requirement to the geometry it affects.

‍

Engineering Intelligence addresses that structural problem.

‍

It applies AI directly to the product, its architecture and the technical rules that determine whether a design is valid.

‍

The automotive industry already owns the knowledge required to move faster. Dessia turns that knowledge into design decisions, earlier validation and measurable engineering speed.

‍

FAQ

‍

How is Engineering Intelligence different from a generic AI copilot?

A copilot retrieves and summarizes documents. Engineering Intelligence reasons over the product itself. It can tell you whether a specific component has already been validated against a requirement, not just what a specification document says about it — the difference between search and engineering judgment.

‍

Why isn't more AI investment enough to fix slow vehicle development?

Because speed isn't a model problem, it's a data problem. Manufacturers already own the CAD models, PLM records, BOMs, and validation history they need, but it's fragmented across disconnected systems, so AI layered on top of it can only summarize, not verify. Connecting that data comes first; a bigger AI budget without it just automates the same delays faster.

‍

What is CAD-aware AI, and why does it matter?

CAD-aware AI reads geometry together with the engineering context behind it, instead of treating a 3D model as just a shape. Two parts can look identical and serve completely different purposes; CAD-aware AI is what tells them apart automatically, closing a gap generic AI tools can't.

‍

How does Engineering Intelligence relate to the digital thread?

The digital thread connects engineering data across a product's lifecycle; Engineering Intelligence makes those connections usable by AI. One links systems together, the other lets AI reason across those links to answer real engineering questions, not just trace data.

‍

Is Engineering Intelligence specific to automotive, or does it apply to other industries?

The automotive industry is the clearest current example, but the underlying problem applies to any industry with complex, long-lifecycle product development, from aerospace to industrial equipment.

‍