AI

Grafik zeigt die Texte „Services“ und „SOFTWARE FIRST“ über einer geschwungenen Linie mit Party-Popper- und Glühbirnen-Icon sowie dem Schriftzug „HiQ“ auf schwarzem Hintergrund.

From Model Year to Always Up-to-Date

How Software Is Redefining the Car’s Lifecycle

But that logic is changing. With the right software architecture, the car can become a platform for continuous improvement – a product that evolves, refines, and optimizes over time, even after it leaves the factory.

This shift requires a new mindset: about products, organizations, and business models.

Decoupling That Changes the Game

At the heart of this change is the decoupling of software from hardware. In traditional vehicle platforms, functions are often tightly integrated with specific ECUs, sensors, and electronics. This makes every component replacement costly in development time and often requires rewriting the same function multiple times.

Being able to develop software independently of the hardware’s product cycle enables a shift from sequential to continuous development. This allows for shorter development cycles, greater scalability, and better use of resources.

But the point isn’t always about adding new functions later – it’s about not having to recreate existing ones. A large share of development time today goes into porting existing functionality to new hardware, often due to discontinued or replaced components. With the right architecture, application logic can be decoupled from the hardware, drastically reducing rework and enabling a more robust development chain. It also allows for more frequent model launches and shorter cycles – challenging the traditional model-year mindset.

Mehrere Autos und ein Lkw fahren auf einer geschwungenen Uferstraße entlang eines Flusses, während im Hintergrund unscharfe Stadtgebäude und Verkehr zu sehen sind.

From Distributed to Centralized Electronics

Traditional vehicle architectures consist of a large number of distributed ECUs, where each control unit manages a specific function. This has led to fragmented systems, complex troubleshooting, and limited reuse.

Modern Software-Defined Vehicle (SDV) architectures are shifting toward centralization – consolidating functions into fewer, more powerful computing units that can manage multiple systems in parallel. This can simplify wiring, reduce physical redundancy, and improve conditions for resource sharing between functions.

However, centralization doesn’t necessarily reduce complexity – it redistributes it, often shifting it from hardware to software and architectural levels. With fewer but more powerful nodes, there’s a growing need for clear architecture for isolation, real-time performance, redundancy, and safety-critical functions. For example, new patterns for partitioning, fallback handling, and quality assurance of software running on shared hardware are required.

The design of zonal architectures and the transition from classic AUTOSAR to, for instance, Adaptive AUTOSAR is crucial in this shift – especially to enable dynamic software deployment and future OTA updates of entire subsystems.

What Does It Take to Get There?

Decoupling is not just a technical detail – it’s a strategic choice that affects the entire development model. A well-designed base software, combined with clearly defined interfaces and a thoughtful component architecture, enables high code reuse across different models and generations.

This requires:

A clear software architecture that defines responsibility, dependencies, and isolation between functions
CI/CD pipelines that support both vehicle-specific verification requirements and continuous integration
Cross-functional teams capable of owning entire functions – from development to in-field operations

Middleware plays a key role here. It acts as a hub between hardware and application, abstracting the complexity of diverse hardware platforms. By managing resources, providing standardized APIs, and enabling virtualization, middleware forms a technical backbone for portability and reuse at a scale previously hard to achieve.

With increased connectivity, OTA updates, and external APIs comes increased responsibility. Cybersecurity and data protection can no longer be added at the end of development – they must be embedded in the architecture from the start.

Safety-critical functions must be isolated, updates signed and validated, and all data handling must comply with strict regulations – including GDPR and industry-specific standards such as UNECE R155 and ISO/SAE 21434. The ability to secure systems over time is key to building both trust and technical resilience.

Organizational Transformation in Practice

Technical transformation must be supported by organizational structures. OEMs that break down function-based silos and create teams with end-to-end responsibility – from idea to OTA update – report improved development speed and product quality.

At the same time, the relationship with suppliers is changing. Instead of linear chains, platform-based collaborations are emerging, where OEMs, Tier 1s, and tech companies co-develop around shared APIs, data models, and lifecycle strategies. This demands new levels of technical leadership and openness across the ecosystem.

The Car of the Future Is Always Current

The strategic business benefits are clear: faster time-to-market, shorter update cycles, and better scalability of innovation across product lines. At the same time, new revenue models emerge through service-based offerings, subscription features, and extended vehicle lifespans.

But with longer vehicle lifespans come new demands. Managing the full software lifecycle – from initial development to support, updates, and decommissioning – becomes a new challenge for the industry.

OEMs need strategies for maintaining compatibility with legacy hardware, handling component obsolescence, and ensuring that even vehicles on the road for ten to fifteen years can still receive critical updates and security patches.

In the end, the value of tomorrow’s vehicle won’t just lie in what’s built into it – but in its ability to continuously improve after delivery.
Data becomes an asset. APIs become business interfaces. The product becomes a platform.

Final Thoughts

The shift to software-defined vehicles is not just about technology – it’s about architecture, systems thinking, and collaboration. To succeed, organizations need the ability to navigate both deep technical complexity and broad business impact.

And it is precisely in this intersection – between system architecture, agile development, and strategic tech capabilities – that strong development partners make all the difference.

Want to know how HiQ can support you on this journey?
Let’s talk – we’d love to share more.

HiQ offers a wide range of Atlassian services and is a successful Atlassian Platinum Solution Partner.

Work Smarter with Rovo Agents

How to Build Your Own AI Assistants: a Step-by-Step Guide for Atlassian Rovo

Just recently, Atlassian announced to make its enterprise AI solution Rovo available to most Atlassian cloud users. Currently, Atlassian is kicking things off with access for Premium and Enterprise users of Jira, Confluence, and Jira Service Management, with Standard is coming soon.

Rovo includes several elements, most notably Rovo Search, which serves as personalized enterprise search, Rovo Chat, an AI teammate that answers questions and offer smart suggestions; and Rovo Agents.

What Are Rovo Agents?

Think of Rovo Agents as out-of-the-box and custom-built specialized AI teammates ready to get to work alongside you . They are built on Atlassian’s platform and go beyond simple automation. They can:

Clean up Jira backlogs
Draft release notes
Organize tasks based on themes
Summarize meetings
Analyze feedback
And much more – fully integrated into your tools and workflows.

You can start with one of 20 pre-built agents provided by Rovo or build your own, tailored to your team’s specific needs. Which brings me to the next point – three practical examples of Rovo Agents.

Examples of Rovo Agents

Transcript Insights Agent

Turn meeting notes into clear, actionable takeaways – automatically.

After a customer conversation, the agent reviews the Zoom transcript and generates a concise summary, reducing post-meeting time from an hour to just five minutes.

Release Notes Generator

Create polished release notes straight from your Jira backlog.

The agent scans completed issues, pulls relevant data, and assembles it into a structured release note format — ready to share.

Bug Triage Assistant

Improve the quality of issue tracking by ensuring reports meet reproducibility standards.

The agent evaluates bug tickets for critical information like reproduction steps or browser versions and flags any missing or unclear details.

How to Build Your Own Rovo Agent

Creating a custom Rovo Agent is quick and easy, even without technical skills. The process follows a guided, no-code setup within the product:

Define the Agent’s Purpose | Start by answering three key questions:

Role: Who will use the agent? (e.g. product manager)
Goal: What should it accomplish? (e.g. clean up a backlog)
Use Case: Where will it operate? (e.g. in Jira)

For example: A product manager cleaning up a backlog might create an agent that identifies flags missing details, duplicate tickets, and enforces labeling conventions.

Configure the Agent | Now add the details:

Name & Description: Choose a clear name like “Release Manager Agent” and provide a short description.
Instructions: Define how the agent should behave using a natural language prompt.
Knowledge: Link relevant content sources like Confluence spaces, Google Drive folders, or Jira projects.
Actions: Add up to five supported actions – for example, creating pages, moving issues, or analyzing content.
Conversation Starters: Help users engage with your agent by adding sample prompts like “What’s the status of open bugs?”

Creating your own Rovo Agents can help in saving time and ensuring quality control.

Rovo Agents offer two particularly valuable benefits for teams. First, they deliver significant time savings by automating repetitive tasks – for example, summarizing meeting transcripts or grouping Jira issues – saving valuable hours every week. Second, they support quality control by consistently applying predefined frameworks, helping teams stay aligned with internal standards and best practices.

Which brings us to the final point of this article: five of leveraging Rovo Agents successfully.

5 Tips for using Rovo Agents

To get the most out of your agents, follow these proven tips:

Start Small and Specific: Build agents with clearly defined tasks. Expand their abilities gradually as needed.
Test Before Automating: Chat with the agent manually first. Validate its behavior before plugging it into critical workflows.
Give Feedback: If the agent makes a mistake, ask why — then refine its instructions. Iteration is key.
Be Transparent: Each agent has a profile that shows its purpose and capabilities. This builds trust and encourages adoption.
Use Examples: Sample prompts help the agent perform better. Show it how to respond with relevant, real-world scenarios.

Conclusion: Small Agents, Big Impact

Rovo Agents make the advantages of an AI assistant tangible for every team – without needing a development background or complex setup. They save time, boost clarity, and help your teams focus on what really matters.

Curious how Rovo Agents fit into your Atlassian ecosystem? Learn more about our full range of Atlassian services.

Want more hands-on tips and live demos?

Join our free webinar on May 22 to learn how to get the most out of Atlassian Rovo in your daily work. With practical use cases and expert guidance.

Secure your webinar spot now!

HiQ covers every aspect of agile software development: from web applications to microservices, mobile development and IoT.

DevOps vs. Vibe Coding

The Future of Software Development in the AI Era?

The term, recently described by Andrej Karpathy, former CTO of OpenAI, refers to an extremely AI-driven coding style where developers largely let AI write the code and interact through natural language. Instead of manually searching for where to adjust the padding in a sidebar, you simply say, “Make the padding half as big,” and let the AI take care of it. Code changes are accepted without review, bugs are fixed through trial and error, and the code grows beyond the direct understanding of the human developer.

This raises a key question: Is vibe coding a disruptive method that can fundamentally change development work, or is it just a quick fix for prototyping? And how does it relate to established DevOps principles?

Is vibe coding a disruptive method that can fundamentally change development work, or is it just a quick fix for prototyping?

What is Vibe Coding?

Vibe coding is an AI-driven way of coding that relies on intuitive interaction rather than strict code control. Instead of manually writing code structures and debugging, the developer lets the AI handle the details. This creates a more direct, conversation-based development experience where the programmer describes what they want rather than exactly how to implement it. This method enables a fast and creative development process, making it particularly useful for prototyping and experimental development.

However, it also comes with a lack of code understanding and a risk of technical debt, as the code is generated at a pace that makes it difficult to keep up. Since the AI creates and modifies the code freely, it can also become harder to scale and maintain larger systems over time.

DevOps – Why It’s Still Necessary

For years, DevOps has been the standard in software development, built on continuous integration and delivery (CI/CD), version control, and automated testing. DevOps ensures that code is reproducible, scalable, and stable.

While vibecoding is primarily about speed and creative flow, DevOps focuses on structure, traceability, and quality assurance. Version control and code reviews ensure that everyone understands the codebase, while automated testing and security analyses minimize the risk of operational issues and security vulnerabilities. Infrastructure as Code (IaC) and CI/CD pipelines also make it possible to scale and update systems in a controlled manner.

At the same time, AI has the potential to improve DevOps. Instead of replacing DevOps entirely, AI can contribute by automating pipeline configurations, analyzing logs, improving code quality, and debugging systems faster than a human developer.

Can Vibe Coding and DevOps Coexist?

Rather than positioning vibe coding and DevOps as opposites, they can be seen as complementary. One possible use case is that vibe coding serves as a tool for rapid innovation, while DevOps takes over when it’s time to scale up and production-proof the system. Another scenario is AI-assisted DevOps, where AI helps streamline and automate more stages of the development pipeline.

Instead of just generating code, AI can assist in building CI/CD pipelines, monitoring operations, and suggesting optimizations. Another potential development is that vibe-coding becomes more common in frontend and UI development, where interactive changes can be quickly implemented, while backend and infrastructure continue to require more traditional DevOps processes.

What Does This Mean for Companies and Developers?

We are at a turning point in software development where companies must balance speed and structure. AI can lower the barriers to development and make it easier to build digital solutions, which presents an enormous opportunity.

However, it also requires a strategic approach to avoid technical debt and ensure long-term sustainability.

Should companies create internal guidelines for AI-driven development?
How can code quality be ensured when it is generated by AI?
And what does this mean for the role of the developer – will we see a shift from traditional programmers to a new profession where prompt engineering becomes just as important as coding skills?

These are just some of the questions organizations need to consider.

Future developers will need to navigate between AI-driven speed and DevOps precision—finding the right balance between intuition and control.

Conclusion

AI-driven development methods like vibe coding are exciting and could revolutionize rapid innovation, but they do not eliminate the need for structured processes in production and scalability. Future developers will need to navigate between AI-driven speed and DevOps precision – finding the right balance between intuition and control. AI is changing how we build software, but not why!

Want to build smarter, faster, and future-ready software?

See how we turn bold ideas into scalable solutions – with agile software development that works. Explore our development services!

At HiQ, we specialize in developing AI solutions that align with real business objectives.

AI: A Shield Against the New Reality of Trade Tariffs?

There is no doubt that we will have to relate to trade tariffs in the years ahead. This has become particularly clear as Donald Trump, in his renewed presidential campaign, has announced plans for sweeping new tariffs targeting countries like China, the EU, and other major industrial powers. The signals are clear: the wave of protectionism that has swept across the globe in recent years is not receding — if anything, it’s accelerating.

For the manufacturing industry, this means planning and strategy can no longer rely on a static trade landscape. Rather, companies must be prepared for rapid shifts, unexpected barriers, and cost changes that can severely impact both production and profitability. This is where AI comes in—not as a future vision, but as a practical tool for the present.

More and more industrial companies are realizing that AI can serve as a smart guide in a geopolitical terrain that is rapidly changing. By combining data from customs systems, logistics flows, and geopolitical analysis, AI can turn uncertainty into actionable insights. As Sofie Perslow, AI expert at HiQ, puts it: “In order for AI to predict the effects of trade tariffs in real time, it needs access to connected, quality-assured data from both internal systems and external sources. Logistics, customs information, and market shifts must be fed into the same data stream.”

Adjust in Real Time – or Lose Ground

When tariffs suddenly alter the cost structure, it is essential to react quickly. AI technology enables real-time analysis of the impact of tariffs at various stages of production, making it possible to swiftly adjust both pricing and component choices.

Companies can identify alternative materials with lower tariffs or adjust their sales prices based on regional conditions. Supplier contracts can be renegotiated automatically as trade policy evolves. It’s not just about numbers — it’s about staying competitive when the rules of the game are rewritten overnight.

From Bottleneck to Advantage in Logistics

Supply chains are often the first to feel the impact when new tariffs are introduced. AI can analyze thousands of logistics variables to suggest alternative routes, relocate inventory to tariff-friendly zones, and even automate customs documentation. During the US-China trade war, companies like Caterpillar and Whirlpool were forced into costly adjustments. With AI, many of these bottlenecks could have been anticipated — and mitigated.

“Agent-based AI makes it possible to automatically monitor supply chain changes, suggest renegotiations, or dynamically redirect flows. But systems must act in line with business logic — not just data-driven, but business-driven,” says Sofie Perslow.

“Agent-based AI makes it possible to automatically monitor supply chain changes, suggest renegotiations, or dynamically redirect flows. But systems must act in line with business logic — not just data-driven, but business-driven.”
Sofie Perslow, Head of AI, HiQ

See the Risks Before They Hit

Properly trained AI models can monitor trade patterns and geopolitical developments to predict upcoming risks. If a trade agreement is on the brink of collapse, or new tariffs are being signaled, AI can provide early warnings. This enables a more diversified supplier base and better-prepared scenarios for potential trade wars. A company like Harley-Davidson, which was hit hard by steel tariffs in 2018, might have made different decisions with access to such tools.

In a complex global economy, production optimization is critical. AI’s strength lies in its ability to weigh factors such as labor costs, energy prices, transportation options, and tariff rates. This enables companies to make strategic decisions about where production should be located — not just based on today’s costs, but tomorrow’s risks and opportunities. Apple, for example, has already begun relocating parts of its manufacturing from China to India and Vietnam. With AI, such decisions can be made faster, more accurately, and with lower risk.

Gelbes Absperrband mit schwarzer Aufschrift „IMPORT TARIFES,“ spannt sich im Vordergrund, während im Hintergrund unscharf gestapelte Frachtcontainer und Hafenkräne zu sehen sind.

The Right Product to the Right Market – Despite Barriers

Customer strategies must also be adapted to new tariff landscapes. AI-driven systems can analyze market data to recommend a shift in focus to regions with lower tariffs or adjust the product portfolio to minimize exposure to highly tariffed goods.

Even marketing efforts can be fine-tuned — for example, by using AI to optimize campaigns that account for price increases caused by tariffs. When the US threatened tariffs on Mexican goods in 2019, the auto industry was immediately affected. With the right AI support, companies could have reacted proactively.

Ready to unlock the full potential of AI in your business?

Discover how our Business-Centric AI solutions can help you navigate complexity, drive efficiency, and stay ahead of the competition. Explore our AI services!

The Bureaucracy No One Wants – But Everyone Must Manage

Trade tariffs bring not only economic consequences but also increased bureaucracy. Classifying products, managing regulations, and generating customs documents require time and accuracy. AI can automate much of this work, update internal systems with regulatory changes in real time, and ensure compliance — without pulling resources away from core operations. But the technology must go hand in hand with human judgment.

“Only when AI is connected to human judgment through transparency, feedback loops, and manual overrides can we build trust and real business value in critical decisions affecting customers, costs, and societal outcomes,” Sofie Perslow emphasizes.

“Only when AI is connected to human judgment through transparency, feedback loops, and manual overrides can we build trust and real business value in critical decisions affecting customers, costs, and societal outcomes.”
Sofie Perslow, Head of AI, HiQ

But – AI Is No Magic Wand Without Data Access

Challenges remain. Many AI tools are built for retail, not for the complex ecosystem of the manufacturing industry. To fully leverage AI in mitigating the effects of trade tariffs, better integration with customs systems, access to updated trade data, and collaboration between tech providers, companies, and government authorities are essential.

Still, the potential is clear: shifting from gut feeling to pattern recognition, from guesswork to simulation.

Conclusion

In a world where conditions change rapidly, AI is becoming the manufacturing industry’s most important tool for adapting — and excelling. It’s no longer about following developments, but about leading them.

Companies that begin using AI as a strategic partner today will have a competitive edge tomorrow — regardless of which way the geopolitical winds are blowing.

More AI Insights

News New Report Shows How We Use ChatGPT

News When AI Meets Cybersecurity: From Risk to Opportunity

News Atlassian Data Center Through 2029

From Model Year to Always Up-to-Date

Decoupling That Changes the Game

From Distributed to Centralized Electronics

What Does It Take to Get There?

Organizational Transformation in Practice

The Car of the Future Is Always Current

Final Thoughts

Contact

Work Smarter with Rovo Agents

What Are Rovo Agents?

Examples of Rovo Agents

How to Build Your Own Rovo Agent

5 Tips for using Rovo Agents

Conclusion: Small Agents, Big Impact

Contact

Christian Six

DevOps vs. Vibe Coding

What is Vibe Coding?

DevOps – Why It’s Still Necessary

Can Vibe Coding and DevOps Coexist?

What Does This Mean for Companies and Developers?

Conclusion

Contact

AI: A Shield Against the New Reality of Trade Tariffs?

Adjust in Real Time – or Lose Ground

From Bottleneck to Advantage in Logistics

See the Risks Before They Hit

The Right Product to the Right Market – Despite Barriers

The Bureaucracy No One Wants – But Everyone Must Manage

But – AI Is No Magic Wand Without Data Access

Conclusion

Contact

More AI Insights