Healthy data helps organisations see clearly, make better decisions and innovate with confidence. This article explores why data quality matters, what healthy data looks like, and how it supports more meaningful digital innovation.
Software as a Strategic Asset, Not a Technical Utility
In my 30 years of building for the web, I have noticed a recurring pattern in how organisations approach software. There is a tendency to view technology as a utility—something you “buy in” or “bolt on” to solve a specific problem. I believe this is a fundamental mistake. Software is not just code; it is the digital manifestation of your Intellectual Property (IP).
When a startup or an enterprise team focuses purely on technical engineering, they often reach what I call a “functional plateau.” The technology works, the servers are stable, and the code is clean, but the business outcomes are flatlining. This happens because the “why” was never designed into the “how”. McKinsey & Company’s research confirms that the best-performing companies treat design as a top-management issue, assessed with the same rigor as revenues and costs. If you don’t spend the time at the beginning to understand the problem deeply, you aren’t building an asset; you’re building a liability.
The “White-Collar Factory” and the failure of intuition
The office spaces of the 1950s and 60s were designed like factories: lines of evenly spaced desks where people were expected to perform repetitive, mind-numbing tasks. It seemed intuitive at the time because it mirrored industrial success. It took the designer Robert Propst at Herman Miller to observe that people aren’t machines; they need varied environments for focus, collaboration, and movement.
I think we are making the exact same mistake today with specialist software. I see digital “factories” being built where features are included based on what feels “intuitive” to the organisation. By prioritising business “needs” (often just data-gathering requirements) over the reality of the people actually using the system, the result is a long list of features that lead to both design and technical debt. Technical debt is bad in so many ways and is the polar opposite of what the goal should be – to build Digital Equity.
Design debt manifests in the modern era as an explosion of complexity, hundreds of screens produced by AI generation tools or overly extensive, unmanaged design systems in Figma. It creates a labyrinth that the user has to navigate and the team has to manage.
This is compounded by Technical debt, where an ever-growing pile of features demands constant maintenance and an ultimately expensive rebuild, just to reach parity. This is made significantly worse by AI’s tendency to produce “bloated code”, functional but inefficient scripts that require human intervention to untangle. This “Build Trap” leads to systems that are impossible to maintain because they were built for a spreadsheet, not a person. My experience with specialist audiences – whether they are Plant Health practitioners in a busy market or sales people managing leaderboards – is that their intuition and the organisation’s intuition are rarely the same.
The Discipline of the Technical Illustrator
My perspective on this was shaped at 17, as a trainee technical illustrator drawing for GKN Westland Helicopters, now Leanardo. The “human-first” direction didn’t come from my pen; it came from the Technical Authors who were former RAF engineers.
They would talk to me about their time in the field, explaining that a specific part had to be removed before another could be accessed. These weren’t aesthetic notes; they were strategic insights based on physical reality. They possessed real-world knowledge that the original designers of the helicopters would have lacked. It taught me that design is an inquisitive process. You cannot “guess” your way to a solution; you have to draw on the experience of those who know how the machine actually works.
Swapping Autonomy for The Matrix
There is a scene in the film The Matrix that serves as a hauntingly accurate forecast for the current state of digital transformation. It describes a world where humans are surrounded by machines that they no longer know how to build, fix, or even understand. They simply exist within a system that has become too complex.
I believe we are on the road to this reality in 2026.
As “vibe coding” and unmoderated AI become the default for rapid development, we are seeing a massive shift towards what cognitive science would describe as the “Autonomy Trap.” When you allow a system to generate code or design interfaces based on broad prompts, you might get a result that looks impressive on a Monday morning. But if you didn’t go through the inquisitive design process to get there, you don’t actually “own” that solution. You don’t have the map.
In design, less is nearly always more. AI, by its very nature, tends to go straight to “More”, more features, more noise, more complexity. Without a “Designer in the Loop”, not just a human, but a specialist trained in articulating challenges, you lose the context required to evolve that software. If you can’t explain why a specific workflow exists, you can’t improve it when the market shifts.
Designing for Reality with AUX
To move beyond the theoretical, we have to look at how a human-first lens changes business outcomes. At Modular, we champion the term AUX (Authentic User Experience). It is the process of finding the “Transaction Tension” and resolving it through observation rather than intuition.
When designing for specialist environments, such as an engineer on a wobbling boat in the North Sea, where every “intuitive” design choice made in a comfortable studio goes out the window.
We have had to solve for high-contrast interfaces that remain readable in direct glare, large-hit areas for users wearing gloves, and robust offline-first syncing for locations where “the cloud” is a distant memory. We’ve learned that in specific regions where users encounter repeating variables, like the same crop or the same common disease, a simple ‘last entered’ data option can save hours of frustration. It allows for rapid input of repeating data, enabling the specialist to skip the “knowns” and spend their time where it matters most: talking to the person in front of them.
If we had simply “engineered” what was asked for in these scenarios, the platforms would have been useless within ten minutes of entering the field. Instead, we design tools that protect the organisation’s IP by ensuring the data is actually recorded accurately in the harshest conditions.
The ROI of “Designing Less”
There is a pervasive myth that “more features equals more value.”
I see teams falling for this every day, believing they must build a Swiss Army knife to justify the investment. The commercial reality is the opposite. McKinsey found that the market disproportionately rewards companies that truly stand out with exceptional design.
The commercial reality is the opposite.
Every feature you build that isn’t aligned with a real-world human problem is Technical Debt you are choosing to take on. Top-quartile MDI (McKinsey Design Index) performers increased their revenues and total returns to shareholders (TRS) substantially faster than their industry counterparts, 32 percentage points higher revenue growth and 56 percentage points higher TRS growth for the period as a whole.
By “Designing Less,” you are actually protecting your future cash flow. You reduce the immediate build cost, yes, but you also reduce the “Maintenance Tax” that eventually suffocates tech-first projects. When we hone the challenge at the start, we often find that the most impactful solution requires fewer screens, less code and more intuitive experiences.
[By designing with intent, the end experience is precise and exacting. Efficient]
The Future of Specialist AI
As we look ahead, the trend is moving toward Domain-Specific Models. Generic AI models often fall short of the nuance required for specialist work. The value is no longer in the code itself, but in the “Designerly” orchestration of these tools. The future belongs to those who don’t just “deploy” AI, but who design the workflows that allow humans and AI to coexist productively.
My Takeaways for Product Owners
Design is not a “nice to have”, or a ‘make it look pretty’ after thought; it is the primary engine of business success in the 2026 digital landscape. To move beyond the “Build Trap,” organisations consider these five principles:
Prioritise Empathy over Intuition:
Human-first design means gathering genuine audience insights before writing code. Insight drives innovation, while assumptions drive technical debt.
Design for Engagement, Not Function:
A tool is only successful if your audience actually uses it. By designing only what is truly needed, you reduce immediate build costs and protect the future of the business by minimising design and technical debt.
The Process IS the IP: The design process is how you garner the knowledge to innovate and maintain a solution. To skip this step is to lose the context that makes your software a durable strategic asset. The value is not in the code; it’s in the understanding of the problem that the code solves.
Hone the Challenge: Every prompt and feature must be sharpened by a designer’s eye. Unmoderated AI and “vibecoding” create noise that complicates implementation and leads to “designslop”. Success is found in the “Designerly” articulation of the problem.
Stop the Cycle of Fixes: Designing it right the first time eliminates the permanent “maintenance tax” of fixing things that should never have been built in the first place. Efficiency is not building fast; efficiency is building once.
Engineering builds the platform, but Design builds the value. Without keeping the designer in the loop, you aren’t just building a digital product, you’re building a digital debt that will eventually have to be paid.
If you’re planning a specialist software project, our team at Modular can help you plan and deliver value rich bespoke software solutions, reach out to hello@thisismodular.co.uk and we’ll be in touch to arrange an initial call.
Author Bio
Adam Millington is the Founder of Modular, a Bristol-based agency designing software for specialist audiences. His 30-year career began as a technical illustrator before evolving into Creative Director roles and entrepreneurship. Adam clarifies complex technical puzzles for niche sectors, from global animal health data for researchers to custom farm systems for seaweed farmers. Bridging the gap between strategy and execution, his “Human-First” approach ensures technology fits the real-world workflows of the specialists who rely on it.
Sources & Further Reading:
Benedict Evans: AI Eats the World – Interface and Strategy (2026)
McKinsey & Company: The Business Value of Design
RAND Corporation: Why 80% of AI Projects Fail (2025/2026 Analysis)
Gartner: I&O Leaders Report on AI Project Stalls (April 2026)
Nielsen Norman Group: State of UX 2026 – Deep Design as a Differentiator
Melissa Perri: Escaping the Build Trap – Outcome-Based Strategy
