Redesigning invisible infrastructures: what startups and traditional companies should copy from each other

Abstract

The so‑called invisible infrastructures—payments, data management, regulatory compliance, internal logistics, support, and back‑office processes—quietly determine the competitiveness of any business. While many comparisons between traditional companies and startups focus on the product or branding, evidence shows that the deepest disruption happens in these non‑visible layers. Digital startups, less constrained by legacy systems, are adopting modular architectures, cloud services, and specialized APIs to redesign how sectors such as financial services, retail, and healthcare operate [1][2]. At the same time, incumbents are beginning to modernize their infrastructures with emerging technologies—from AI and low‑code to cloud computing—achieving significant gains in efficiency and resilience [1][2][3][4].

This white paper analyzes, across three dimensions—business model, technology architecture, and impact on user experience—how the approaches of startups and traditional companies differ. Drawing on cases in fintech, e‑commerce, and healthtech, it identifies cross‑cutting patterns and derives practical recommendations for both sides. The central message is clear: sustainable competitive advantage does not lie only in a pretty interface, but in how the invisible infrastructures that support it are designed, governed, and evolved.

Background

Invisible infrastructures are the “nervous system” of any organization. They include everything from the transactional core (such as the banking core or the hospital information system) to payment systems, customer identity management, data platforms, regulatory compliance workflows, and support and back‑office mechanisms. End users rarely see these components, but they immediately feel their effects: waiting times, errors, onboarding friction, lack of personalization, or low price transparency.

In practice, these infrastructures define what is and is not possible for a business. A bank may want to offer instant accounts, but if its core requires overnight batch processes and manual checks, the promise will be barely credible. A retailer may invest millions in a new app, but if inventory data is not updated in real time, customers will keep finding “available” products that they then cannot pick up. Similarly, a hospital may announce telemedicine, but if medical records remain on paper or in local systems, physicians will not have all the information they need during a video call.

Over the last two decades, technological acceleration—cloud, APIs, data lakes, artificial intelligence—has created a broad array of new “building blocks” for these invisible layers [1][2]. Startups and traditional companies have leveraged these blocks with very different strategies. While many incumbents are constrained by legacy systems, rigid hierarchies, and a culture resistant to change [6], startups operate with agile structures, risk tolerance, and a mindset of continuous experimentation [7]. This cultural difference translates into divergent innovation rhythms precisely at the operational base of the business.

Recent research indicates that companies integrating emerging technologies into their infrastructure—for example, AI in operational management or scalable cloud platforms—achieve substantial improvements in productivity, cost reduction, and crisis responsiveness [1][2][3][4]. In solar energy, applying AI to the operation of photovoltaic plants has enabled optimization of generation and maintenance, improving profitability and sustainability [1]. In construction, the combined use of BIM, drones, AI, 3D printers, and robotics has accelerated projects, reduced errors, and created new competitive advantages [2]. In industry and services, low‑code platforms have enabled faster and safer application development: Schneider Electric developed 60 applications in less than two years, while Petrobras increased development productivity by 60% by modernizing more than 300 applications [3]. During the COVID‑19 crisis, Booking.com used cloud infrastructure to analyze data in real time, achieving 20% growth in bookings as the market began to recover [4].

However, innovation in invisible infrastructures is not free of risk. A study published in April 2025 shows that 73% of fintechs fail in their first three years due to avoidable regulatory compliance issues [5]. Cases such as Aereo or Lumosity illustrate that ignoring the regulatory and evidence components can lead to shutdowns or sanctions regardless of the degree of technological innovation [5]. Thus, while startups often excel in technological agility, incumbents retain strengths in risk management, data governance, and operational continuity.

Organizational culture acts as either a catalyst or a brake on these invisible transformations. Cultures that promote flexibility, collaboration, and openness to change make it easier to adopt new technologies and review internal processes [6]. In traditional companies, rigid hierarchies and risk aversion can block modernization, as illustrated by periods in the history of manufacturers like Ford compared to more agile competitors [6]. Conversely, startup culture—based on autonomy, rapid iteration, and customer orientation—supports deep redesigns of internal layers [7]. Evidence indicates that diverse teams combining different disciplines and experiences produce more original work with greater long‑term impact, which also translates into more innovative infrastructure designs [8].

This context creates an opportunity: to develop a comparative analysis that goes beyond the visible product and focuses on the silent layers that enable—or block—new business models and superior user experiences.

Methods

This white paper is based on a qualitative synthesis of secondary sources and on the use of company archetypes (rather than specific brands) to illustrate patterns. The aim is not to describe exhaustive individual cases, but to derive design principles for invisible infrastructures that are applicable across multiple contexts.

First, recent literature on the integration of emerging technologies into enterprise infrastructures has been used, covering sectors such as solar energy, construction, manufacturing, and digital services, where quantitative impacts on productivity, cost, and resilience are documented [1][2][3][4]. These studies provide evidence on how technological modernization of internal layers translates into tangible competitive advantages, and allow principles to be extrapolated to information‑intensive service sectors such as banking, retail, and healthcare.

Second, analyses of startup failure due to regulatory issues and compliance best practices have been incorporated, showing the importance of integrating compliance into the design of processes and systems from the outset [5]. This evidence is combined with work on the impact of organizational culture on innovation and competitiveness, which explains why startups and traditional companies differ in how they approach internal transformation [6][7][8]. The literature on startup culture and digital transformation in corporations helps clarify which elements are transferable across contexts and which depend on sectoral or regulatory factors [7].

Finally, based on these inputs, three sectoral cases (fintech, retail, and healthcare) are constructed and organized around a common comparative framework: business model, technology architecture, and impact on user experience. In each case, hypothetical scenarios are proposed based on patterns observed in the market and linked to available data. The cross‑cutting patterns section distills the shared elements identified, and the recommendations sections translate these findings into concrete actions for executives, product managers, and innovation strategists.

Comparative framework: model, technology, and experience

The analysis is organized around three interdependent dimensions:

1. Business model and monetization: how revenue is generated (fees, subscriptions, usage, advertising), what level of recurrence exists, and how value capture is balanced against user alignment. This dimension also includes revenue diversification, resilience to economic cycles, and pricing clarity for the customer.

2. Technology architecture and way of building: degree of modularity, use of public/private clouds, adoption of APIs and third‑party services, level of automation, and observability. Here we analyze decisions such as “build vs. buy,” the use of low‑code platforms, continuous integration, and business continuity mechanisms.

3. Impact on user experience: speed (response and approval times), friction (steps, forms), personalization (segmentation, recommendations), and trust (transparency, perceived security, compliance). This dimension is measured both functionally (what the user can do) and emotionally (what they perceive and how they feel).

These dimensions are not independent. The technology architecture determines which business models are viable (for example, whether it is possible to measure recurrence and calculate LTV in real time) and what kind of user experience can be offered (onboarding in minutes vs. days). At the same time, the business model incentivizes certain technology investments: companies that depend on recurring revenue tend to prioritize data platforms and advanced loyalty mechanisms; organizations whose monetization depends on strict compliance (for example, health insurers) will invest more in traceability and data governance.

User experience, in turn, is the visible manifestation of decisions made in the other two layers. A bank with a monolithic core and manual processes will find it very difficult to offer instant accounts, even if it wants to strategically. A hospital with paper medical records cannot enable online self‑service appointment booking without completely redesigning its back‑office. Understanding these interrelationships is critical for any executive or product manager evaluating where to invest in modernizing their business and how to manage internal expectations of impact.

Key Findings

Case 1 – Fintech and financial services

In traditional banking, the core banking system is typically a monolithic, on‑premise installation from decades ago, to which layers and patches have been gradually added. Each new product entails lengthy development cycles, complex testing, and high risks of breaking existing functionality. Payment processing still follows batch cycles, with daily closings and maintenance windows. This directly impacts the experience: account openings taking days, card limits that are hard to adjust, limited transparency in real‑time transactions, and slow claims processes.

In contrast, the fintech archetype adopts from the outset an API‑first, modular architecture where the current account, card, consumer credit, and KYC (know your customer) are all modules connected via internal and external APIs. This approach enables the use of banking as a service (BaaS) to delegate regulated functions to specialized third parties while retaining control of the user experience. Integration with external providers for payments, risk scoring, or identity verification reduces time‑to‑market for new products from months to weeks and facilitates A/B testing at the level of financial products.

At the business model level, the traditional bank continues to rely on a mix of intermediation margins and opaque fees: account maintenance, transfers, cards, overdrafts, bundled services. This model, inherited from a low‑competition, pre‑digital context, incentivizes intensive cross‑selling of products not always aligned with the client’s best interest. Fintechs, by contrast, experiment with transparent pricing, basic no‑fee accounts, revenue from card interchange, premium subscriptions (for example, bundled insurance or reduced FX fees), and in some cases freemium models where only advanced features are paid.

These business differences translate into radically different experiences. Onboarding at a traditional bank may require in‑person visits, physical signatures, and approval times measured in days. A fintech with digital KYC, automatic document verification, and decision flows based on rules and models can complete account opening in minutes from a mobile phone. Fintechs can offer dynamic limits, automatic spending categorization, and real‑time notifications because their data architectures are built for streaming events and near real‑time analytics. Again, invisible infrastructure enables the visible value proposition.

Regulation, however, introduces important nuances. Fintech is one of the sectors most exposed to regulatory risk. The Hare Strategy Group study shows that 73% of fintechs fail in their first three years due to avoidable regulatory compliance issues [5]. This reveals that many financial startups prioritize speed and product innovation over integrating compliance into their internal processes. In some cases, such as Aereo, clashes with the prevailing legal framework led to business closure despite technological sophistication [5]. Lumosity, for its part, was sanctioned for making scientifically unsupported claims, demonstrating that marketing and product design are also part of the regulatory infrastructure [5].

Traditional banks, with decades of experience in risk management, audits, and regulator relations, have developed capabilities that help them avoid many of these mistakes. Their compliance processes, although often manual and bureaucratic, are deeply integrated into daily operations. The challenge for incumbents is to transfer that discipline into more modern architectures without suffocating innovation; for fintechs, it is to incorporate compliance‑by‑design: business rules, traceability, and automated controls embedded directly into systems, along with documentation practices, internal audits, and ongoing training similar to those of banks [5].

Case 2 – Retail and e‑commerce

A traditional omnichannel retailer typically operates with legacy CRMs by business line, different point‑of‑sale (POS) systems by country, and non‑integrated databases across physical store and e‑commerce. This creates data silos: in‑store purchases are not linked to web browsing, online returns do not feed into in‑store replenishment algorithms, and the call center rarely sees the customer’s full history. The visible consequence is a fragmented experience: promotions that do not apply across channels, poorly personalized loyalty programs, and generic, ill‑timed marketing communications.

A digital‑native startup designs a data‑centric infrastructure from the outset: a Customer Data Platform (CDP) that unifies browsing, purchase, support, and campaign events; cloud‑based data lakes that store large volumes of semi‑structured data; and real‑time tracking tools. The goal is not just to collect data, but to activate it: automated recommendations, dynamic segmentation, and re‑engagement flows orchestrated by rules or machine‑learning models. This ability to orchestrate data is a key piece of its invisible infrastructure.

In terms of business model, the traditional retailer tends to measure performance based on units sold and margin by store or category, with annual plans and seasonal campaigns. Customer value is viewed retrospectively and in aggregate. Direct‑to‑consumer (DTC) and digital‑native startups, by contrast, build their strategy around Lifetime Value (LTV), customer cohorts, subscriptions, and product bundles. This focus on long‑term value and recurrence is only possible because the data infrastructure tracks customer behavior over time and estimates expected value.

The operational difference is clearly reflected in UX. A traditional retailer may offer a decent website and app, but the experience breaks down when a customer tries to return in store a product bought online or redeem an email coupon in a different channel. These failures are not just process issues; they stem from systems that do not share a unique customer identifier and back‑office flows designed by channel, not by journey. The digital‑native startup, with a unified view of the customer, can offer consistent cross‑channel experiences, dynamic loyalty programs, and extensive automation in communications and support.

Data underline the value of modernizing these infrastructures. Traditional companies that have integrated emerging technologies into their processes—such as AI to optimize inventory, BIM and robotics in construction projects, or automation in logistics—have accelerated projects, reduced costs, and improved quality to the point that these tools have become critical competitiveness factors [2]. During the pandemic, Booking.com’s ability to analyze data in real time thanks to its cloud infrastructure allowed it to quickly adjust its offering, contributing to 20% growth in bookings as the market started to recover [4]. Although not physical retail, this shows the direct impact of modern data infrastructure on business responsiveness.

Case 3 – Healthcare / Healthtech and regulated sectors

In traditional healthcare systems, such as public hospitals or large insurers, it is common to find paper medical records, non‑interoperable local systems, and manual authorization processes. Patient information is scattered across specialists, facilities, and insurers, leading to duplicated tests, repeated forms, and communication errors. Moreover, regulatory compliance is often handled via manual processes and periodic audits, adding administrative burden to already overstretched professionals.

A digital health startup takes the opposite approach. Its infrastructure is built on cloud platforms, interoperable electronic records, APIs, and open data standards. Compliance is not treated as an afterthought, but as a set of automated rules and workflows (compliance‑by‑design) embedded in the platform itself: consent management, data anonymization, access logs for medical records, and full traceability of clinical or administrative decisions. This allows faster adaptation to regulatory changes without redoing manual processes each time and provides systematic documentation of compliance for inspections.

From the patient’s perspective, the difference is obvious. A traditional system offers appointments with weeks‑long waits, limited control over scheduling, little visibility into one’s own records, and fragmented communication among primary care, specialists, and insurers. A healthtech startup can offer on‑demand teleconsultations, self‑service appointment management, online access to records and test results, and remote monitoring via wearables or apps. These capabilities enable new business models: corporate wellness subscriptions, data‑driven preventive programs, and remote second‑opinion services.

The risks, however, are significant. Regulated sectors penalize legal misalignment especially harshly. Cases such as Aereo—shut down after losing a legal battle despite its technological innovation—or Lumosity—sanctioned for unsubstantiated scientific claims—show that innovation without regulatory governance can rapidly destroy value [5]. The study that identifies 73% of fintech failures as due to regulatory problems [5] is conceptually applicable to healthtech: ignoring regulation, scientific evidence, and data ethics can turn a technological advantage into an existential threat.

At the same time, healthcare incumbents can learn from startups’ agility. Integrating emerging technologies into infrastructure—such as AI for triage, robotics in operating rooms, or low‑code tools to automate administrative processes—has already produced significant improvements in efficiency and quality in other sectors [1][2][3]. The key question is not whether to modernize, but how to do so gradually, with strong governance and a focus on critical patient journeys, minimizing clinical and regulatory risks and making the most of professionals’ accumulated experience.

Table 1 – Synthetic comparison by sector

Comparative Analysis

Dimension 1: Business model and monetization

Traditional companies tend toward business models that reflect the regulatory and technological history of their sectors. In banking, credit intermediation and service fees dominate; in retail, unit and store margin; in health, fee‑for‑service or procedure‑based payments. These models were rational in environments with limited information and poorly integrated systems. However, they often create incentives that undermine the best user experience: tariff complexity, service overuse, or misalignment between what the customer pays and perceived value.

Startups, by contrast, explore models based on recurrence and data: subscriptions, pay‑per‑use, bundles, and added‑value analytical services (for example, personalized recommendations, preventive programs). This approach is only viable because their invisible infrastructures can measure user behavior, calculate LTV, and adjust prices or packages dynamically. The combination of real‑time data and contractual flexibility makes it possible to experiment with freemium models, free trials, or personalized discounts that would be hard to manage with legacy systems.

The risk is that in the pursuit of growth, many startups underestimate the importance of revenue diversification and sustainable margins—an area where incumbents, used to managing economic and regulatory cycles, have more experience. Conversely, traditional models, although more robust, can become opaque and disconnected from current expectations of transparency and alignment.

Dimension 2: Technology architecture and way of building

In technology architecture, the central difference is modularity. Incumbents start from legacy, often monolithic systems that have been patched for years. Changing these systems involves multi‑year, high‑risk projects. As a result, in many sectors visible innovation (apps, websites, campaigns) moves much faster than transformation of the operational core.

Startups, by contrast, build from scratch using public clouds, APIs, and specialized third‑party services, following a logic of combining “build” (differentiating core) with “buy” (commodity components). This approach allows them to concentrate resources on the parts of the stack where they add the most value—such as risk algorithms, recommendation engines, or UX flows—while delegating standard functions (payments, messaging, basic analytics) to established providers.

The adoption of emerging technologies in invisible infrastructures has already produced compelling results. In energy, integrating AI to manage solar plants has reduced operating costs and improved sustainability [1]. In construction, BIM, drones, AI, 3D printing, and robotics have sped up projects and improved quality [2]. In industry, low‑code platforms enabled Schneider Electric to create 60 apps in less than two years and Petrobras to increase development productivity by 60% when modernizing over 300 applications [3]. During the pandemic, Booking.com’s cloud infrastructure was key to adjusting its offering almost in real time and capturing market recovery [4]. These data show that technological modernization is not a “nice to have,” but a determinant of competitiveness.

Dimension 3: Impact on user experience

Ultimately, customers do not see the architecture, but they do see the speed, friction, personalization, and trust it enables. Incumbents tend to offer experiences marked by long times (for approval, delivery, or appointments), bureaucratic friction (redundant forms, in‑person processes), and limited personalization. This is not necessarily due to lack of will, but to the constraints of inherited systems and processes designed around internal limitations rather than user needs.

Startups treat the desired user experience as the starting point for process design (service design). From there, they configure their invisible infrastructures to support that vision: onboarding in minutes, frictionless returns, on‑demand teleconsults, automated recommendations, and so on. Trust is built differently: not only through brand and regulation, but via radical transparency (clear pricing, direct communication on regulatory changes) and visible practices of data protection and compliance [5]. However, when the innovation narrative exceeds reality (as in Lumosity’s scientifically unsupported promises [5]), trust can erode quickly.

Table 2 – Trade‑offs between incumbents and startups

Case Studies

Case 1 – Modernizing payments at a regional bank

A 40‑year‑old regional bank was suffering from card issuance times of 5–7 days and high abandonment rates in online application processes. Its payment systems were tightly coupled to the monolithic core, and any change required long development and testing cycles. Management decided to address invisible infrastructure modernization with a layer separation strategy: they kept the existing core but introduced an integration layer exposing controlled internal APIs.

On top of this layer, they integrated a modern payment provider and an external digital KYC service. Rather than attacking the core directly, they completely redesigned only the card onboarding journey. The result: issuance times fell to under 24 hours for most customers, and abandonment dropped significantly. Inspired by fintech models, they introduced real‑time notifications and app‑based adjustable limits while preserving traditional risk and regulatory controls. This approach showed that it is possible to combine incumbent‑style compliance discipline and operational robustness with more agile experience layers.

Case 2 – DTC startup learning resilience from a retailer

A DTC personal‑care startup grew rapidly, supported by advanced data infrastructure, a centralized CDP, and customizable subscription models. However, it depended almost entirely on a single logistics provider and multiple third‑party services for payments, analytics, and customer support. A sudden change in logistics pricing and a prolonged outage of one of