The war for the last mile of user experience in health, banking, and energy: startups vs. traditional giants

Summary

The “last mile” of user experience has become the decisive battleground in hyper‑regulated sectors such as healthcare, banking and energy. In this stretch—from the moment the user says “yes, I want this” until they can use the service, pay, solve an issue or cancel without friction—perceived value, trust and likelihood to recommend are cemented [1]. Traditionally, this phase has been dominated by bureaucracy: forms, manual checks, waiting times and misaligned channels that erode the brand promise at the most critical moment [1][2].

At the same time, a new generation of startups has understood that the advantage lies not only in the core product, but in how the last mile is orchestrated. Supported by more flexible business models and modular technology architectures (APIs, microservices, KYC integrators, smart meters), they are redesigning onboarding, contracting, payment and support to make them almost invisible to the user [1][2]. This white paper comparatively analyzes how incumbents and startups configure that last mile in healthcare, banking and energy; explains the constraints of business models and technological legacy; and extracts actionable patterns and lessons for product, innovation and strategy leaders on both sides of the board.

Background

In regulation‑intensive services, user experience has historically been a side effect of regulatory compliance and internal operational efficiency. The priority has been “do not fail” (neither with the regulator nor with financial or clinical risk), above “do not frustrate” the user. Hence, in many countries, interacting with a hospital, a bank or an energy utility is still mentally associated with queues, paperwork and endless calls to a call center [1][2].

However, the competitive context is changing rapidly. In healthcare, the healthtech wave combines artificial intelligence, telemedicine and B2B2C platforms to reduce clinical errors and facilitate access. For example, AI‑based solutions launched in 2020 have achieved diagnostic accuracy levels close to 95% in specific pathologies, reducing diagnostic errors by around 30% and benefiting hundreds of thousands of patients per year in large metropolitan areas [2]. Although these data refer to the clinical core, the same technological logic—automation, real‑time analytics, data integration—is being applied to onboarding and initial use, enabling, for example, digital check‑ins, personalized reminders or smoother coordination between insurers and providers [1][2].

In banking, massive digitalization has changed the standard. A 2024 retail banking study places customer satisfaction at around 71%, driven by automation, digital channels and more personalized experiences [3]. This jump in satisfaction stems less from new financial products than from much lower friction when opening accounts, making payments or resolving issues. Fintechs and neobanks, built on “API‑first” architectures, have shown that it is possible to open an account in minutes, move money almost in real time and manage disputes from a mobile phone, reshaping users’ expectations across the industry [1][3].

In energy, network digitalization and the incorporation of smart meters and consumption‑management platforms are opening the door to “energy as a service” models, where the key is no longer just billing kilowatt‑hours, but transparently orchestrating usage, billing and support [4]. At the same time, the literature on last‑mile electrification in rural areas emphasizes that the challenge is not only technical, but also about designing operating models that can reach dispersed users, with clear information and simple channels for sign‑ups, payments and claims [4].

The e‑commerce experience provides a useful mirror: the logistical “last‑mile battle”—from warehouse to customer’s door—has turned what used to be a cost center into a decisive competitive advantage [5][6]. Leading brands integrate multiple delivery operators (multi‑carrier strategies) and manage distributed inventory to offer fast and reliable deliveries [5][6]. They have understood that post‑purchase experience and frictionless returns are as important as the initial purchase for customer loyalty [7]. Similarly, in healthcare, banking and energy, the last mile of UX is shifting from being an inevitable externality to a strategic lever where the market is won or lost [1].

Methods

This white paper synthesizes evidence and trends from three types of sources. First, it draws on recent sectoral studies that quantify customer satisfaction and the impact of digitalization, especially in retail banking and digital health [2][3]. These studies make it possible to anchor the analysis in verifiable metrics (for example, satisfaction levels, onboarding times, reduction in diagnostic errors or incidents).

Second, it analyzes technical and academic documents on electrification and access to energy services, which describe the physical and regulatory barriers of the last mile in energy [4]. These works provide a complementary perspective: whereas in banking and healthcare the last mile is mainly digital and procedural, in energy there is an unavoidable physical component (network infrastructure, meter rollout, installation maintenance) that conditions any redesign of the experience.

Third, it uses analyses of the “last mile” in e‑commerce to extrapolate operational design and user‑experience principles applicable to hyper‑regulated services [5][6][7]. The literature on multi‑carrier delivery, distributed inventory, post‑purchase experience and frictionless returns helps identify orchestration patterns that, although born in retail, are conceptually transferable to healthcare, banking and energy [5][6][7].

Based on this material, a three‑layer comparative framework—business model, technology and user interaction—is constructed and applied to three sectoral cases: healthcare, banking and energy. For each case, quantitative data are combined where available with plausible scenarios consistent with current regulation and technological capabilities. The aim is not to describe the situation of a specific country or player, but to identify structural patterns and design decisions that can be transferred across markets.

Finally, mini case studies are included to articulate concrete journeys (for example, opening an account or switching supplier) from the perspective of the user and the underlying architecture. The text avoids unjustified extrapolations: when comparable data are lacking, explicit hypotheses are formulated, grounded in business logic and in the regulatory constraints described in the literature [1][2][4].

Comparative framework: business model, technology and UX in the last mile

The last mile is not designed in a Figma screen; it is designed at the intersection of how money is made, how data flows and what room for maneuver regulation allows [1]. It is therefore useful to look at three chained layers that mutually reinforce or block each other.

Layer 1: business model and incentive structure

The business model determines who the real customer is, what unit of value is monetized and what time horizon matters. In healthcare, an insurer that charges annual premiums has incentives to control care costs and actuarial risk, whereas a private clinic focused on fee‑for‑service optimizes schedule occupancy and billing. A digital subscription healthtech, by contrast, lives or dies by monthly recurrence and the NPS of its end users [1][2].

These differences translate directly into the last mile. When revenue depends on large B2B contracts (for example, corporate health insurance), the pressure to simplify individual onboarding can dissipate: the end user is a “beneficiary,” not the paying customer. Individual friction is seen as an acceptable cost as long as the corporate relationship holds. A B2C or B2B2C startup whose churn spikes with every bad experience, by contrast, is forced to remove friction or disappear; it cannot take shelter in framework agreements [1].

The same applies in banking. A traditional bank with strong revenue from fees and cross‑selling can tolerate a heavier last mile if it believes this reduces default or fraud risk and protects margins. Its horizon is relationship‑based and long‑term; an expensive onboarding is amortized if the customer buys multiple products. A neobank with a low variable‑cost structure and a focus on a single star product (for example, a no‑fee account plus card) needs onboarding to be as simple as installing an app, because conversion at first contact defines the model’s viability [1][3].

In energy, the difference is even starker. An integrated utility that bills kilowatt‑hours under long‑term contracts may prioritize stability over commercial agility; a “captive” customer, due to infrastructure or low perceived competition, reduces the urgency to polish the last mile. A digital retailer that sells energy as a service or allows penalty‑free cancellations monetizes trust and flexibility, not just consumption volume. In short, the business model defines which frictions are considered an “acceptable cost” and which are existential.

Layer 2: technology architecture and legacy

On top of the business model sits the technological infrastructure. Incumbents carry decades of monolithic systems, point‑to‑point integrations and fragmented databases that make any change in onboarding, billing or support flows very costly. Even if they have built digital layers (apps, portals), beneath them there are still batch processes, manual checks and information silos [1][3]. Frequently, the modern “front” is connected via fragile integration layers to cores designed for another era, where even a minor change triggers multimillion‑dollar projects and lengthy internal approval cycles.

Startups, by contrast, start from modular architectures based on APIs and microservices, or on coupling multiple specialized SaaS tools (KYC, payments, communications, electronic health record, ticketing). This allows last‑mile journeys to be orchestrated almost as configurable flows: if an additional verification step or a new payment method is needed, a new provider is plugged in; the core does not need to be rewritten. In banking, open banking has shown how exposing account data via APIs allows third parties to build fluid experiences on top of traditional infrastructures [3]. In energy, the combination of smart meters and real‑time data platforms paves the way for consumption dashboards and proactive alerts [4].

The key is not only how “modern” the technology is, but how malleable. A legacy system can be fully compliant and robust, but if every small change requires months of project work and risk committees, the last mile is frozen in time. Startups, on the other hand, design for constant change: feature flags, A/B testing, automated deployments and usage metrics embedded in the journeys themselves [1].

Layer 3: user experience in the last mile

The third layer is what the customer sees: screens, forms, response times, tone of voice and channel consistency. This is where the most visible differences emerge. Startups and neobanks design journeys backwards: they start from the moment of maximum pain (for example, an erroneous charge or an urgent medical appointment) and ask which steps can be automated or removed. Incumbents tend to transfer the existing internal process to the digital channel: where there used to be a physical folder of documents, there is now a PDF form.

This is not just about visual design. The ability to offer “three‑click” onboarding, in‑app payments, real‑time status updates or one‑step cancellations depends on whether the underlying tech layer allows synchronous queries, federated authentication, digital signatures and automated business rules. UX is just the tip of a regulatory and technological iceberg [1][2].

The table below summarizes the framework succinctly:

Key findings

Healthcare: from endless paperwork to invisible onboarding

In the traditional healthcare model, the last mile is marked by fragmentation between insurers, provider networks and hospital systems. The user enters a circuit where each actor requests its own data and forms: insurer enrollment, hospital admission, informed consent, test authorizations and, in some cases, manual reimbursement processes. The lack of real interoperability between medical records means information is duplicated again and again, multiplying the risk of errors and delays [1][2].

This complexity has legitimate regulatory roots (data protection, clinical traceability, anti‑fraud checks), but also technological ones: many hospitals operate legacy information systems with poor internal integration, and insurers process authorizations in batch. The result is that the moment when the patient most needs fluency—an urgent appointment, a critical diagnostic test—is precisely when they encounter bureaucracy. Even though AI‑based solutions already show that it is possible to analyze large volumes of medical data with high precision, reducing diagnostic errors by around 30% where they have been deployed since 2020 [2], that clinical sophistication does not always translate into faster onboarding or simpler claims.

Healthtech startups have chosen to attack this gap. Digital‑first clinics and B2C or B2B2C telemedicine platforms allow users to register, verify their identity and schedule an appointment within minutes, via an app [1][2]. By building on unified electronic health records (where regulation and infrastructure allow) and APIs with insurers, they can automate basic authorizations, display coverage information in real time and collect co‑payments in‑app. The experience resembles booking any digital service: three taps, a clear summary of what is covered and push notifications with follow‑up.

These UX improvements require solving deep challenges: strict compliance with health data protection regulations, interoperability agreements with traditional providers and robust mechanisms for informed consent. But the pattern repeats: the startup does not try to rebuild the entire healthcare system, it encapsulates its complexity behind a consistent experience layer. Where the traditional hospital sees an “admissions process,” the healthtech designs an “invisible check‑in” that reuses existing data and minimizes patient input.

Banking: from the document folder to a neobank in 10 minutes

In retail banking, the contrast is especially stark. For decades, opening an account meant visiting a branch, presenting multiple physical documents, signing lengthy contracts and waiting for manual checks. Even when parts of the process were digitized, many institutions simply moved long forms to the web, keeping manual KYC/AML review. The consequence has been a fragmented experience: mobile apps with modern interfaces, but onboarding journeys that stall at in‑person steps or multi‑day waits [1][3].

Competitive and regulatory pressure has driven change. The adoption of artificial intelligence and automation in back‑office processes has helped make banking one of the leading sectors in digital customer satisfaction in 2024, with benchmark figures around 71% overall satisfaction [3]. Yet this figure is an average that hides a gap between incumbents and neobanks. The latter have built their value proposition precisely on the promise of “an account in 10 minutes,” relying on digital identity verification (video ID, biometrics), open banking to import data from other institutions and alternative scoring that reduces paperwork needs [1][3].

In issue resolution, last‑mile differences are just as pronounced. In a traditional bank, detecting and correcting an erroneous charge may require calls to the call center, branch visits and days‑long resolution times. Internal policies, designed to minimize risks and comply with complaints regulations, translate into friction for the user. A typical neobank lets customers report the charge from the app, temporarily block the card and open a ticket in seconds. The customer sees real‑time status updates, receives progress notifications and often gets a provisional refund while the case is investigated.

This fluency is not “fintech magic”; it is the result of API‑first architectures, rule engines that automate real‑time risk decisions and business models that can absorb some fraud cost in exchange for a superior experience [1]. Incumbents, with billions on their balance sheets and tight supervisory scrutiny, tend to be more conservative; but as regulation incorporates digital identity mechanisms and open‑banking standards, the technological gap stops being a justification and becomes a question of willingness to redesign processes.

Energy: from the incomprehensible bill to proactive consumption management

The energy sector combines capital‑intensive physical infrastructure with a historically opaque user experience. Traditional utilities have operated with long‑term contracts, complex billing—mixing regulated charges, network fees, taxes and rebates—and overloaded service channels. Switching providers, understanding a sudden spike in the bill or activating a new supply point are often perceived as tedious processes, with timelines extended by distributor‑side procedures or network limitations [4].

The barriers here are not purely regulatory; there is an unavoidable physical component: activating a supply point depends on network infrastructure and technical staff. But that does not preclude improving the digital last mile: the problem is that many incumbents have replicated, online, the internal logic of their commercial systems, keeping long forms, scattered information and a lack of clear consumption visualizations. The bill remains the main (sometimes only) touchpoint, and its design reflects accounting requirements more than user understanding.

New retailers and energy startups are transforming this experience [1][4]. Leveraging smart meters and near‑real‑time consumption data APIs, they offer dashboards that let customers view hourly consumption, receive proactive alerts when anomalous patterns appear and simulate the effect of switching tariffs or installing self‑generation [4]. “Energy as a service” models and dynamic tariffs make the relationship more like a digital subscription than an opaque supply. Contracting is 100% online, in minutes, even if the technical activation takes days: the user receives clear timeline estimates, status tracking and notifications as milestones are completed.

Structural limits remain. The startup can simplify the front end, but cannot fully control work‑order schedules, meter replacement or regulatory switching windows. Last‑mile design therefore focuses on expectation management, making progress visible and offering alternatives (for example, saving recommendations the customer can apply from day one). The goal is to turn a process that is partly out of its control into an experience perceived as transparent and supported.

The following table summarizes, comparatively, some last‑mile traits in the three sectors:

Comparative analysis

Sector differences: regulatory intensity vs. UX flexibility

Although healthcare, banking and energy all face high regulation, their mix of risk and market structure creates different trade‑offs. In healthcare, the main risks are clinical and privacy‑related. Regulation protects confidentiality and safety, but tends to be less prescriptive about how the onboarding interface must look. This leaves room to innovate in UX as long as consent and data‑protection frameworks are respected. That is why we see healthtechs achieving nearly invisible onboarding without touching the clinical core, while hospitals remain stuck in admission processes designed for paper [1][2].

In banking, KYC/AML regulation is extremely specific about what information must be obtained and how to verify it, limiting the possibility of “removing steps” unless they are replaced with robust digital‑identity technologies. Neobanks and fintechs have taken advantage of regulatory changes—such as the recognition of advanced electronic signatures and non‑face‑to‑face identification—to offer onboarding in minutes; but where these changes are not in place, even startups are forced to add friction [3]. In energy, regulation focuses more on tariffs, wholesale markets and networks; the commercial layer has some room to simplify language and processes, but is chained to physical timelines and conditions (for example, switching windows) that no design can erase [4].

Actor‑type differences: incumbents vs. startups

Beyond regulation, the major divergence arises from organizational starting points. Incumbents manage millions of customers, with complex structures and internal policies designed to minimize exceptions. This gives them scale but hinders last‑mile personalization: any change must work for everyone, across all branches and channels. Startups operate with smaller customer bases and can afford to iterate quickly, even temporarily handling some processes manually to learn before automating [1].

In banking, for instance, an incumbent can leverage its customer base and reputation to introduce incremental improvements (such as digital signatures or chatbots), but finds it hard to break the “standard journey” for contracting. A fintech can design, from scratch, a journey optimized for a specific segment (young people, freelancers, digital SMEs) and accept that other segments will be outside its initial focus. In healthcare, a large insurer has bargaining power and capacity to invest in interoperability, but often prioritizes internal efficiency projects over simplifying patient check‑in. A healthtech that lives off recurring use of a telemedicine app has no such luxury: the last mile is its product [1][2].

Key trade‑offs: risk, cost and control of the last mile

Radically redesigning the last mile entails trade‑offs. Automating decisions (for example, authorizing a medical test or approving a small loan in seconds) reduces friction but potentially increases fraud or overuse risk. Startups tend to accept more unit risk in exchange for growth, relying on advanced data models to contain it. Incumbents, with far larger balance sheets and public scrutiny, tend to over‑regulate the experience internally, adding verification steps to mitigate any doubt.

In energy, offering dynamic tariffs and data‑driven consumption recommendations may require substantial investment in data platforms and smart‑meter integration. For a traditional utility, that capex competes with other priorities (network maintenance, new generation assets) [4]. A digital retailer that does not carry the cost of physical infrastructure can devote a larger share of resources to the experience layer. In return, it depends on the network operator meeting deadlines and quality standards it does not control.

The underlying strategic question is who really owns the last mile. In e‑commerce, retailers have learned that fully outsourcing delivery experience to a single logistics operator is risky, and are building multi‑carrier setups and distributed inventory to gain resilience and speed [5][6][7]. In healthcare, banking and energy, similar moves will appear: incumbents outsourcing parts of the experience to specialized providers (KYC, telemedicine, consumption management), and startups that choose to “get their hands dirty” and integrate more deeply into core systems to truly control the last stretch [1].

Case studies

Case 1: Health insurance enrollment and first appointment

Imagine Laura, a 35‑year‑old professional who gets traditional health insurance through her employer. She receives a paper welcome pack with a policy number and a list of clinics. Weeks later she needs a specialist appointment. She calls the insurer, which tells her she must first choose an in‑network doctor, then call the clinic, and that some tests require prior authorization. At the clinic, she fills out an admission form with data she already gave the insurer. Authorization for a diagnostic test takes two days; the appointment is rescheduled. For a moderate issue, the end‑to‑end journey stretches over several weeks.

In parallel, Ana signs up for a digital health subscription with a healthtech. She downloads the app, verifies her identity through a remote onboarding process that cross‑checks her data with official sources, and in under 10 minutes gains access to a catalog of professionals. She books a video consultation for the same day. If the doctor recommends an in‑person test, the app shows partner centers, estimated prices and availability. Authorization is handled automatically via API with the associated insurer, and Ana receives a QR code in the app for clinic check‑in. Results and billing are integrated into her digital record. The difference is not just the channel, but how the last mile has been orchestrated, integrating traditional players under a single experience layer [1][2].

Case 2: Account opening