Technology and Service Innovation: AI Reality Check, Digital Services, Platforms, and the 7 Most Common Technology Anti-Patterns

36% of German enterprises use AI — doubled within a year¹. Yet only 7% of companies earn half their revenue from digital products and services². Volkswagen’s Cariad consumed EUR 14 billion, Lidl’s SAP project eLWIS EUR 500 million — both failed for the same reason: technology without a service vision.

This article shows what technology can actually deliver in service innovation, where the line between hype and reality runs — and which seven technology anti-patterns systematically prevent innovation.

Technology Is a Tool, Not a Goal

The decisive sentence for every technology decision in service innovation: Technology enables services, but it does not replace a service strategy.

Vendrell-Herrero et al. demonstrate in a systematic analysis of 26 studies: the sequence matters³. Organizations that first build their service capability and then digitalize outperform those that deploy technology and hope services will emerge from it.

Evgeny Morozov calls the opposite “technology solutionism”: complex social and business phenomena are recast as neatly defined problems with computable solutions⁴. The result: expensive technology investments without service impact.

AI in Service Innovation: Hype vs. Reality

What AI Can Actually Do Today

Capability	Maturity	Evidence
Pattern recognition in service data	Proven	Identifying unmet needs from behavioral data
Personalization at scale	Proven	Individualized service delivery without proportional cost increase
Predictive services	Proven (industrial)	Anticipating failures/needs before customer awareness
Process automation	Proven (commodity)	Removing friction from service delivery
Generative design	Promising	AI as co-designer in service ideation
Autonomous service delivery	Speculative	AI replacing human judgment in complex services

The Gartner Hype Cycle 2025

GenAI is in the “Trough of Disillusionment” — less than 30% of CEOs are satisfied with AI ROI⁵. 42% of companies abandon the majority of their AI initiatives before production.

Amara’s Law describes the pattern: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” The proven value of AI lies in infrastructure (data quality, AI engineering, governance) — not in speculative frontier applications.

DACH Data on AI Adoption

The Bitkom study 2025 (n=604) shows¹:

• 36% use AI (doubled from 20% the previous year)
• 81% see AI as the most important technology, 83% as an opportunity
• But: usage concentrates on customer interaction (88%) and marketing (57%)
• R&D only 21%, production only 20% — where real service innovation happens
• Barriers: Legal uncertainty (53%), technical knowledge (53%), talent shortage (51%)

The DIHK Digitalization Survey 2026 (n=4,897) adds⁶:

• 35% have AI in use, another 34% plan adoption within three years
• 78% of AI users deploy GenAI for text, image, and code generation
• 41% of active users rate the productivity effect as high
• Self-assessed digitalization grade: 2.8 (on a 1—6 school grading scale)

The central insight: AI adoption is growing rapidly, but deployment concentrates on the periphery (marketing, customer interaction) rather than the core of value creation. For service innovation, this means: AI is an enabler, but without a service strategy, it produces only more efficient versions of the status quo.

Digital Services: Digital-First vs. Digital-Enhanced

Not every service needs a digital revolution. The distinction between two fundamental types⁷:

Dimension	Digital-Enhanced Service	Digital-First Service
Definition	Existing service improved through technology	Service that could not exist without digital technology
Technology role	Enabler (reduce friction, increase reach)	Constitutive (the service IS the technology)
Example	Insurance consultation with digital closing	Algorithmic credit decision in real-time
Development logic	Service logic first, then technology	Technology and service logic simultaneously
Risk of confusion	Overengineering: too much technology for a simple service problem	Tech without user understanding: perfect code, no adoption

Warning signs of confusion:

• A digital-enhanced project has 80% technology budget and 20% service design budget → overengineering
• A digital-first project starts without user research → technology solutionism
• The project team cannot answer: “What changes for the customer?” → digital-washing

Platform Service Innovation: When Is a Platform Worth It?

B2B marketplaces in Germany are growing 17.5% annually⁸. This tempts companies to copy platform models. But: most B2B service contexts lack genuine network effects.

Five prerequisites for platform viability:

1. Each additional user makes the platform more valuable for all others (network effects)
2. Genuine multi-sided market dynamics exist (not just a supplier-customer relationship)
3. Marginal costs per additional user approach zero
4. The intermediation function creates value that bilateral relationships cannot
5. The ecosystem has sufficient actors for critical mass

If fewer than three prerequisites are met, you are not building a platform — you are building a customized portal. That can be valuable, but it follows different logic and a different investment framework.

Siemens Insights Hub (formerly MindSphere) shows how industrial platforms require years and massive ecosystem investment⁹. The platform connects machines, plants, and systems — but the road there was long and expensive.

Technical Debt as Innovation Blocker

92% of organizations are afflicted by technical debt¹⁰. 85% report that legacy systems hamper their ability to launch new solutions. The DIHK survey confirms: the biggest digitalization barriers in Germany are time (60%), complexity (54%), and money (42%).

The asymmetry: Every euro spent maintaining legacy systems is unavailable for innovation. Organizations spending 70—80% of their IT budget on maintenance have structurally no capacity for digital service innovation — regardless of their strategy.

7 Technology Anti-Patterns That Prevent Innovation

1. Technology Solutionism

Technology looking for a problem. VW Cariad invested EUR 14 billion in software without a clear service vision — “never set up as a true product company”¹¹. Lidl eLWIS tried to force technology onto unique business processes instead of starting from need. Diagnostic question: “Did this initiative start with a customer need or a technology capability?“

2. Digital-Washing

Superficially digitizing existing services without changing the value proposition. Between 2020 and 2022, several banks launched “digital branches” — behind the glass walls, manual reconciliation and paper-based approvals continued¹². Diagnostic question: “If we removed the word ‘digital’ from this initiative, would anything change?“

3. AI Hype Cycle

Overestimating short-term capabilities, underestimating long-term potential. 42% of companies abandon AI projects before production. The proven AI applications (data quality, ModelOps, governance) are unspectacular — the spectacular applications (autonomous service delivery) are unproven. Diagnostic question: “Are we investing in AI infrastructure or AI hype?“

4. Platform Envy

Copying platform models without network effects. Many Mittelstand companies pursuing “platform strategy” are actually building customized portals¹³. Assumptions from B2C platforms (Amazon, Uber) do not transfer to B2B service contexts. Diagnostic question: “Will each additional user make the platform more valuable for all others?“

5. Hoarding Data Without Generating Insights

60—73% of enterprise data remains unanalyzed¹⁴. Collecting data is not value — translating data into service decisions is value. Data lakes without an analytics strategy are cost centers, not innovation enablers. Diagnostic question: “Which service decision changes because of this data?“

6. Getting Build vs. Buy Wrong

Lidl (EUR 500M) and VW (EUR 14B) show both sides: Lidl over-customized a standard product, VW over-built a commodity¹⁵. Otto Group shows the successful path: building proprietary platform technology based on existing culture and capabilities. Decision logic: Build for strategic differentiation, buy for commodity functionality, partner for ecosystem dependency.

7. Accepting Technical Debt as Permanent State

92% of organizations suffer from technical debt, 79% must divert resources from core objectives¹⁰. Legacy systems are not an inevitable fate — they are decisions that can be revised. Diagnostic question: “What percentage of our IT budget flows into innovation vs. maintenance?”

Technology Assessment for Service Innovation

Before investing in technology, check five dimensions:

Dimension	Weight	Key Question
Need validation	30%	Is the customer need demonstrated (not assumed)?
Technology maturity	20%	Is the technology proven or experimental?
Organizational readiness	25%	Does the organization have the capabilities to operate the technology?
Business case	15%	Does the expected benefit justify the investment?
Risk	10%	What happens if it fails? Is the damage contained?

The weighted order is decisive: Need validation and organizational readiness together account for 55%. The technology itself (maturity) only 20%. This reflects empirical reality: most failed technology projects fail not because of the technology but because of missing need or missing organizational capability.

More on organizational readiness: Innovation Culture and Organization. Getting started: Getting Started with Service Innovation.

FAQ

Do I need AI for service innovation? No. AI is an enabler, not a prerequisite. The most successful service innovations begin with deep customer understanding (user research), not technology. AI becomes relevant when scaling, personalization, or predictive capabilities are needed.

What distinguishes digital service innovation from digitalization? Digitalization optimizes existing processes with technology. Digital service innovation creates new or fundamentally changed value propositions. Digitalization is an enabler but not a substitute: making processes faster is not the same as creating new services.

Should I build a platform? Only if genuine network effects exist. If each additional user makes the platform more valuable for all others — yes. If not, you are building a portal. Portals can be valuable but follow different logic and a different investment framework.

How do I avoid technology solutionism? Three tests: (1) Did the initiative start with a customer need? (2) Would we pursue this if the technology did not exist? (3) Who is the user, and did they ask for this? If any answer is “no”: pause and develop the service strategy first.

What does not digitizing cost? 82% of companies believe Germany’s economic crisis stems from slow digitalization². 73% report lost market share. The question is not “Can we afford to digitalize?” but “Can we afford not to?”

How do I start with data-driven services? Do not start with data — start with the service question the data should answer. Then check: Does the data exist? Is it accessible? Is it of sufficient quality? 57% of organizations estimate their data is not AI-ready⁵.

What about shadow AI? 78% of employees bring their own AI tools¹⁶. This is not just a security problem — it is a signal: the organization is not providing what employees need. The answer is not bans but the provision of secure, capable AI tools.

---

Footnotes

1. Bitkom AI Study 2025. 604 companies, representative. 36% use AI (doubled from 20%). 81% see AI as most important technology. ↩ ↩²

2. Bitkom Digitalization Survey 2025. 603 companies. 82% see economic crisis caused by slow digitalization. Only 7% earn half their revenue digitally. ↩ ↩²

3. Vendrell-Herrero et al., systematic analysis (DASOBI Framework, 2024). Servitize first, then digitalize. 26 quantitative studies. ↩

4. Evgeny Morozov, To Save Everything, Click Here, Public Affairs, 2013. Technology solutionism as interpretive framework. ↩

5. Gartner Hype Cycle for AI 2025. GenAI in “Trough of Disillusionment.” Less than 30% of CEOs satisfied with AI ROI. 57% of data not AI-ready. ↩ ↩²

6. DIHK Digitalization Survey 2026. 4,897 companies across all sectors, end of 2025. Grade 2.8 for digitalization status. ↩

7. Synthesis from: Youngjin Yoo, “Digital First: The Ontological Reversal,” MIS Quarterly, 2020. Digital-first vs. enhanced framework. ↩

8. Xpert.digital, platform economy analysis. B2B marketplaces: 17.5% annual growth. Marketplace share: 55% of e-commerce. ↩

9. Siemens Insights Hub (formerly MindSphere). Industrial IoT-as-a-Service platform. Gartner Visionary in Magic Quadrant 2022. ↩

10. Unqork / Morning Consult, 2024. 500 decision-makers. 92% afflicted by technical debt. 85%: legacy hampers new solutions. ↩ ↩²

11. German Autopreneur, “Cariad: VW’s Software Failure — Lessons.” EUR 14 billion invested, 6,000 employees, then 2,000 job cuts. ↩

12. ScienceDirect 2025, “Digital Washing.” Banking example: digital branches with manual backstage processes. ↩

13. Springer 2023: “Rare empirical studies show that transferring B2C platform assumptions to B2B leads to incorrect management decisions.” ↩

14. MIT Sloan / Forrester: 60—73% of enterprise data remains unanalyzed. Data lakes without analytics strategy are cost centers. ↩

15. Lidl eLWIS: EUR 500M SAP project failed (over-customization). Otto Group: successful proprietary platform (1/3 of workforce = tech experts). ↩

16. Bitkom Shadow AI Report. 78% of employees use their own AI tools. Gartner: By 2030, 40%+ security incidents from shadow AI. ↩