Energy prices, skills shortages, NIS-2 and personal liability weigh on Germany’s Mittelstand. Proliance CEO Alexander Ingelheim explains why “compliance at the push of a button” fails audits, and how a connected platform with human expertise holds up.

Economic pressure, bureaucracy and cyber risk are overstretching Germany’s Mittelstand, and the personal liability of managing directors is a particular worry. Some software vendors promise regulatory conformity at the push of a button, yet deliver templates that prove worthless in an audit. In an interview at it-sa 2026 in Nuremberg, Proliance co-founder and CEO Alexander Ingelheim explained why automated fig leaves can raise liability risk, and how a consolidated platform combined with expert advice can create genuine information security.

Pressure from all sides

Running a mid-sized company in Germany means balancing operational problems against regulatory demands. Rising costs, strained supply chains and a persistent skills shortage already absorb considerable management capacity. At the same time, European rules such as the NIS-2 Directive, the Cyber Resilience Act and the EU AI Act raise the bar for digital resilience. According to figures from Germany’s Federal Office for Information Security (BSI), which Ingelheim cited in a June article, roughly 30,000 organisations in Germany fall under NIS-2, yet only about 11,000 had registered by March 2026.

Lawmakers increasingly hold executives personally accountable when security is neglected. If protective measures are missing, or an incident is handled poorly because no emergency plan exists, the consequences extend beyond downtime to sanctions. For a company that depends on a few central IT systems, an outage without clear crisis instructions can mean the end of the business.

Ingelheim describes the mood in management teams plainly. “On the one hand, the German Mittelstand is under extreme economic pressure,” he said. “Energy prices, skills shortages, supply chains: these topics take up so much focus that very little time is left for unloved subjects like information security, data protection, AI and compliance. Many managing directors are overwhelmed.” His company therefore applies a firm rule: “We no longer enter a customer relationship if the managing director isn’t in the meeting and says, ‘I have understood that this has to be a priority.’ Otherwise it becomes a three-year chewing gum and everybody ends up unhappy.”

The 99-euro fig leaf

Faced with complex rules, many companies look for the easy route. Offers are multiplying that promise full conformity through generative AI tools at a low price. Under scrutiny by external auditors, such shortcuts often prove ineffective: standard policies generated without reference to actual business processes protect neither against incidents nor against legal consequences.

“AI washing is visible in every field now,” Ingelheim said. “If I hear that a platform makes you NIS-2-ready for 99 euros, I can guarantee you will not pass an audit. It’s a small fig leaf, policies that don’t fit the company at all.”

He pointed to a US case from earlier this year without naming the vendor. Public reports from March describe allegations that a US compliance-automation startup produced near-identical audit documents for hundreds of clients, which came to light through a publicly accessible spreadsheet; the company has rejected the accusations. According to Ingelheim, the episode caused a major loss of trust, and some affected customers turned to Proliance. His advice: check twice what you buy and where AI is used.

Proliance says it avoids the problem with two principles. AI supports repetitive work known from its consulting projects, but a human stays in the loop: nothing is inserted that an employee cannot review. And the AI draws on curated knowledge, such as an asset database maintained by Proliance staff and about nine years of project experience, rather than on generated text, according to Ingelheim.

One asset register instead of ten frameworks

A common mistake is to treat each rulebook in isolation. When data protection, ISO standards and sector rules are handled in separate systems, administrative effort and silos follow, and overstretched IT departments inherit tasks for which they have neither capacity nor primary responsibility. Relief comes from consolidating the relevant information on one platform. Once infrastructure and data flows are recorded in a shared inventory, different requirements can be built on top.

“We have to speak of connected compliance and security to keep this manageable,” Ingelheim said. “Otherwise you have ten people working on different frameworks that should talk to each other but don’t, and you keep building silos. If you have a clear view of the infrastructure and a correct, up-to-date asset register, you can build on it so much that you only have to do it once.”

He stresses that this is an organisational project, not an IT task: from day one, colleagues from the specialist departments are invited into the software to enter their information. For an IT lead with two administrators, he said, adding NIS-2 on top of daily operations would pull them away from their core duties.

Ingelheim’s June guest article in Healthcare Digital, aimed at hospitals, suggests a pragmatic start: define the scope, run a gap analysis, assess the risks. Writing down the five most important business processes, naming their dependencies and assigning responsibilities already provides a basis for priorities. He argues that progress beats perfection.

Shadow AI: enablement instead of bans

Beyond classic threats, unmanaged use of AI tools by staff is a growing risk. Out of ignorance or convenience, employees paste sensitive company data into freely accessible online services. Bans usually fall short and only push usage out of sight. The better route is a sanctioned working environment and clear rules.

Proliance tries to identify all AI-native applications and rate them by use case. “You very quickly get a view: these are the five AI systems I must keep on my radar because they have explosive potential,” Ingelheim said. He is candid about the limits: detecting shadow AI technically is currently very difficult, and Proliance cannot do it itself. Its approach is enablement. Employees move onto one managed platform they are happy to use as a gateway, and clear policies define what may be used, for what and by whom.

Evidence and European data sovereignty

To stand up to auditors and supervisors, decision-makers need to see their compliance status at any time. According to Ingelheim, the Proliance 360 platform has four areas: an asset register, a risk hub, a control hub and a policy center, where policies are generated and training is tracked. Users see each relevant framework with a score from 1 to 100 and can drill down by department. Reports on risks, controls, trained staff, the asset register and emergency plans can be exported as PDF, so management can document how it meets its oversight duties.

The location of processing matters as much as the rules. Last year Proliance moved to the French provider OVHcloud, a switch Ingelheim describes as a one-to-two-week effort. According to the company, costs fell, security stayed at the previous level, and the data is held entirely on European servers with no dedicated line to the US. It now runs open-source AI models there and says they work well, because curated data matters more than the newest model.

The topic has gained weight since the US Supreme Court ruled on 29 June in Trump v. Slaughter that the independence of the Federal Trade Commission is unconstitutional. The FTC’s oversight role underpins the EU-US Data Privacy Framework. The framework remains formally in force, but the European Data Protection Board has asked the European Commission to assess the ruling’s effect. Ingelheim advises companies to inventory their data flows to the US, review standard contractual clauses and transfer impact assessments, and examine a European alternative with every new implementation. He cites Bitkom: 85 percent of companies consider Germany too dependent on US cloud providers

What it takes, and what it costs

How long implementation takes depends on how strongly management pushes, Ingelheim says. Privacy projects have been completed within a week, information security takes three months for small firms with a young tech stack and about a year as standard, and an EU AI Act project can be done within a week, according to the company. A privacy solution costs from about 2,000 euros per year for the smallest firms, rising to roughly 20,000 to 30,000 euros per year for a larger ISO certification at a company with 100 to 150 employees. Proliance, based in Munich, has more than 50 experts and focuses on the Mittelstand.

A structured approach, Ingelheim argues, joins legal certainty with efficient data handling and can lead the Mittelstand out of the bureaucracy trap. The precondition is the one he sets at the start of every engagement: the managing director has to be at the table.

By Dietmar Müller

Dr. Dietmar Müller is an experienced IT journalist and editor-in-chief of Beyond Buzzwords. He has decades of expertise in the German and international IT markets, as well as extensive editorial experience at renowned publishing and media companies such as Springer, IDG, and CBS Interactive. Whether writing cross-media technical articles, conducting in-depth analyses, or conducting insightful interviews—whether at his desk or in direct dialogue with industry leaders—Dr. Dietmar Müller is right at home in the world of enterprise IT. His key areas of focus include enterprise IT and OT architectures, the cloud, ERP, MES, data centers, cybersecurity and compliance, industrial automation, and AI governance, among others.

Leave a Reply

Your email address will not be published. Required fields are marked *

WordPress Cookie Notice by Real Cookie Banner