INTERVIEWS
Security operations are often described through systems, tooling and predefined workflows. In practice, however, the defining characteristic of a real incident is not technical sophistication but the environment in which decisions are made.
“They have one very common denominator… adrenaline… stress,” said Nicol Daňková, Head of SOC, Incident Response & CTI at Henkel. What she is describing is not chaos, but intensity – a compressed window in which information is incomplete, consequences are immediate, and judgement must operate under pressure.
In theory, organisations prepare for this moment through simulations and tabletop exercises. In reality, the change from rehearsal to live incident alters the equation entirely.
“The things are usually not going as they have been decided.”
The difference lies not in the playbook, but in the context. During a drill, failure is contained. During a live incident, every action has operational, financial, and reputational implications. That awareness changes how people behave.
At the same time, attackers are not bound by predefined scenarios.
“The attacker can go absolutely anywhere… can do something very different than for what you are having playbooks.”
Preparedness remains essential, but the presence of a playbook does not guarantee coherence. What incidents reveal is whether coordination structures hold when the script dissolves.
Where Organisations Begin to Strain
When escalation begins, it is rarely the technology that falters first. Systems may continue to function; monitoring may still generate signals. The stress fractures appear in coordination.
Under pressure, decision latency increases. Ownership becomes blurred. Individuals hesitate to escalate.
“No one wants to be the person who is really triggering the red buttons.”
This hesitation is understandable. Escalation carries responsibility, and in high-stakes environments responsibility carries risk. Without clearly defined authority and trust in process, teams default to caution.
The effect is cumulative. Information fragments. Parallel investigations emerge. Instead of a coordinated response, activity becomes distributed and reactive.
Daňková describes how easily organisations can lose structural cohesion under these conditions.
“Plenty of the organisations are turning into… the bunch of a headless running chicken.”
The imagery is vivid, but the underlying issue is architectural. Incident response is not merely a technical function; it is an organisational one. When coordination mechanisms are weak, even strong technical teams struggle to operate effectively.
Assumptions Under Pressure
One of the most destabilising dynamics during an incident is the gradual substitution of verification with assumption.
“One of the biggest issues… is practically assuming instead of verifying.”
Under time pressure, teams rely on partial signals. Hypotheses harden into perceived facts. Without a shared source of truth, each group works from its own interpretation of events.
“If plenty of the people… are working only based on assumptions… it is creating mess in the quality of the information.”
The result is duplication of effort and erosion of time. The same questions are answered repeatedly. Validation loops multiply. What appears as activity may in fact be stagnation.
“If this kind of a verification is happening five times… you can really imagine what kind of a slowing down effect it is having.”
For attackers, delay creates opportunity. For defenders, time is the scarcest resource.
Visibility as a Structural Advantage
It is tempting to believe that expertise compensates for structural weakness. Daňková rejects that assumption.
“You can have the absolutely best playbooks in the world… the best experts… but in the moment when there is no proper coordination… even the best people really cannot succeed.”
The differentiator is not individual capability but collective visibility.
“Really knowing what is going on… having some kind of a clarity… and having some kind of a central point of knowledge.”
Centralised visibility allows teams to operate in parallel without divergence. It ensures that verification is shared, that assumptions are challenged collectively and that decisions align across functions.
Without it, expertise becomes isolated. With it, escalation becomes manageable.
The Leadership Variable
At the centre of this structure sits the incident response lead; a role often underestimated in technical discussions.
“It is all dependent on the guy who is behind the steering wheel.”
The role extends beyond task coordination. It requires psychological steadiness under stress, the ability to translate technical realities into executive language, and the discipline to maintain focus on foundational facts before pursuing more complex threads.
“It is a role which is… talking to all of the C-level people… defending the technicians from the upper management.”
The incident lead becomes a stabilising interface between operational teams and strategic leadership. Their effectiveness depends not only on technical fluency, but on resilience, judgement and the ability to preserve clarity amid noise.
Beyond Containment
Incident response does not conclude with technical containment. Regulatory obligations, disclosure requirements and legal exposure extend the lifecycle of an event.
“If you haven’t fulfilled some kind of regulation demands… those fines can be astronomical.”
This introduces a second dimension to response: governance. The organisation must account for what happened, demonstrate procedural compliance, and maintain stakeholder trust.
Incidents therefore function as systemic audits. They expose weaknesses in coordination, communication, and escalation pathways as much as they reveal vulnerabilities in code.
Technology remains critical. But under pressure, structure determines outcome.
“Really knowing what is going on.”
AI Is Amplifying Difference – And Systems Are Built to Reward It
Artificial intelligence is often described as connective infrastructure – a system that increases access, accelerates understanding and reduces friction.
Peter Mousaferiadis approaches the question differently.
In conversation ahead of Tech Show Frankfurt, he situates AI within a longer arc of technological compression — where time, distance and information barriers collapse. The critical issue, in his view, is not connectivity itself, but what happens when systems scale difference without shared context.
Compression Without Convergence
Mousaferiadis traces the modern inflection point to 1989 — the fall of the Berlin Wall and the birth of the World Wide Web.
“Tim Berners Lee gifted us the worldwide web… that led to what we call this hyperdiverse age where time and space were compressed.”
The compression of distance created unprecedented access. But access does not automatically produce alignment.
“You would think with the worldwide web the world would be becoming a better place… but… peace has been steadily in decline since 2008, 2009.”
The divergence, he suggests, is not accidental. It is systemic.
The Architecture of Amplification
AI systems do not operate in isolation. They are embedded within economic and engagement models that reward attention.
“These algorithms have been developed in such a way to polarise us because they generate income out of that.”
Engagement-driven optimisation tends to reinforce existing viewpoints rather than interrogate them. At scale, that reinforcement becomes structural.
This is less a question of individual bias and more a question of design incentives. Systems optimise for measurable interaction. Measurable interaction often correlates with emotional intensity. Emotional intensity can deepen division.
The outcome is not neutral amplification. It is selective amplification.
The Baseline Assumption Problem
At the centre of Mousaferiadis’ argument is a design assumption.
“Because if those algorithms are still being built with the assumption that we’re all starting from the same place socially and culturally, we are damned.”
Digital systems frequently assume a shared baseline — shared literacy, shared context, shared access to opportunity. In practice, societies are uneven.
When AI systems are deployed at scale without acknowledging those asymmetries, they can magnify disparities rather than reduce them.
The issue extends beyond representational bias. It concerns structural alignment between system design and social reality.
Governance, Not Just Capability
The discussion, therefore, moves away from whether AI is technically impressive.
“If technology is not there to serve humanity, then what purpose does it serve?”
For Mousaferiadis, this is not a philosophical aside. It is a governance question. How systems are optimised, how success is defined and how accountability is structured determine their long-term impact.
AI’s capabilities will continue to advance. The more consequential question is whether the frameworks surrounding those capabilities evolve alongside them.
Direction Is a Design Choice
AI does not create fragmentation in isolation. It interacts with existing social and economic conditions.
In a hyperconnected environment, differences become more visible. Whether that visibility leads to cohesion or polarisation depends on how systems are designed, regulated and incentivised.
The architecture being built today will shape not just information flow, but how societies interpret and respond to complexity at scale.
“It Should Be the Last Resort” – Rethinking Net Zero in Data Centres
Sustainability in data centres is no longer defined by efficiency metrics alone. As AI workloads intensify and power demand rises, the challenge is expanding beyond design into energy systems, cost structures and regulatory alignment.
Across Europe in particular, operators are facing growing pressure to demonstrate how facilities interact with the grids and communities around them. What was once treated as an optimisation exercise now intersects with grid stability, energy sourcing, and regional planning.
Ahead of Tech Show Frankfurt on 6-7 May at Messe Frankfurt, Ed Ansett, Global Director of Data Centre Technology and Innovation at Ramboll, examines how sustainability is being reshaped by rising density, regulatory scrutiny and the practical limits of infrastructure. The discussion moves beyond aspiration and into systemic constraint.
From Efficiency to Contribution
For many operators, sustainability is broadening in scope.
“We’re seeing organisations having to tackle things like demand response, grid support, and heat exports,” Ansett explained. “Data centre developers are increasingly under pressure by local authorities to think more about how they’re contributing back to society.” Ansett added that Ramboll often supports operators and local authorities in developing practical solutions for demand response, heat reuse and grid integration.
Facilities are increasingly evaluated not only on how efficiently they operate, but on how they participate within wider energy ecosystems. In mature European markets, that expectation is already influencing planning decisions.
Where “Net Zero” Is Being Reconsidered
At the same time, established sustainability narratives are being reassessed.
“A lot of companies… claim to be net zero… this has really been achieved primarily through something called carbon offsetting.”
Offsetting remains part of the equation, but Ansett argues that reduction must precede compensation.
“It really should be the last resort rather than the first resort when it comes to net zero.”
The emphasis moves toward operational and embodied carbon reduction — through energy sourcing, design decisions, and material choices — before offsets are used to close residual gaps. Net zero becomes a process grounded in measurable change rather than declaration.
AI and the Power Constraint
Rising compute density is amplifying the tension.
“we’re currently at around 100 and something kilowatts per rack… we’re talking about a megawatt per rack.”
As density increases, cooling, and power distribution requirements scale with it. The central constraint becomes energy availability.
“The increase in density means changes in the way that provision them… Then of course on top of that is where’s this power coming from? Because it was about vast amounts of power.”
The sustainability question therefore extends beyond facility-level optimisation. It reaches into grid capacity, generation mix and regional infrastructure readiness.
Cost as the Structural Barrier
While debates often centre on resilience or redundancy, Ansett highlights economics as the more immediate tension.
“I think the bigger challenge is actually the trade-off between sustainability and cost.”
Improving sustainability frequently introduces additional capital and operational expenditure. Operators face competitive pressure while navigating rising expectations around carbon, biodiversity, and water usage.
“You want to do the right thing… but you also have to compete with other service providers.”
The distance between intention and implementation is often shaped by financial constraints rather than technical feasibility.
Retrofitting Reality
Retrofitting existing sites is frequently presented as a straightforward solution. In practice, the picture is more complex.
“It’s unfair to simply say to people… you have to retrofit… it may not be practical.”
Older facilities were not designed for current rack densities or energy loads. Lease terms, physical layouts and legacy power infrastructure can limit what is feasible. In many cases, new builds are driven by structural limitations rather than strategic preference.
Toward Greater Transparency
Industry communication is also tightening.
“One thing that’s got to be slammed out altogether is the greenwashing that’s been going on for a long, long time.”
Regulatory oversight and local authority scrutiny are increasing expectations around reporting. Energy sourcing, water usage, and embodied carbon are subject to more consistent measurement.
Shared benchmarks are gradually emerging as part of that maturation. The conversation moves from narrative to evidence.
The Direction of Travel
“it’s got a very, very long way to go… because there’s a heck of a lot to do.”
Ansett explains that the industry is progressing, but the workload remains substantial.
Sustainability in data centres now operates at system scale. It depends on how facilities integrate with energy networks, economic realities and regulatory frameworks under increasing demand. The central question is no longer whether operators prioritise sustainability, but whether infrastructure ecosystems can support it at scale.
Where AI Strategy Collides with Organisational Reality
Tiankai Feng, data and AI leader and author of Humanizing AI Strategy, tends to resist the idea that AI failure begins with the model.
Across organisations, he sees something quieter happening first. Teams launch pilots, early demos generate momentum, but once systems meet the realities of shared drives, partial documentation and inherited processes, the energy begins to thin out.
The friction rarely looks dramatic. It shows up in small ways: information isn’t maintained consistently, ownership drifts, updates are applied in one place but not another. What seemed coherent in a workshop setting feels less stable inside everyday workflows.
For Feng, that is where the real constraint sits.
AI depends on context — on structured data, traceable decisions and a degree of operational clarity that many organisations have never had to formalise. Traditional software could be absorbed into existing processes. AI exposes the gaps inside them.
When Momentum Meets Structure
He describes a pattern he calls “pilot theatre”: an impressive proof of concept, a period of enthusiasm, and then gradual neglect. The system continues to function, but its relevance fades. No one is quite sure who is responsible for maintaining it, or how it connects to measurable outcomes.
The technology remains intact. The environment around it does not.
This is also why resistance can be misunderstood. When teams hesitate, it is often read as cultural reluctance. Feng suggests it is just as likely to be structural misfit. If a system demands new forms of collaboration, new documentation standards or new accountability models, those shifts carry risk for the people involved.
Embedding AI, therefore, requires more than deployment planning. It asks organisations to examine how information moves, how incentives are aligned and how responsibility is distributed. In many cases, AI does not create new weaknesses — it makes existing ones visible.
What determines whether experimentation endures is less the sophistication of the tool and more the willingness to reshape the structure around it.
Organisational Readiness:
Lessons from Accessibility, Talent and Infrastructure Leaders
Across accessibility, AI deployment, healthcare analytics, cloud engineering and early-career talent developament, a shared theme becomes visible. Technical capability is advancing rapidly. What determines impact is whether organisations are structured to use that capability responsibly and at scale.
Much of this comes down to organisational readiness for AI. Tools alone are not enough. Governance, ownership, culture and operational discipline shape whether technology delivers durable value.
Each conversation in this series touches a different domain, yet the underlying challenge is similar. Embedding accessibility into product design requires operational discipline. Building sustainable AI capability requires executive alignment and shared ownership. Developing the next generation of cloud engineers requires structural investment. Scaling analytics in healthcare depends on trust, transparency and user-centred design
Embedding Accessibility Into Systems, Not Statements
David MacArthur, Digital Accessibility Consultant at Accessibility Now, approaches accessibility as an operational question rather than a branding one. In his experience, the most significant shift happens when teams consider inclusion at the start of a project rather than treating it as a final check before release.
That shift affects more than compliance. It changes how teams think. Once engineers, designers and product owners begin looking consciously for accessibility issues, they start to notice patterns they would previously have ignored. Small design decisions around colour contrast, navigation, media controls or procurement choices begin to carry more weight.
Importantly, this is not framed as a technical complexity problem. Much of the work involves incremental improvements and deliberate perspective changes. The deeper issue is whether organisations build inclusion into their operating model or treat it as a reactive exercise. Where accessibility is embedded into architecture and governance, it becomes sustainable. Where it is layered on afterwards, it rarely is.
The principle extends beyond accessibility. Early ownership matters.
Building the Workforce That AI Requires
Emily Hall-Strutt, Director at Next Tech Girls, brings a longer time horizon into the discussion. Interest in technology among young women exists. The friction appears later, when that interest must convert into opportunity.
Early-career roles have tightened. Unpaid placements remain common. Apprenticeships and degree apprenticeships are still unevenly available. Structural barriers, including financial constraints and limited visibility of non-traditional pathways, continue to shape who enters the industry.
In this context, inspiration alone is insufficient. Visibility of role models is important, but it does not replace paid, accessible entry points. Retention also demands attention. Cultural signals inside organisations influence whether young women see technology as a long-term career or a short-term experiment.
The conversation here moves beyond representation targets. It raises questions about how organisations design policies, career ladders and early-career programmes so that talent development is deliberate rather than assumed.
Future capability depends on present infrastructure.
From Proof of Concept to Production Value
Chief Data & AI Officer, Kevin Cassar, reflects on enterprise AI and the gap between technical build and realised value. The views expressed in his interview are his own and do not reflect the opinions, positions or perspectives of his existing or previous employers.
A recurring issue in AI programmes is the proliferation of prototypes without clear pathways to production. Organisations experiment widely, yet struggle to embed outcomes into core operations. Ownership often sits narrowly within data teams, while executive alignment and cross-functional sponsorship remain underdeveloped.
He emphasises foundations. Clear articulation of why a model exists, who is accountable for its outcome, and how it will be measured matters as much as model accuracy. Governance frameworks, when introduced early, help teams understand constraints and avoid investing in initiatives that cannot scale safely.
The technical components of AI are advancing quickly. Scaling responsibly depends on organisational clarity.
Engineering for Constraint and Continuity
Ryan Kirk, Head of Cloud & DevOps at Formula One, describes a markedly different environment at Formula 1, yet the underlying theme remains similar. Infrastructure must function under immovable deadlines, physical transport constraints and unpredictable external conditions.
Trackside systems are powered down, shipped and reassembled across continents. Weather events, latency requirements and broadcast commitments introduce complexity that cannot be deferred. Automation, guardrails and disciplined planning are therefore central, not optional.
He also speaks about experimentation in practical terms. Testing ideas, accepting that some will fail, and refining processes iteratively form part of the engineering culture. The focus is not on perfection at first attempt, but on steady improvement supported by clear safeguards.
High-performance systems are not only technically sophisticated. They are supported by teams that understand both risk and adaptation.
Designing Analytics That Influence Decisions
Chris Beeley, Head of Data Science, The Strategy Unit (NHS), shared that his experience in NHS analytics introduces another dimension. Accurate outputs alone do not guarantee adoption. Analytical products influence decisions when they are designed around the people who must use them.
Time constraints, professional backgrounds and emotional context all shape how information is received. Dashboards and models that are technically sound may still fail if they do not align with user realities.
He also argues for reproducibility and transparency. Sharing code and making analytical processes visible strengthens collective learning and reduces duplication. In environments where trust is central, openness becomes part of the delivery model rather than an optional extra.
Here again, the limiting factor is rarely technical capability. It is whether organisations design with users and accountability in mind.
A Converging Pattern
Taken together, these conversations illustrate a broader shift.
Embedding accessibility requires early structural consideration. Sustaining AI initiatives requires shared ownership and production discipline. Developing talent requires financial and cultural investment. Delivering analytics in healthcare requires transparency and user-centred design. Operating high-performance cloud systems requires automation grounded in practical constraints.
In each case, technical tools are advancing quickly. Organisational readiness determines whether those tools translate into durable value.
The next phase of digital maturity will not be defined solely by new models or platforms. It will be shaped by whether institutions are prepared to evolve their governance, culture and operating practices at the same pace as their technology.
Continuing the Conversation
These interviews form part of the wider conversations taking place around Tech Show London 2026, but the questions they raise extend well beyond any single event. They reflect the realities facing organisations across cloud, AI, data, infrastructure and talent today.
The technology is advancing quickly. The real work lies in ensuring institutions evolve with it.