News | Reflections

We have learned to be source-critical. The problem is that source criticism requires time, engagement and focus — while information is being produced faster than ever. In a stream of mass-produced texts, images and “insights”, it is no longer sufficient to ask who is saying something. The pace is simply too high.

AI does not solve this. It can formulate, summarise and visualise, but it cannot understand intent or carry responsibility. AI wants nothing, and it cannot determine whether something is true, relevant or manipulative. It reproduces probabilities. Using AI as a truth filter is therefore a mistake.

At the same time, objectivity has acquired a price. Sponsored text and imagery are not necessarily propaganda, but they are rarely neutral. The boundary between journalism, opinion and marketing becomes increasingly blurred as attention turns into hard currency.

In theory, we can always go back to the primary source. In practice, almost no one does. This does not make people lazy — it makes the system vulnerable. The focus must therefore shift from slow source criticism to rapid source referencing, source data and traceability. We must be just as quick to examine origins as we are to type a prompt into an AI. New technology is not needed to determine the truth, but to discipline information quality.

The 2026 trend is not about believing less — it is about demanding better origins, in real time.

Most AI initiatives begin with an answer.
The problem is that no one has yet asked the right question.

Across many organisations, we see the same pattern. A new tool is introduced, a model is tested, a solution is rolled out. AI produces suggestions, insights and recommendations at speed. Yet something peculiar happens: decisions do not improve, customers are no more satisfied, and the business is no clearer. What is missing is rarely intelligence — it is curiosity.

• We have become very good at explaining what we can deliver.
• We are far less good at understanding which problem actually needs to be solved.

In a recent consulting engagement, an advanced AI solution for customer dialogue was presented. Technically, it worked. But the conversations missed the mark. When we stepped back, it became clear that the client did not need better answers — they needed better questions. What are their customers worried about right now? What are they really trying to achieve? Where is the friction in everyday operations?

Once those questions were clarified, both the solution and the outcome changed. This is where many organisations go wrong: they allow AI to speak before they have fully listened to the customer. Solutions are presented too early, often because they are familiar, proven and easy to explain — not because they actually address the customer’s problem as defined by the customer.

The organisations that succeed are not those with the greatest AI capability, but those that ask the best questions. Those that begin with the customer’s reality, not with a model’s output. And those that understand AI as a tool for exploring options — not as a substitute for judgement and experience.

For a long time, marketing has been about reach. Messages have been broadcast widely, often in the hope that something might lead to business. The consequences are well known: large numbers of uninterested recipients, irritation among audiences who perceive the communication as irrelevant, and ultimately unsubscribes, blocking and a weakened brand. Many companies have described this type of marketing as a necessary evil — “you have to be visible”. The real cost, however, has not been distribution, but lost trust.

A few years ago, many organisations took a step forward. Instead of firing a scattergun, customers were grouped into more homogeneous segments. Compared to mass marketing, this felt like a breakthrough. Accuracy improved, response rates increased and irritation decreased. In practice, however, the underlying logic remained the same: one message per group — the only difference being that the groups became slightly smaller.

The real shift occurs when organisations move beyond segmentation and begin to treat communication as individual. Not personalisation in the trivial sense — “Hi Anders” — but individualisation in content, timing and intent. Marketing moves away from broadcasting messages towards asking relevant questions. What is this company interested in right now? What has changed in their business? What do we already know — and what do we genuinely need to ask?

Modern tools, including AI, make individualised communication possible at scale. In my experience, however, technology is rarely the limiting factor. Structure and data quality are. Without clear customer records, history, consent and an understanding of the relationship, individualisation risks becoming arbitrary — or, at worst, intrusive. When done properly, communication is no longer perceived as marketing, but as relevant dialogue.

As communication becomes individual, the way marketing is measured also changes. Campaigns become less important; relationships become more important. Instead of open rates and clicks, questions like these move to the centre:
• Does the communication lead to better conversations?
• Are sales cycles shortened?
• Does trust increase over time?

Marketing ceases to be a separate machine and instead becomes an integrated part of the business process.

You may also want to read about my programme:

In many conversations with company owners and executives, the same pattern emerges. Once the decision has been taken to explore a sale, finding buyers is rarely difficult. Industrial players and financial investors are usually available. What quickly becomes apparent, however, is how demanding and frustrating the sales process can be for those who are unprepared. The problem is seldom the business idea itself, but the structure surrounding the business.

Many underestimate how little buyers care about presentations and how much they care about underlying evidence. A sale is not a sales pitch; it is a systematic examination of the company’s reality: revenues, costs, customer relationships, contracts, history, risks and dependencies. Owners often experience the process as unnecessarily burdensome, not because the business is weak, but because answers cannot be produced without manual work, interpretation and caveats.

A recurring source of frustration is that the information exists, but does not fit together. Customer lists in one system, contracts in another, revenues in Excel and accounting data in a third. Historical data that cannot be reconstructed, and key figures that require explanation after the fact. For the buyer, this signals not necessarily a poor business, but elevated risk.

In practice, preparation is about credibility. A well-prepared company does not need to be perfect, but it must be comprehensible. When structure, data and documentation align, the dialogue becomes rational. When they do not, every question turns into a discussion, and every discussion introduces uncertainty that affects price, terms and timing.

For many owners, this requires a mental shift: from viewing the company from the inside to seeing it through the buyer’s eyes. What works in day-to-day operations is not always sufficient when the business must stand on its own, without verbal explanations.

The conclusion is clear. Finding buyers is rarely the hard part. Being ready when the questions begin is. Preparation is not about presenting the business attractively, but about building a company that can be understood, scrutinised and taken over. That is why the most successful transactions often begin long before any formal process is initiated.

You may read more on this topic here:

Läs gärna mer om detta ämna här:

• Second Opinion on transactions
• Prepare your company or investment—for a successful transaction

Excel is not wrong — it is simply insufficient. In a world where business data is complete, historical and machine-readable, reading summaries is no longer enough. A CEO who still leads through spreadsheets operates on a reduced version of reality, relying on statistics instead of facts and assumptions instead of tested scenarios.

With databases, programmable analytics and AI as a practical aide, there is no longer any reasonable justification for making decisions without first simulating the data and conducting impact analyses. Pricing, discounts, cash flows and margins can be tested against real accounting data before decisions are taken. This does not require the CEO to become a developer — but it does require an understanding of how data is actually generated and processed.

Excel was once leadership’s most important tool. Today, it is primarily a presentation layer, and expectations have shifted accordingly. When AI can write, explain and review code, lack of technical understanding is no longer a technological limitation — it is a leadership choice.

In many organisations, the ambition to increase efficiency and automation is high, yet the frustration remains the same: figures that cannot be trusted, manual tasks that persist, and uncertainty about what data actually exists. When these issues are examined in depth, they rarely prove to be a matter of tools or skills. Instead, data quality emerges as the true bottleneck.

In several en_gagements, GDPR has acted as a catalyst rather than an objective. When organisations are required to account for which personal data is stored, why it exists, and how long it is retained, it becomes clear that data has evolved without coherent ownership. Duplicates, outdated roles and unclear consent are not exceptions, but a structural condition.

The fundamental principles of data protection — accuracy, relevance and data minimisation — are often described as legal requirements. In practice, they function as enablers of collaboration. When it is clear what data exists, how it may be shared, and for what purpose, organisations can work securely with a wider range of external parties: suppliers, consultants, platforms and analytics providers. Uncertainty is replaced by clearly defined boundaries.

A frequently overlooked effect of data cleaning is precisely this: freedom of action. Data with a clear structure, documented justification and known limitations can be selectively shared, constrained and used where it creates value. In several projects, risk has been significantly reduced while collaboration has increased — simply by removing information that lacks a legal or business rationale.

In practice, data cleaning is therefore not a project, but a mindset. A way of defining what the organisation knows, what it can share, and what it no longer needs to carry forward. When this is in place, GDPR ceases to be a list of prohibitions and instead becomes a framework that enables faster, clearer work with more partners — without compromising accountability.

Data cleaning is thus not a concession to regulatory compliance. It is a prerequisite for rational decision-making and for confidently collaborating in a data-driven reality.

In consulting engagements, I increasingly see traditional organisational structures being challenged. Not because people have become redundant, but because the division of labour no longer makes sense. As AI and digital tools become accessible to more people, the conditions for how work can be organised fundamentally change.

Previously, roles were often defined by function: marketing, finance, operations, IT. Skills and responsibility followed organisational charts, and in practice this meant that many ideas were never realised — not because they were poor ideas, but because they “fell between the cracks”.

With AI as support, a different logic is now emerging. Organisations are beginning to operate more like consultant networks, even internally. The key question is no longer what title someone holds, but what they are genuinely good at — and what they want to amplify.

A CEO with a strong interest in marketing can, with the right AI support, work more hands-on with messaging, analysis and campaign logic without having to go through multiple layers. A warehouse worker with an aptitude for optimisation can, using data and tools, contribute to improved flows and reduced costs — without needing to become a software engineer.

In this context, AI acts as a wingman, not a replacement. It amplifies individual strengths rather than forcing everyone into the same mould.

I see this leading to a loosening of traditional hierarchies, where decision-making moves closer to those who actually understand the problem. Organisations become less dependent on information being “reported upwards” and more dependent on the right people having the right tools. This does not mean leadership disappears. On the contrary, leadership becomes clearer: setting direction, prioritising, and deciding who should work on what — rather than micromanaging how the work is done.

What is particularly interesting is how closely this mirrors the way consultants have always worked. Companies engage expertise to solve a specific problem, for a limited period, with a clear mandate. AI now makes it possible to apply the same logic internally. Employees can step into different roles as needs arise, rather than being locked into job descriptions, titles or traditional organisational doctrine. This does, however, require maturity. Without clear direction and accountability, organisations risk fragmentation — and flexibility quickly turns into chaos.

The decisive question, therefore, is not whether an organisation adopts AI, but how it allows AI to shape its organisational form. The issue is no longer simply which systems are used, but how people are empowered.

So, who is your AI wingman?

Is it a tool for marketing, analysis, optimisation or decision-making?

And are the right people using it — or are new possibilities being forced into old organisational boxes?

Most organisations assume they own their data, but in reality the situation is often more complex. When enterprise systems are introduced, they bring not only functionality but also a data model and an implicit view of how the business should be understood. This is where the conflict arises. While the company formally owns its information, the vendor frequently determines how that data may be structured, used and interpreted.

Across almost all businesses, the same core datasets recur: transactions, customers, suppliers, products/services, and financial accounts. These are commonly referred to as the five golden tables. The critical question is who controls their structure, history and accessibility. In several consulting engagements, we have seen enterprise systems where data is stored correctly, yet the company lacks real control over relationships, historical continuity and data extraction. Operationally, the organisation becomes dependent on the vendor’s model, licences and constraints. The data exists – but it is not free.

This becomes particularly problematic during system migrations, mergers and acquisitions, new business models, regulatory demands, or AI and analytics initiatives. When system logic begins to replace business logic, a subtle shift occurs. What does not fit the system’s data model becomes difficult to analyse and, eventually, difficult even to articulate. Pricing is simplified to suit the system, services are packaged according to available fields rather than customer needs, and analysis relies on manual workarounds. The enterprise system no longer supports the business; it becomes its interpretive frame.

In more mature data environments, we observe a different approach. The core tables are owned by the business within an independent, neutral data model. Enterprise systems act as producers and consumers of data, not as its owners. This allows the data model to survive system changes, preserves historical continuity, and ensures that analytics and AI are not constrained by a single system. The company retains control over its own concepts. Financial accounting then serves as the reference point, but not as a prison. When accounting data is connected to the wider business dataset, genuine governance emerges. When isolated, it is reduced to retrospective control.

The dispute over data ownership is therefore not merely an IT issue, but a strategic choice. It concerns who holds interpretive authority over the business – the company itself, or the system vendor.

Experience from numerous consulting engagements is clear: long-term freedom of action requires that the company owns the structure of its data.

Systems are necessary – but replaceable.

At KLRAB, we work daily with technologies such as AI, integration and data analytics — but we don't believe in smokescreens or overhyped buzzwords. AI is, at its core, statistics and pattern recognition. That doesn't make AI any less useful -- but it doesn't make it intelligent either. Teaching an AI Västgötgötska is considerably harder than teaching it Python. Why? Because dialects, like real activities, are full of nuances, context and exceptions. That's where AI has it most difficult—understanding human variation and business logic.

At KLRAB, we have therefore developed AnyNode — a methodology that combines business strategy with technical integration, where AI is part of the toolbox — not the whole solution. AnyNode focuses on building digital ecosystems that support the direction of business, not the other way around.

We are inspired by actors like TIC, who clearly and transparently explain how AI actually works — without promising too much. That's the kind of honesty we want to see more of in the AI conversation.

📣 We have a simple principle:

• Use technology where it works.
• Don't sell AI as magic.
• And above all — understand the business before you automate it.

We believe that this is how sustainable and smart digital solutions are built.

The annual audit fulfills an important function: it ensures that the financial statements are accurate and comply with the law.
But it does not say anything about how the company is doing — commercially, strategically, organizationally or technically.

It does not tell you about the risks in the business model, dependencies in the customer base, outdated infrastructure or untapped potential.

This is where me and KLRAB come in.
While the auditor reviews what has been, a strategic due diligence focuses on the future. I look at business model, growth capability, risk profile, financial quality, AI structure and technical capabilities — all of which determine how the business actually performs, and what stands in the way of the next step.

Inspecting the company strategically every two or three years should be as obvious as calling in the auditor.

I offer two clear formats:

• The 100-day program: from current analysis to concrete change
• Tansactional strategy: where value and risk need to be clarified prior to acquisition, sale or change of ownership

Want to know how your company is doing behind the numbers? Get in touch.
I inspect more than accounting — and do it strategically, discreetly and concretely.

Daniel Källberg, former CEO of a listed company and experienced management consultant, launches Kinnekulle LedningsAB (KLRAB), based in Stockholm with roots in Kinnekulle. The company offers strategic advice to management teams, boards and owners with a focus on digitalization, artificial intelligence (AI), IT strategy, sourcing and transformation management.

KLRAB combines classic management consulting with modern AI tools to help organizations navigate a rapidly changing business environment. By integrating AI into consulting, the company can offer data-driven insights and innovative solutions that strengthen clients' competitiveness. Daniel Källberg's own experience spans publicly traded groups, family businesses and entrepreneur-driven companies, which gives KLRAB a broad perspective in working with different types of operations.

KLRAB strives to be an extension of the client's management team rather than a traditional consultant. The focus is on building long-term partnerships where the advice is characterized by integrity, professionalism and measurable results.

To offer broad expertise and tailor-made solutions, KLRAB cooperates with a network of selected specialists in areas such as AI, brand strategy, business development and sustainability. This model ensures that each client receives the right skills for their unique challenges and support throughout the change journey — from initial analysis to final implementation.

Among the services offered by KLRAB is a structured 100-day program that helps organizations identify opportunities, increase efficiency, and achieve measurable improvements. Furthermore, inbound marketing strategies adapted to the digital shift, transactional strategy advice and objective second opinion assessments are provided for important business decisions. KLRAB also offers interim solutions where experienced leaders can temporarily step into key roles in the client's organization.

Integrating AI into consulting is crucial for companies to maintain their competitiveness in the future,” says Daniel Källberg, founder and CEO of Kinnekulle LedningsAB. “With KLRAB, I want to give our clients the best of both worlds — solid strategic expertise combined with the latest AI tools — to create lasting business benefit.

The launch comes at a time when digitalisation is accelerating and AI is high on the agenda in the business world. Källberg stresses that it is crucial for both management teams and owners to understand the potential of AI to ensure continued competitiveness and growth.