The obsession with automation is hiding the real change
For a long time, artificial intelligence has been presented almost exclusively as a tool for automating tasks and reducing operating costs. The dominant narrative revolved around autonomous processes, the elimination of manual work and immediate productivity gains.
However, when companies start to implement AI in their day-to-day operations, reality is usually quite different. Automating complex processes is not as simple as connecting a language model and waiting for results: the MIT NANDA report itself points out that the main cause of this failure is not the quality of the models, but the lack of integration with existing systems and processes.
Many internal procedures still depend on human decisions, interpretation, accumulated experience or information scattered across multiple systems and documents. In those scenarios, AI cannot work properly unless the organisation already has a clear structure for how it works and how it manages its knowledge.
That is why the real change that is beginning to take place is not just about automating tasks, but about building organisations capable of working alongside artificial intelligence in a much more integrated and efficient way. At LurNova we tackle this phase, prior to any automation, as part of our AI consultancy services.
Why does AI not solve internal organisation problems on its own?
Many companies try to add advanced automation on top of internal structures that still depend on the informal knowledge of certain people, scattered documentation or poorly defined procedures. And that is where the problems begin.
Artificial intelligence does not automatically organise a company. If information is disorganised, documents do not follow consistent criteria or processes keep changing depending on the department or the people carrying them out, automation becomes far more complex than initially expected.
In quite a few organisations, much of the knowledge still lives in unstructured folders, emails, historical files or simply in the accumulated experience of certain employees. This means AI has no reliable context to work with and forces companies to face an uncomfortable reality: before automating, they first need to put their own operations in order.
That is why a good number of artificial intelligence projects also end up becoming internal organisation projects. Knowledge must be structured, documentation correctly classified and processes clearly defined so that the technology can deliver value consistently and sustainably.
Many companies still do not really know how they work
One of the most interesting effects of artificial intelligence is that it forces organisations to look at themselves in more detail. When a company tries to automate a process, questions that previously went unnoticed quickly arise: what information is really involved, which decisions are objective and which depend on human interpretation, which tasks follow a clear procedure and which work out of habit, or which departments share information and which work in isolation.
That is where many companies discover something important: much of their real operations had never been fully documented. For years, many processes have worked thanks to the tacit knowledge of certain employees, informal relationships between teams or decisions based on experience that is hard to transfer. McKinsey's report "The State of AI in 2025" confirms this idea on a global scale: a large share of business processes still depends on tacit knowledge that is neither documented nor available in an executable format, which limits the ability to scale agents and automation.
Artificial intelligence is forcing organisations to map processes, structure knowledge and define internal criteria that had remained implicit for years. In a way, AI is also acting as an accelerator of organisational maturity.
Corporate knowledge is becoming the most important asset
For years, companies have invested enormous amounts of resources in ERP, CRM, management software or data systems. However, much of the real operational knowledge still depended on specific people who accumulated experience, context and the ability to interpret.
The problem arises when that knowledge is neither structured nor accessible to the rest of the organisation. Many companies discover too late that part of how they operate depends on specific profiles that are hard to replace and that much of their internal value is not actually documented.
Artificial intelligence is changing that perception. For the first time, many companies are starting to understand that their real differentiating value lies not only in the software they use, but in the internal knowledge accumulated over years of experience.
And precisely for that reason, the companies that make the best use of AI will probably not be those that automate the most, but those best able to organise, protect and activate their corporate knowledge and put it at the service of the whole organisation.
The greatest return sometimes comes from much simpler solutions
There is a perception that the impact of artificial intelligence only appears when a company manages to automate entire processes from start to finish. In practice, however, the fastest and most profitable return usually comes in a different way.
Many companies are starting to obtain real results through intelligent assistants capable of helping teams in their daily work: locating internal information in seconds, consulting procedures, summarising technical documentation, preparing proposals, analysing regulations, drafting reports or resolving operational queries without constantly depending on other people.
This is where one of the most significant changes that AI is bringing about within organisations appears. It stops being merely a system designed to automate tasks and starts to become a permanent support layer for work teams.
These kinds of solutions tend to be adopted much more naturally because employees see them as useful tools for working better, not as systems imposed to replace human roles. In addition, the cumulative effect across hundreds of small daily tasks ends up generating significant operational improvements without having to tackle extremely complex projects from the outset.
What is RAG and why does it change the role of corporate assistants?
One of the most common mistakes when talking about artificial intelligence in companies is to think that all assistants work the same way. Many organisations associate AI solely with general-purpose tools capable of answering open questions, but the real business leap begins when these systems are connected to the company's internal knowledge, through technologies such as RAG systems (Retrieval-Augmented Generation) or knowledge graphs.
Instead of answering only with generic information learnt during the model's training, an assistant with RAG consults corporate documentation in real time (internal procedures, regulations, technical manuals, project histories) before generating an answer. The assistant stops being a simple chatbot and starts to behave like an intelligent interface over corporate knowledge, contextualising each answer with specific, validated information from the organisation. This is exactly the kind of architecture we deploy in our RAG solutions.
Knowledge graphs make it possible to go even further: they connect not only documents but also relationships between people, processes, regulations, customers, products, incidents or projects. This allows AI to understand how information is really related within the company and to provide much more accurate answers.
For example, an assistant connected through RAG and corporate graphs can identify which procedures relate to a specific regulation, which departments are affected by a technical change, which previous projects contain similar cases or what documentation an employee needs to resolve a specific incident. That is where one of the biggest changes that artificial intelligence will bring to companies begins: turning internal knowledge into something accessible, searchable and usable in real time.
Collaboration between people and AI will be the real change
For years, much of the imagery surrounding artificial intelligence has been related to the replacement of human work. However, the model that is starting to prevail in organisations is different: AI does not come in only to replace tasks, but to collaborate continuously with people through AI agents capable of searching for information, analysing documentation, generating content, summarising data or supporting decision-making in real time.
This transforms the operational dynamics of companies. Teams no longer spend hours looking for scattered information or carrying out repetitive tasks and can focus on higher-value activities, such as analysis, supervision, creativity or customer relationships.
In addition, this model helps to reduce one of the major risks many organisations currently face: the loss of corporate knowledge caused by retirements, staff turnover or excessive dependence on certain key profiles.
Automating without control can create more dependence than efficiency
Many companies are adopting artificial intelligence tools very quickly, but without a clear strategy behind them. Different departments use different platforms, employees bring in external tools on their own initiative and much of the real use of AI happens outside any corporate policy.
This creates a scenario in which the organisation loses visibility over which tools are being used, what information is being shared and what dependence on external platforms is starting to build up. The more AI is integrated into daily processes, the more important it becomes to keep control over that technological infrastructure, because dependence is not only technical: it can also be operational, economic and strategic.
Some companies could end up building part of their daily operations on tools they do not control, external models over which they have no supervisory capacity or platforms whose operating criteria can change at any time.
Automating with public models also carries risks
At the same time, another phenomenon is quietly growing within organisations: the uncontrolled use of general-purpose AI tools to automate internal tasks or consult corporate information.
It is increasingly common for employees to enter contracts, technical documentation, internal procedures, customer data or sensitive information into external platforms without really knowing what happens to that information. The problem is not only technological, but strategic: part of the most valuable corporate knowledge begins to leave the organisation in a distributed way that is often invisible to the company itself.
In addition, using general-purpose AI models to automate complex corporate processes introduces another problem: the lack of real business context. These models do not know the internal operations, the specific criteria of each company or the particularities of its processes. That is why automating directly on public tools without a layer of control, adaptation and governance can lead to errors, inconsistencies and operational risks that are hard to detect.
The future will be hybrid: automation, assistants and private AI
The natural evolution of artificial intelligence within companies will probably not be a scenario in which everything is fully automated, but rather hybrid environments in which automation, intelligent assistants and human supervision coexist.
There will be highly automated processes, but also tools designed to increase people's operational capacity and make it easier to access corporate knowledge. This balance will allow organisations to take advantage of the speed of automation without losing the human judgement needed to supervise complex or contextual decisions.
For this to work securely, many organisations will end up needing private AI environments capable of keeping control over their data, their processes and their internal knowledge. Because the question is no longer just what artificial intelligence can do, but how to integrate it into the organisation without losing control over what makes each company unique.
CEO of LurNova, with extensive experience in technology and digital transformation. Involved in developing technology projects and applying artificial intelligence in organisations, he combines technical knowledge with business vision.
