Cybersecurity and AI: the investment that boosts your results
Spanish executives are clear: without more investment in cybersecurity, AI will not deliver the expected results. We analyze what this means for an SME.
5 min read

In this article
Accenture presented some conclusions from its Pulse of Change report at its Cybersecurity Reinvention Center in Madrid, and it did so with a message that should resonate in any SME that is bringing AI into its processes: security is no longer a brake, it is the lever that makes artificial intelligence actually work. The data handled by the consultancy point in a very specific direction, and at ForgeNEX we believe it is worth reading them calmly because they directly affect the day-to-day work of a small IT team.
What the numbers say (and why they matter in an SME)
According to the report, 69% of Spanish executives believe that increasing investment in cybersecurity and risk management would improve their organizations' results in AI. That figure is ten points above the European average, which stands at 59%. Translated into more familiar language: more and more managers understand that a conversational assistant, a data analysis model or an automated agent cannot be deployed without first having a minimum security baseline.
55% of Spanish executives place AI-powered malware among their main concerns, and 57% place the strengthening of cyber infrastructure and resilience among their top three priorities, compared with 54% in Europe. This is not alarmism: it is acknowledging that the same engine that helps us automate tasks also helps attackers generate more credible phishing emails, malware variants or automated reconnaissance.
Agustín Muñoz Grandes, head of Cybersecurity at Accenture in Spain and Portugal, summarized it this way during the event: investment has been progressing, but with the massive deployment of AI there is a greater perception of the importance of increasing resources allocated to cybersecurity to accompany that investment. And he added that Spain generally has a high level of maturity, above the average of its environment in many indicators.
The Shadow AI problem: company data in third-party models
One of the points that most concerns executives is Shadow AI, the uncontrolled use of language models from different vendors within the organization. The executive himself acknowledged that there are employees who upload company data to AI models to request analysis or help, without a sufficient level of control. 47% of executives have doubts about whether their company manages this issue with the necessary rigor, which directly affects data security.
In an SME this is very easy to see: someone copies a customer list into a public tool so it can help them draft a proposal, or pastes a contract so it can summarize the clauses. The intention is good, but the risk to intellectual property and customer data is real. And you do not need to be a large corporation to suffer from it.
Accenture's recommendation is not to prohibit, but to accompany: it is not about putting a brake on development, but about ensuring that AI is implemented securely. At ForgeNEX we share that approach, because prohibiting outright only pushes usage into the shadows.
What the Spanish market is already doing
The report also offers clues about the measures already being taken. 54% of the executives surveyed state that their organization is improving its monitoring and detection systems, compared with 43% in Europe. 51% invest in AI-driven cybersecurity tools, compared with 43% in Europe. 40% strengthen the protection of models and of the data used to train them, and 39% incorporate AI into detection, investigation and threat response operations.
As for the objectives of cybersecurity strategies, 59% of Spanish executives prioritize reducing the risk of cyberattacks and breaches, compared with 50% in Europe; 54% highlight the protection of sensitive data, intellectual property and digital identities, compared with 49% in Europe; and 47% emphasize support for digital and AI transformation, compared with 44% in Europe.
What we recommend at ForgeNEX
Beyond the headlines, what this report hints at is a gap between intention and execution. Many SMEs want to do things right, but they do not always know where to start. These are the lines of work we propose:
- Define a clear and simple AI usage policy: which tools are allowed, with what data and for what purpose. You do not need a hundred-page manual; a document that the team understands and signs is enough.
- Classify information before connecting any model: what is public, what is internal and what is confidential. Without that classification, any control is worthless.
- Review AI providers with the same questions we would ask any critical provider. If you are considering AI-assisted pentesting, you can start with these 15 questions before choosing a provider.
- Monitor and detect: protecting the model is not enough; you have to watch what happens around it. Supply chain incidents, such as the cyberattack on the cold chain, show that risk comes from where you least expect it.
- Train the team in the responsible use of AI, with real day-to-day examples and not abstract theory.
The Accenture event ended with a real-time cybersecurity crisis simulation, a very graphic way of reminding us that a coordinated response matters as much as prevention. In an SME, that coordination is usually simpler because teams are small, but also more fragile if there are no written procedures.
The conclusion we draw is clear: AI is not going to wait for security to be resolved, and security cannot become the brake that prevents taking advantage of AI. The balance lies in investing with judgment, starting with the basics and scaling according to each organization's real risk.
Source: Computerworld. Analysis and adaptation: ForgeNEX.
Keep reading
- AI Pentesting: 15 Questions Before Choosing a Vendor18 November 2026
- AI Agents Hunt for API Keys in Exposed Apps: Stolen Keys17 November 2026
- SANSAR: Linux malware that hides GPU mining in your SMB16 November 2026
- Nvidia without Jensen Huang: what it means for your SME14 November 2026
- TypoPrompting: 186 fake AI domains that hijack your browser13 November 2026