Trust and resilience in the digital enterprise: five tech trends shaping 2027

As 2027 looms on the horizon, thoughts turn to what the next year might bring in digital trust and security. It’s hard to look beyond the impact of AI – and specifically autonomous AI agents. While this technology helps enterprises to improve productivity, it also introduces new security and governance concerns.

Early in 2026, the Thales 2026 Data Threat Report explored how the evolution of AI is changing the way organisations protect their data, systems and digital identities. Among its findings were:

• 61 percent of organisations rank AI as their top data security risk
• 59 percent have experienced deepfake attacks
• 48 percent report reputational damage linked to AI-generated misinformation
• 70 percent identify the pace of change in AI as a leading security concern.

The use of autonomous agents looks set to add more complexity to the threat landscape. Agentic systems will increasingly perform tasks and make decisions on behalf of people. In response, enterprises will need to control how agents identify themselves, what actions they are authorised to perform and how they access/handle data.

Gartner’s Top Strategic Trends for 2027 report offers a forward-looking perspective on these challenges. It analyses the security risk posed by AI systems and shows how organisations can “move from a passive observer of model capabilities to an active architect of organisation, human and machine boundaries.”

We inspected the report to reveal five take-aways relating to digital security and trust…

1. AI is moving from assistance to action

The agentic era is already here. According to Gartner’s analysis, AI agents in the enterprise are moving from passive assistance to active execution. Rather than just answering prompts, they are performing multi-step workflows and acting independently on behalf of individuals. 

Gartner predicts that by 2030, 50 percent of organisations will use AI agents with minimal human involvement in IT operations, compared with almost none in 2026.

This shift raises important questions about whether traditional identity and security models are equipped for increasingly autonomous systems. These systems were built for humans. Now, they have to answer fundamental questions such as:

• Is an agent or a human performing this action?
• Which human is the agent acting for?
• What can and can’t the agent do?
• How can we keep an audit trail?
• How can we revoke this agent?

In short, organisations must be able to trace every agentic action to a person, prompt and model. Gartner believes new AI ‘context platforms’ will mitigate the risk. They will give agents access to enterprise knowledge so they can make decisions based on organisational information rather than just a prompt. 

2. Security is growing more adaptive

In the coming agentic enterprise landscape, every agent must be assigned a verifiable identity, clear ownership and a set of permissions that are appropriate to its role. Without appropriate safeguards, organisations may become more exposed to misuse, unauthorised access and other security threats. Gartner forecasts that by 2028, AI agent abuse could be linked to 25 percent of enterprise security breaches.

One remedy is to make security more adaptive and contextual. The idea here is to develop systems that take into account the identity of the requester and the risk associated with the action. For example, a low-risk request could be granted without checks while an unusual request to access sensitive data will demand stronger verification.

Of course, AI tech can be used for defence too. Security teams can deploy AI tools to protect their systems. For example, in the report, Gartner highlights the ability of AI to fix vulnerabilities in real time. It says this kind of on-the-fly automation typically comprises elements including:

• Continuous detection of potential exposures
• Prioritising risk based on business criticality
• Automated remediation of patches
• Verifying that fixes have worked, vulnerabilities have been addressed and affected services remain operational

The priority for security leaders is not to remove humans from the process. Instead, they should reserve human attention for complex, high-impact decisions while automation handles routine tasks.

3. Digital sovereignty is becoming a business priority  

The expansion of cloud computing and AI has created new opportunities for efficiency. In the process, it has also increased organisations' reliance on external providers.

They now want to reduce their exposure. Enterprises are concerned about where data is stored, and what happens if a partner becomes inaccessible. They are also conscious of geopolitical uncertainty and regulatory changes. As a result, they are seeking greater visibility and control over their data – aka digital sovereignty. In the report, Gartner predicts that by 2030, more than 75 percent of organisations outside the US will adopt digital sovereignty strategies that comprise:

• Data control
  Greater visibility into where data is stored, processed and transferred, and which regulatory regimes apply to it. 
• Infrastructure and operational control
  Oversight of who operates data systems, has privileged access, holds encryption keys etc.
• Technology independence
  An understanding of how much the enterprise depends on tech partners to manage data, and how it can switch providers if access is disrupted.

Of course, sovereignty does not mean building everything in-house. 

The objective is autonomy, not independence. 

4. Trustworthy AI depends on secure data

Previous sections have set out the importance of controlling what AI agents can do. But what about the data they expose? How can organisations ensure this data is accurate, fit for purpose and secure? Effective data governance is essential to mitigate risks relating to:

• Training data
  If the data feeding a model is compromised, the AI’s actions will be flawed or vulnerable to malicious manipulation.
• Regulation
  AI systems must be built on data repositories that respect consent and data sovereignty. Regulations such as the EU Cyber Resilience Act, whose main obligations will apply from December 2027, are adding to the expectations around the security of digital products.

Strong data governance will ensure that AI models only process accurate, permitted information, and that this data does not expose intellectual property or personally identifiable information (PII).

5. Resilience is becoming a design principle 

This article has already explored how data sovereignty will be an important goal for enterprises across the rest of the decade. However, data sovereignty is just one facet of the wider need to improve organisational resilience in all domains. 

Gartner argues that enterprises should factor resilience into almost every big decision they make. It should not be an afterthought. Why? Because, thanks to corporate uncertainty, geopolitics, regulation, energy transition and more, enterprises can no longer rely wholly on existing systems and suppliers.

Gartner recommends a transition to what it calls ‘SovereignTech’. It describes this as “hardware, software, databases and AI models technically and legally architectured to offer realistic exit strategies from critical providers.” SovereignTech aims to help organisations maintain operational continuity and access to critical capabilities if a provider becomes unavailable.

But provider independence is only one facet of the approach. It also incentivises organisations to build security and recovery into every facet of their architectures. They should be able to make continuous technology changes without requiring wholesale replacement of systems and suppliers. 

In other words, SovereignTech turns resilience into a design principle. It directs organisations to actions such as:

• Evaluating on-prem alternatives to standard public clouds
• Using new generation technology to share capabilities with trusted partners
• Identifying emerging protocols and standards
• Minimising risk exposure. Assess whether a provider is willing to meet sovereign delivery options and offer no-kill switch assurances.

In the new world of autonomous non-human systems, organisations must rethink how they establish trust, maintain control and ensure resilience. The good news is that creative new solutions are emerging. The challenge for enterprises is to adapt to the new reality, stay secure and keep innovating.