Four challenges every organisation must solve to adopt AI safely

By James Thomas | 26 August 2026

Confident AI starts with security

As organisations embrace AI, many are facing the same challenge: how do you give employees the freedom to use AI tools while ensuring sensitive business data remains protected?

It’s a bit like opening up a new motorway network. Every new route helps people move faster, but it also becomes harder to see where traffic is heading, which roads are being used and whether drivers are staying within the rules.

That is the challenge many organisations face as AI adoption accelerates. In our demo-led webinar, Confident AI Starts with Security, we explored how greater visibility and smarter controls can help businesses embrace AI with confidence, rather than relying on policies alone. Using Cisco Secure Access, a cloud-delivered security platform, we demonstrated how organisations can gain that visibility and apply more effective controls.

Below, we’ll explore four key highlights from our Cisco Secure Access demonstration that show how organisations can gain visibility of AI usage, protect sensitive information and enable AI adoption without sacrificing security.

1. You can’t secure what you can’t see

For many organisations, the challenge isn’t deciding whether AI is being used. It’s understanding where that usage is happening and which AI applications are already active across the business.

Without that visibility, policies, guardrails and risk discussions are often built on assumptions rather than evidence. The demonstration below shows how Cisco Secure Access helps organisations uncover AI activity across their environment, providing the foundation for more informed governance decisions.

Instead of debating whether shadow AI might exist somewhere in the business, organisations can make decisions based on evidence. That visibility helps IT teams prioritise where to focus their efforts, review new AI applications more effectively and build policies that reflect how AI is actually being used across the organisation.

2. Control that keeps pace with AI

Visibility is only part of the equation. Once you understand how AI is being used across the business, the next challenge is ensuring the right safeguards are in place around that activity.

Just like the motorway analogy, visibility shows you where the traffic is flowing. Control is what helps ensure it stays on the right route. As AI becomes embedded into more applications and workflows, organisations need security controls that can adapt to new ways of working without forcing them to start from scratch.

What this demonstrates is that AI governance doesn’t need to sit outside of an organisation’s existing security strategy. Instead, it can build upon the controls and policies already in place, extending them to cover new AI-related risks. That makes AI adoption far more achievable, allowing organisations to strengthen governance while continuing to support innovation.

3. What happens when users find another route

Most security incidents don’t happen because somebody sets out to break the rules. More often, people are simply trying to get a job done and will look for an alternative route when the first option isn’t available.

That creates a challenge for AI governance. If an employee is prevented from sharing information one way, what happens if they try a different route instead? The demonstration below explores how organisations can protect sensitive information consistently, even when users change the method they’re using to work.

What makes this example valuable is how closely it reflects real-world behaviour. The challenge isn’t stopping people from solving problems. It’s ensuring sensitive information remains protected regardless of how those problems are solved. As organisations adopt more AI tools and workflows, that distinction becomes increasingly important. The format may change, but the information being protected remains the same.

4. Find the right balance between productivity and protection

One of the biggest challenges facing organisations today is finding the right balance between productivity and protection. Employees want to take advantage of AI, but businesses also need confidence that sensitive information remains secure.

The question isn’t whether AI should be used. It’s where organisations draw the line between acceptable use and unnecessary risk. The demonstration below shows how Cisco Secure Access helps create those boundaries, allowing employees to benefit from AI while helping protect information that should never leave the business.

The key takeaway is that not every AI interaction carries the same level of risk. Organisations don’t need to block AI entirely to reduce risk. Instead, they need the ability to apply the right level of control to the right type of information. When those boundaries are clear and consistently enforced, employees can embrace AI with confidence and organisations can support innovation without compromising security.

Four questions that reveal your AI readiness

The four demonstrations explored different aspects of AI governance, but they all point back to the same question: how prepared is your organisation for AI adoption in the real world?

As AI adoption continues to accelerate, every organisation should be able to answer:

Do we know which AI tools are being used today?
Can our controls adapt to modern AI workflows?
Will our policies still work when users take alternative routes?
Can we enable AI safely without blocking productivity?

Organisations that can answer “yes” to all four are often the ones adopting AI most confidently.

The challenge for many businesses is that they don’t yet have the visibility needed to answer those questions with certainty. Knowing where AI is being used, how information is flowing and whether existing controls remain effective is often the first step towards building a more secure AI strategy.

That’s where an AI Readiness Assessment can help. By establishing a clear baseline of AI usage across your organisation, you can identify opportunities, understand potential risks and explore whether capabilities such as Cisco Secure Access could strengthen your approach to AI governance.

If you are ready to understand how AI is being used across your organisation, book our complimentary AI Readiness Assessment and explore whether a Cisco Secure Access trial could help you adopt AI with greater confidence.

Book your Data and AI Readiness Assessment

James Thomas, Solutions Architect

James Thomas brings over 10 years of experience helping organisations design and deliver secure, scalable network and infrastructure solutions. As a Solutions Architect within Bistech’s Professional Services team, he takes a hands-on approach to turning technical designs into real-world outcomes. He ensures every solution is practical, secure, and aligned to long-term operational needs. By combining deep technical expertise with a clear focus on delivery, James helps customers build resilient environments that support performance, security, and growth.