How Deskpro Private Breaks Down the Barriers to Secure Help Desk AI: Interview with CEO Brad Murdoch

AI adoption is accelerating across support operations, but support teams in regulated industries face a tough choice: use AI in a limited capacity due to compliance requirements, or don’t use it at all–at least until they can bridge the divide between AI innovation and security. Deskpro CEO Brad Murdoch discusses how support organizations are approaching AI today and how the challenges around AI, data security, and data sovereignty led to Deskpro Private.

What are the biggest barriers preventing enterprises from adopting AI in their support operations currently?

Two main things are holding enterprises back. The first is that AI isn’t fully effective without access to corporate data. Its success in support operations depends on having access to relevant indexed information. Otherwise, you're getting generic answers based on the large public dataset that the model was trained on–and that’s unlikely to do much to improve customer service or increase a human agent's productivity.

The second thing is related: many organizations have concerns about the security model in adopting AI, which is primarily delivered via public cloud services. With public AI models, data flows outside of an organization’s controlled environment. While that data is encrypted in transit, it’s unencrypted in the public cloud so the AI model can access it, introducing governance, data residency, and data leakage risks.

How is the current CX and EX technology market forcing companies to choose between innovation and compliance?

It's not the market so much as the reality most organizations are forced to deal with. For example, with Deskpro Cloud, our customers can either use Deskpro’s AI service, which is powered by OpenAI, or they can choose to “bring their own AI” that they have approved in their organization.

However, all of those models typically run in a public cloud environment delivered by a hyperscaler like AWS, Google Cloud, or Azure. While all of the hyperscalers invest significantly in security, the fact that the data needs to be unencrypted to be processed by the AI foundation models in the public cloud can be a huge challenge.

There are many industries, particularly regulated ones such as financial services, healthcare, aerospace, defense, government, and critical infrastructure, where security and compliance are limiting factors in how these public cloud services can be used. So organizations in these industries either don’t allow AI usage or have to look at alternative ways to adopt AI.

Additionally, many organizations are subject to data sovereignty rules that would preclude the use of public cloud services located outside a specific sovereign or geographic region.

Until recently, these organizations were effectively prevented from using AI to power their help desk, whether for internal or external support. We're now breaking down the barriers to AI with Deskpro Private.

What is Deskpro Private?

Deskpro Private is our self-hosted help desk offering combined with the ability to support private AI: it can be configured in a virtual private cloud, sovereign cloud, on-premise, or even with an air-gapped deployment model. With our launch, we introduced the ability to connect any commercial, custom, or private AI model within those environments and have that AI drive the help desk AI features.

We began building Deskpro Private to address a gap we were seeing between our cloud and self-hosted customers. We introduced our AI features in 2024 and, before the launch of Deskpro Private, literally none of our self-hosted customers were using those AI features, whereas many of our cloud customers were.

As we worked through solving that challenge for our existing customers, it became clear there was a much broader market of organizations with the same security, compliance, and data sovereignty requirements as our current self-hosted customers. Deskpro Private is a growing opportunity to serve regulated industries and other organizations that have security, compliance, and data sovereignty requirements with a fully featured, AI-enabled help desk platform.

We recognize that one size does not fit all, and that is why we have built support for deployment in Virtual Private Clouds (VPC’s), Sovereign Clouds, private clouds and on premise in an organization’s own data center.

How does private AI change the way enterprises approach innovation?

Investment in AI is a very meaningful part of corporate strategy for many organizations. Whether for commercial reasons or security and compliance reasons, many are investing in private AI. They may be using commercial AI deployed privately such as Cohere or Mistral, or building on top of open-source models like Meta Llama and DeepSeek.

This investment in private AI can support many of their business functions, but if they use a help desk from anyone other than Deskpro, they're not going to be able to leverage their own private AI to drive their Customer Service and Employee Service functions. Being able to bridge those two worlds so that the investment an enterprise is already making in AI can actually be used to help them with their customer and employee experience is a very significant advantage.

The Deskpro team has been using the phrase “the help desk redefined.” What does this phrase mean to you?

When I think of a help desk, I think of submitting a form on a portal or sending an email. If I have direct access to a help desk, I'll go to a web portal, enter my details, and submit a query. Otherwise, I will email the help desk mailbox. That is the paradigm that most of us are used to and have been using for years.

The “Help Desk, Redefined,” is when AI enables a help desk to become a function rather than a product, and any user, customer, employee, or partner should be able to get direct support from the tools they are using on a day-to-day basis.

For example, at Deskpro, we use Slack. If I have a query and need to be able to ask our HR department, “How many days of PTO do I have left?, I should just be able to go into the HR channel on Slack and ask the question.

AI will recognize that question and determine if an AI agent can get this answer for me without bothering a human. If so, it'll say, “Brad, you have three days of PTO left this year,” right in my own Slack channel. If the AI determines that this is a more complex query that cannot be answered directly by an AI agent, then it will route it to a human agent who can also respond in Slack.

Of course, it’s not just Slack; it’s any channel that suits the user, such as Microsoft Teams, email, social channels, and so on. That is how the help desk is being redefined: it now becomes a function rather than a product.

How does Deskpro Private reflect the company's values and culture?

Deskpro has always been a company that has tried to provide a capability that is extremely flexible and highly customizable, truly serving the requirements of our customer base for either customer experience or employee experience.

This flexibility and customizability now extends to supporting private deployment with private AI with our recent Deskpro Private launch. This means that we are now enabling 100% of organizations to leverage AI in their support functions, not just those comfortable with using public cloud services.

The security model allows our customers to offer the best possible customer and employee experience by giving access to corporate data inside whatever security perimeter they are comfortable with. They do not have to adapt their security, compliance and data sovereignty needs to suit their help desk platform: the platform works for them.

As the ONLY help desk platform that enables full support of AI in private environments, Deskpro Private allows enterprises to connect to the AI of their choice while remaining confident that sensitive data will not leave their security perimeter. Learn how your team can stay secure and compliant with Deskpro Private by booking a demo with us today.