The Network Automation Parallel: Decoding AI Adoption Part III

What we as people do every day in our work and personal lives and how we go about doing those things constitute a large part of our culture. As such, when we talk about how technology changes our daily activities at work, we are really talking about culture. Failing to consider this cultural component is a common reason why new technology implementations fail.

This blog, part 3 in the series that examines and predicts how Artificial Intelligence (AI) will be adopted, focuses on cultural acceptance of AI by examining what we have learned about cultural acceptance around network automation. Both technologies fundamentally change what we do and how we do it, making this a lesson in managing cultural change.

For a successful network and/or infrastructure automation program to succeed long-term, a few things must be true: 

• Employees must trust the technology to do what the technology is built to do
• Employees must have a clear understanding of the impact that the automation will have on their day-to-day activities
• Building on the prior point, employees must also have an understanding of what their role will look like in the future, as the technology performs the repetitive, high-volume work

Lessons from Network Automation

Trust

Network Engineers must be confident that automation will correctly perform tasks such as implementing device configuration changes. Without that trust, they simply won’t use it, and the initiative will wither. One key to building this trust is training. When engineers understand the mechanics of how automation works, they are far more likely to rely on it.

Skill Set Expression

Automation empowers Network Engineers to express their most valuable skills. Repetitive and laborious, high-volume tasks like manually configuring devices or performing OS upgrades are perfect candidates for automation. Freeing engineers from this repetitive work allows them to focus on higher-value activities that reduce technical debt, such as:

• Focusing on proper long-term network design
• Implementing permanent fixes instead of temporary band-aids
• Collaborating with development teams to create and update configuration templates

In some cases, engineers can even transition into network automation development roles. 

Applying the Lessons to AI

There is already a lot of conversation around how AI will affect our culture at large. AI’s impact on our work culture is a very important subset of that larger discussion.

Trust

As with automation, employees must trust that AI will do what it’s designed to do. However, AI adds a layer of complexity. While most network automation is deterministic (the same input always produces the same output), some AI is probabilistic, meaning it incorporates randomness and can produce different outputs.

This non-deterministic nature can create problems with trust. For example, an AI coding assistant recently made headlines for deleting a company’s entire database and then trying to hide its tracks. This highlights a crucial point: it’s not just about trusting the AI itself, but also trusting its implementation, including the safeguards built around it. For AI to be a trusted partner in business decisions, its deployment must be explicitly managed.

Properly implementing AI is important so that it can be trusted to assist in making day-to-day and important business decisions.

Skill Set Expression

Employees in an environment with a comprehensive AI architecture will also see the expression of their skill sets change. As with network automation, the employee will still need the skill set, but it will be expressed differently. For example, where a worker used to perform a task manually, they may now find themselves supervising a fleet of AI agents doing that specific type of work at scale. 

AI Predictions

• Extensive AI deployment will require establishing trust in AI models and their safeguards. This may be complex due to the probabilistic nature of some AI models.
• Organizations need a clear AI training program for all employees, teaching them how AI functions, when to use it, how to interpret its results, and how it will transform their roles.
• Leadership must treat AI adoption as a marathon, not a sprint, requiring consistent guidance and support every step of the way.