8 urgent updates your IT playbook needs to survive the AI era

Follow ZDNET: Add us as a preferred source on Google.

ZDNET’s key takeaways

• Your technology playbook could be outdated because of AI.
• Be prepared for revisions that help people check their ideas.
• Focus on key areas, such as use cases, data sources, and training.

Do you or your team use a technology playbook? If so, what’s in it? There’s a good chance your playbook is becoming rapidly outdated.

Also: 10 ways AI can inflict unprecedented damage in 2026

That’s the challenge posed by Thomas Erl, a prolific tech author and educator, in a recent interview with Matt Strippelhoff, partner and CEO at Red Hawk Technologies. Erl calls for fresh playbook revisions and tried-and-true practices to help AI proponents and developers vet their ideas, run safe pilots, and prove the return on investment of their projects.

Playbooks, whether formal or informal, detailed or simple checklists, ensure everyone works from the same page strategically for consistent operations and deployments, with strong security policies. However, in today’s fast-moving digital world, if you or your team is working with AI, you may need to revisit those guidelines.

Also: Nervous about the job market? 5 ways to stand out in the age of AI

A playbook for the 2026 enterprise has several new requirements, but it also builds on previous IT guidelines. Strippelhoff and Erl reviewed some of these considerations.

8 guidelines for the AI era

1. Start with a meaningful problem: Identify where AI will truly make a difference, versus AI for AI’s sake. “Some companies are looking for a way to apply AI, but they haven’t identified the problem they want to solve,” said Strippelhoff. “So, they have a solution looking for a problem. Traditional strategic planning is critical to make sure you’re identifying a meaningful problem.”
2. Start with the desired outcomes up front and prepare the business case: This approach was common for earlier technologies, but it takes on additional urgency with AI initiatives. “The most important piece is understanding the organization’s readiness for the idea itself,” said Strippelhoff. “Someone needs to take the time to craft and define what that vision is. Then you need to incorporate subject matter experts around those systems, data sources, and more, and determine if you are actually prepared, if it’s time to make that investment. Oftentimes, a lot of organizations are not quite as ready as they may think they are.”
3. Incorporate an additional layer of caution: AI isn’t just about building and running software. Deploying AI also means diving into an organization’s deepest wells of knowledge. Training data comes out of those wells, Strippelhoff said: “This also includes a means to validate the generated responses or what’s being produced by the AI.”
4. Build in space for exceptions: This area is where even the most well-planned AI systems can grind to a crawl. Insufficient data quality, for example, can create significant inconsistencies in AI outputs, Strippelhoff cautioned: “Exceptions in the quality of your data could create a lot of challenges for training the AI model.”
5. Include time for AI model training: People need assurance that training data is refreshed and accurate. For example, Strippelhoff said, in the healthcare sector, a wide array of billing codes makes automating revenue-cycle management challenging, “as there are thousands of codes to choose from.” As a result, the process needs to be closely monitored by humans until there is assurance that codes are properly classified through an ongoing feedback loop.
6. Make sure your data is ready: “Some companies may assume, with their digital assets, standard operating procedures, and governance, they are prepared to move forward with an AI initiative, only to find out that their data is in such a poor state that they have to ‘can’ the project. I’ve seen that lead to projects being stalled or permanently shelved.”
7. Keep humans in the loop, always: AI may seem synonymous with total automation, but that’s not the case. An essential part of the AI output validation process is maintaining human oversight at key points. This likely will be a “subject matter expert who validates the output,” said Strippelhoff, adding that “it takes time to train.”
8. Check for platform limitations: “If your solution is dependent on extracting and moving data across systems through API endpoints, there may be limitations to the number of calls, the availability, and the type of information that you can get, as well as the frequency at which you can get it,” he said.