![Intelligent augmentation cover art](http://www.zscaler.com/cxorevolutionaries/sites/default/files/images/blogs/AI%20augmentation%20-%20header%20image%20-%20Simpson.jpg)
With the hype around large language models (LLMs) like Chat-GPT, knowing what a business should do can take time. Here’s my take on what every business needs to do regarding artificial intelligence today.
First, every business needs an AI strategy. At a minimum, you must consciously decide if you will not pursue an AI strategy. Assuming you wouldn’t be reading this article if that was your decision, let’s look at my four-step process for an AI strategy.
- Pillars: First, Get more specific than just “AI”… AI is an entire field of solutions. What are your business objectives? I recommend breaking AI down into a few “pillars” that align with your business objectives. For example, if you are focused on cost out, maybe you have a pillar to focus on robotic process automation. If you have a large pool of data and a team of data scientists, machine learning might be one of your pillars for supporting any use cases, such as optimization and predictive analysis.
- Action: In addition to defining your pillars, define actionable projects with concrete business deliverables in less than six months. A good way for your AI strategy to be shut down is for it not to deliver any real value. You need pragmatic short-term deliverables tied to your long-term strategic vision.
- Innovation: You need a base “innovation” team besides your primary pillars. This is where I’d place Chat-GPT today. You should experiment with generative AI and LLMs to make your business more productive. Microsoft already has enterprise trials of Microsoft 365 Copilot, their OpenAI-powered productivity tool. A small agile team or partnering with a university can drive the innovation pillar. University partnerships are a great way to get talent at a low cost and can establish a hiring pipeline. By separating innovation from action, you can have different metrics for measuring the success of your innovation work relative to your action pillar since innovation projects take higher risks. If all your innovation projects deliver, you aren’t reaching high enough.
- Communication and education: Communicate and engage stakeholders across the business. This means at all levels. Engage and educate your board and CEO staff. Invite employees to a company-wide hackathon. Don’t limit it to technology employees. You may find passionate supporters in other functions that can add value to your efforts.
This is a classic flow-down. AI is at the top, but it leads to pillars that flow down to specific projects with measurable milestones. At the bottom of the pyramid is a base of innovation. (Side note: All of these projects should be done using an agile methodology. This field changes too quickly to treat these projects with any other type of project management methodology.) So, build your pyramid today. Here’s a simple template based on the four steps above.
![AI action plan template](http://www.zscaler.com/cxorevolutionaries/sites/default/files/images/blogs/ZS-Diagram-AI_Innovation.png)
AI has been and will be a rapidly changing area with significant ramifications. I always tell people to think of AI today as “augmented intelligence” instead of “artificial intelligence” since the projects with the most significant impacts are where strong performers will leverage AI to give themselves a superpower, accelerating their deliverables. Chat-GPT is a great example. You don’t want to trust it to be 100% accurate with everything it generates, but if it can help a strong employee get a ten-page strawman done, then the employee can make corrections and edits instead of starting from scratch; it can be a superpower, allowing the employee to finish their task dramatically faster.
If it hasn’t already, generative AI will soon play a significant role in producing your next presentation or helping you analyze some data in your spreadsheet. Who wouldn’t like some of those hours back they spent building a pitch or crafting an Excel formula as they analyze a pile of data? There are excellent tools released seemingly daily. With Gamma.app, you can generate a presentation with a simple prompt, like “Why every IT leader should adopt a zero trust architecture.”
![Sample Gamma.app zero trust presentation](http://www.zscaler.com/cxorevolutionaries/sites/default/files/images/blogs/2023-06-29_10-59-39%20-%20AI%20-%20GS.jpg)
With Bing Image Creator and Midjourney, a short prompt can generate a photo-realistic image in seconds.
![AI-created image by Midjourney](http://www.zscaler.com/cxorevolutionaries/sites/default/files/images/blogs/AI%20augmentation%20-%20photo-realistic%20image%20-%20Simpson.jpg)
These are just a few examples of the explosion of AI services (many of which are free) here today. Use them for “augmented” intelligence.
AI is a lot like the cloud. Many people will overuse the term, but a few will charge ahead and implement real AI-based solutions where it makes sense and stay on top of this wave as it matures. Others will look up one day and say, "Do we need an AI strategy?"
Don’t ask this question two years from now. Ask it now.
(And yes, the first image of this blog was created using AI. Specifically, Microsoft Designer)
What to read next
How personalized ChatGPT will give superpowers to every CIO (video demo)
Leveraging large language models (LLMs) for corporate security and privacy