Do you need an AI strategy?

Did you catch the Moves in the Cloud blurb in my last newsletter where I applauded global-telco Orange on its large-scale test of new generational artificial intelligence (GenAI) capabilities? Orange is part of an early access program (EAP) for Microsoft 365 Copilot, which is designed to leverage the Microsoft 365 suite. Orange plans to roll out at scale across its internal organization, and will have just about every function using it: sales, marketing, human resources, IT, consulting, operations, assistants, project managers, legal, finance, “and more,” according to the blog post. What an ambitious project!

But analyst Dean Bubley, @disruptivedean, suggested that companies don’t really need an overarching AI strategy, with this posting:

I disagree with Dean—a lot. GenAI is SO new, and it’s changing SO fast, that I believe every enterprise needs an AI strategy. Why? Because this shit is hard to do, dude. Remember, us “commoners” have been using GenAI on the regular for only about a year. Chances are, very few people inside your org are skilled at using it, especially the non-techie folks. So don’t expect to toss a bunch of AI tools at your organization and watch people to just “get it.” Why is that? Because the inertia of your organization is strong, and to change work habits, you’re going to have to drive AI through the organization.

Your strategy to drive widespread internal adoption doesn’t have to be complex, but you do need a plan. My first suggestion: treat this as a quick, iterations-type project, not a big-bang approach that takes years to implement. But get going: the sooner, the better. One of my favorite starter blogs is this one from Bain, which offers a cheat sheet for getting underway. Once you do that, follow this four-step plan to roll out GenAI in your organization.

Step 1.  Where are you going to use GenAI?

Step 2. What tools are you going to use?

Step 3. How will you train people?

Step 4. How will you measure success?

1. Where are you going to use GenAI?

Short answer: In as many places as you can! GenAI can be applied in almost every area of an organization. Take another look at the list above from Orange and think about how workers across your organization could speed the creation and review of reports, spreadsheets, presentations, marketing copy, etc. Review this blog from Microsoft Azure about how our industry can use AI. To Dean Bubley’s point, AI is eventually going to be everywhere. Deciding where to begin is its own challenge.

Lots of telcos are trying it in customer support organizations, where they already have a ton of content and data to work with, as well as a mechanism for measuring success. (More on that later.) Start with your internal help desk, or add a chatbot to your external chats. Areas of low-hanging fruit for GenAI are call transcription, summarization, and categorization, as well as drafting answers to customer queries based on existing information. While it will marginally improve your top worker’s productivity, it will dramatically improve the quality of work of average to below-average workers. Imagine turning your whole workforce into a full set of powerhouses? Whoa.

2. What tools are you going to use?

It’s overwhelming to keep up with all the creation that’s happening with AI tools. Get a handle on it by assigning one or more AI Champions in your org who are aligned with the vision of using the tech in as many places as possible. Their first task: research all the AI tools out there and create a second brain, which is a document where you keep all your AI tips and tricks about a particular topic. In this case, it’ll include all the GenAI tools and ideas your Champion finds on LinkedIn, X, blogs, etc., a record of which ones they’ve tried, and the result.

Your Champion should spend 30-60 minutes each day researching all the changes that have happened. (Here’s a tip: have them follow Rowan Cheung and subscribe to The Rundown AI for daily updates. He also puts out a super informative email newsletter each business day.) So much is happening so quickly that your Champion will need to scour the sites on a daily basis to see what new things have come out.

Once your Champion has a shortlist of ideas to try, sign up and take them for a spin. Almost all will have a free trial or a low introductory cost; buy a few seats and experiment rapidly. See which ones produce the best quality and turn out the best results. Once you find the winners, buy the tool for the entire department.

3. How will you train your people?

It’s hard to get people to stop working the way they’ve worked their entire lives, which is doing everything manually. So, channel Jim Abolt! Jim is a leadership guru (and my former colleague and pal), and the smartest guy I know when it comes to getting people to embrace change. He shared his “framework for change” with me more than 30 years ago. In short, it’s a five-step process where company leaders communicate the plan, create the desire among workers, model the new behavior, change the work, and make it stick. That means your C-suite, your VPs, and all of your directors become cheerleaders, telling everyone that you are going ALL-IN on AI. To get your org to change its inertia everyone has to be singing from the same song sheet, saying, “We’re changing the way we work. We’re investing in new initiatives. We’re going to be AI-first.” Then, your leaders have to be the pioneers, forging ahead to learn and demonstrate the new way to operate, and then require the same of everyone else. If they start using it, then everyone will start using it. In that way, the new way becomes the only way.

As a part of that effort, leadership will need to set business goals around how AI will affect the organization. As a part of that, you’ll need a goal for how many people are going to use AI. Then, offer free training. There are tons of options out there from Amazon Web Services, Google Cloud, and Microsoft Azure (which has different courses for developers and for business users). Then: get everyone to start using it. This is where the rubber meets the road. If you can measure usage, even better. Track who’s doing it, and who’s not. Share successes and reward the people who are embracing the new way. At TelcoDR, my goal is 100% of our organization using AI—all roles, all people.

4. How will you measure success?

With AI, the key performance indicator (KPI) will be productivity improvements. This is another reason customer support is an easy place to start. You are probably already set up to measure all kinds of metrics, and you can easily compare the before and after of your implementation. Here’s a list of things that will improve with the right application of GenAI:

• First response time
• Resolution rate
• Number of interactions per ticket
• Tickets solved per day, per agent

In my organization, we’ve set a goal that we want people using AI 50% of the time or more. Additionally, we’ve created second brains for all our company functions. We’re experimenting with tools, and keeping the ones that work. We’re training people. We require recruits to take an AI test, and we start training new hires on AI immediately once they start at our company.

And we’re measuring it. We’ve implemented GenAI in our customer-support systems for Skyvera and Totogi, and are now solving 42% of our tickets with AI, freeing humans to work on thornier issues that AI can’t solve. Our customers are happy to get answers faster. Our support teams are psyched to have time freed to work on more complex problems, giving up the repetitive mundane tasks for the AI bots to manage.

Want to talk more about it? Try the TelcoDR ChatBot on my home page! (Launch it with the little orange bot in the lower right of the screen.) We trained it on all my blogs, podcasts, and speeches from the last three years to give you an authentic chat-with-DR experience.

You could ask it if you need an AI strategy. But you already know the answer: the answer is YES, so get going.

Recent Posts

Get my FREE insider newsletter, delivered every two weeks, with curated content to help telco execs across the globe move to the public cloud.

"*" indicates required fields

Email*

Email

This field is for validation purposes and should be left unchanged.