Implementing an Artificial Intelligence (AI) system can have a substantial impact on your workforce. This impact can be positive or damaging, depending on how you prepare them. When beginning the AI implementation process, you’ll want to keep in mind that not everyone will be initially excited about software assuming traditionally human roles. What’s more, stakeholders may be skeptical about spending so much time and energy on yet another software implementation if they’re not provided strategic direction.

In this post, we’ll walk you through the best way to prepare your organization for an AI system implementation — from pre-implementation knowledge-sharing, to post-implementation phase-in. At every stage of the implementation process, it’s important to remind your employees that AI is there to serve them, not replace them, and that by buying into the new solution, they’ll be helping to create a more enjoyable work environment.

Pre-Implementation Preparation

The key statistic that employees will need to hear before you unveil your plans for AI is the following: According to Gartner Research, AI will create 2.3 million jobs, while eliminating only 1.8 million jobs by 2020, and by 2025 AI will have resulted in 2 million net-new jobs. Your employees will want to know immediately that your company is dedicated to reskilling them to handle AI-related tasks, as opposed to turning AI into a tool solely for reducing headcount. By pounding this theme home, your employees can breathe a sigh of relief, and ultimately keep an open mind about the power of AI.

The same issues will arise with stakeholders, whether they be managers, line-of-business executives, financial officers, or investors. AI implementations — including those for digital colleagues like Amelia — take time, and they require financial and labor output unlike most other software implementations. Let them know ahead of time that implementation may take up to three months, that you’ll need workers who are solely focused on the project, and that you may not realize Return on Investment (ROI) until several months or even a year after the project goes live. Let them know that you’ll start with a small implementation as a test to see if the solution can actually yield results.

Post-Implementation Team-Building

As we’ve mentioned in previous posts, when the AI system is ready to begin interacting with humans, let a group of super users be the first employees to interact and engage with the technology. This should be a small collection of passionate and imaginative workers who are willing to withhold their prejudices. This group can watch your AI project evolve, learn and develop human characteristics over time, and will quickly be impressed by AI’s capabilities. Super users will realize that thanks to your digital colleague, their work will be less repetitive and they’ll be armed with powerful information at all times. They’ll tell their co-workers about the tasks and duties that have been taken off their plates. They’ll recount just how easy it is for them to find information through cognitive AI. They’ll say the technology is making their job easier, not eliminating or diminishing it.

The employees who haven’t had a chance to work with the tool will actually request access.

Once you’re done testing with super users, invite a larger group of somewhat skeptical employees to test the solution. They’ll work with the system and realize that it’s not a job replacement tool. They’ll share their experiences with the most skeptical users in your organization. Word will spread that a new solution is being developed that isn’t a job killer and actually makes work easier. The employees who haven’t had a chance to work with the tool will actually request access. Now you’re ready for a full-scale launch.

To keep your stakeholders happy, let them know that the post-launch experience is the time to focus on ROI. In the past, we’ve written about when and why to expect ROI. Once the system is up and running (beyond testing and piloting stages), you should expect to begin saving money immediately by having the system handle simple and repetitive processes that were once handled by humans. These automated processes are massive time-savers that allow your business to allocate human labor to more complex and creative work. Hopefully, when human employees focus on other tasks, they can figure out new ways to generate revenue or make the company more innovative and competitive. Employees can spend more time with customers, or move to sales-oriented roles, or dream up new products and services simply because of the time gained from AI. Tell your stakeholders to focus on these initial victories, and ask them to give you several quarters to prove consistent  ROI.

Scaling for Continued ROI

The first two years of AI adoption are generally focused on experimenting with use cases and proving that AI can provide massive value for your organization. The second phase of stakeholder preparation is to focus on assigning the AI system to areas where it can generate money, or further prevent your business from having to spend it. This will no doubt please your stakeholders.

For example: In fraud prevention, AI can thwart phishing and denial-of-service attacks without manual intervention. These types of attacks could cost your company millions. In the case of direct selling, AI can capture valuable data from customer interactions, especially those where customers have expressed frustration or satisfaction with your products or services. Those would be ideal moments for AI to make product recommendations for upselling and cross-selling.

If you can prove these use cases valuable, you’ll want to start preparing your stakeholders for an end-to-end AI deployment. This means integrating AI with every front and back-end systems so that it learns everything it needs to know about your processes, your customers, and whatever it may need to collaborate with customers and employees. Once everything is finalized, you’ll have a digital colleague who can welcome customers to your website, guide them through the correct automated interactions, or escalate issues to the correct human agents. On the back-end, you’ll have a digital colleague who can retrieve data for employees at a faster rate that someone searching through Excel docs. You’ll have a digital colleague who can pull up historical customer data to bring context to human-to-human service calls. You’ll also have a digital colleague who is available 24/7, 365, on any channel, ready to assist customers for almost any issue. And which one of your company stakeholders wouldn’t like that?