It is understandable that people are generally concerned about automation replacing jobs, but gaining buy-in for the integration of AI is not just about alleviating fears. Instead, the job of leadership is to build confidence that this tech can help us meet our goals faster and with a larger upside. If companies are to successfully navigate this transition, people need to know that leadership is incorporating AI in a way that improves and maximizes their contributions and optimizes company growth, ultimately leading to increased experience and compensation for the team.

We have a curated list of projects for the next year and people have already dedicated many months of their time to prioritize them. So, I am very transparent about what we can gain from AI and machine learning, which is building upon our existing roadmap with the end result of a more robust and powerful suite of products. This will not happen all at once. A staged approach to integrating AI into workflows and products emphasizes strategic resource allocation and a commitment to continuous learning. Leaders are on a journey with AI as much as anyone else, so it starts at the top.

Related: Building an AI-Augmented Workforce While Remaining Human-Centric

Integration: A staged approach

The first goal of integrating AI should be understanding the quickest way for it to start having a positive monetary benefit. While our AI project is still a work in progress, we are expecting to increase revenue anywhere from $2 million to $20 million as a result of a first round of investment of under $100,000. But to achieve that type of result, leaders need to get comfortable with AI and figure out the challenges and complexities they might encounter.

I am hiring outsourced firms to build the smallest prototype possible so I can be part of an ongoing conversation about the most feasible way to integrate AI. It was similar to when I would meet with my CTO every morning to talk about building the main product of our business today. We would combine his knowledge of programming and my knowledge of our business goals to make daily improvements and get a minimal viable product out to the market for testing as quickly as possible. Getting real-world feedback and seeing real work performance is when you truly start to improve and refine your product.

My initial goal with testing AI is to increase our push messaging system. We send millions of messages daily with just one person handling the creation, scheduling and optimizations of hundreds of different messages. With AI, we can customize them to a previously unthinkable degree, leading to meaningful performance optimizations that AI can handle automatically. We also have other products that should have a machine-learning component at their center. By exposing my product and dev teams, as well as myself, to the potential of AI, we won’t waste resources without seeing a solid increase in performance. And for all that to work, communication must be consistent throughout the process.

Related: 3 Practical Ways to Let Artificial Intelligence Work for You

Communicating our direction

I am a huge proponent of transparency in all facets of business. I want employees to know how any changes we consider will impact them. But if I am not articulating my goals clearly, misconceptions can take root. It’s important for leaders to express certainty about their direction with AI integration and provide a clear vision for where their investments are leading the company.

As a company scales and more people’s ideas are competing for limited resources, leaders need to be as accurate as possible in how they prioritize. For example, spending $1 to make $14 should be prioritized over projects that cost $1 for a $3 return. That may seem obvious but a threefold return on investment is persuasive. Prioritizing the project with the better yield first will then fund building many more of the product with the smaller return.

When people see that we are risking the least money to get positive results the quickest, we are more likely to get buy-in from everybody. Start small and build from there. If we can increase our marketing message response rates by 2%, then we will have more confidence applying the prototype to our external products. It is also an opportunity to reinforce our messaging.

Paying attention to red flags

As leaders socialize ideas for AI integration across a wide range of people, they may come across major red flags of what you are missing. Everybody has different perspectives. If you are a glass-half-full kind of person, listening to the glass-half-empty kind of person offers a complementary point of view.

Whenever I have ideas to really move the numbers, I tend to act fast. It is crucial that people understand that I am not fast-tracking AI integration because I am unhappy with our current process or people. It is because I am happy that I will not risk what we already have unless I am fully sold on the range of the upside — and I want to expedite the learning process to get to those benefits faster.

I still want to talk to as many people as I can — employees, developers, marketing folks, product managers, external investors — both for the tone of responses and any major issues. Those red flags may be great things to consider or I need to give people more information. Either way, my response can alleviate their concerns.

Related: I Transformed My Company Through AI 15 Years Ago — Here’s What I’ve Learned

AI and human: Complementary growth

I believe people should embrace AI in much the same way they embraced computers 20 years ago. Computers opened the door for more new jobs and greater diversity than they replaced, so leaders need to emphasize that AI represents a major opportunity for employees with a growth mindset — while balancing efficiency and capability gains with the human elements of creativity, empathy and ethical judgment. Think about ways to engage curious employees through workshops, courses, webinars and industry conferences. With transparent communication and a commitment to continuous learning, people will come to see AI as a complement to our human endeavors.