Change Takes More Than a Megaphone: Communicate, Experiment and Educate to Drive Transformation

AI is the tip of the iceberg —  what we read in the news, see on billboards and hear in the boardroom and from employees. But that’s only the part that’s above the surface. The real challenge is navigating what lies below. 

According to BCG, top-performing organizations recognize the iceberg. They follow the 10-20-70 principle, dedicating 10% of their efforts to algorithms; 20% to data and technology; and 70% to people, processes and cultural transformation. In a recent webinar, Delivering on AI ROI: Ensure Your Story Is Fact, Not Fiction, Snowflake and BCG leaders discussed that challenge. 

Transformation — of people, processes and culture — doesn’t happen overnight. And it won’t happen with just a vocal leader, no matter how big the megaphone. In his recent book Change: How to Make Big Things Happen, Damon Centola, a sociology professor at the University of Pennsylvania, argues that change does not spread like ideas do. Big changes or transformations are not viral. Instead, they grow through a snowball effect, building through repeated exposure and strong ties. To drive change, it’s not enough to tell people about something; rather, you must show them that others who are important to them are doing it — and, more importantly, are deriving value from it. Whether it’s adopting new agricultural methods, reducing energy consumption, using data to inform decisions or relying on AI to generate code or emails, social influence matters. 

When change represents new behaviors rather than just a new way of thinking, it’s not enough to have a leader with a megaphone or a few influencers promoting the change. Real buy-in and behavioral change happens when people see others doing it and hear about real results. Better yet, education and participation give them a helping hand and stake in the game. 

Successful organizations bring all these forces together to create a fear of missing out (FOMO), a proven motivator and change agent. In a recent discussion, a data leader at a large US food distributor described how the company drove excitement for its AI program: 

1. Catalyze change: The food distributor  designated an AI catalyst within each function or line of business in the company. The catalysts promote AI within the business context and act as a sounding board for new use cases. The catalyst community offers a safe place to present ideas, encouraging ideation far and wide within the organization. Casting that wide net is important.

2. Foster FOMO: The organization offered workshops, supported proofs of concept and promoted successes. Then, the data leader doubled down on recognition. As they said, "I celebrate the !@#$ out of the project leaders and shine the light on them." At that point, others raised their hands and asked when they would get a workshop. 

3. Create buzz: And last but not least, they promoted her team by creating a buzz. "If we were in office we’d have mugs, but now we have a logo and Teams backgrounds to promote the team." Go team!

Across Snowflake customers, true changemakers don’t reach for the megaphone. They launch programs to communicate, experiment and educate, to transform their people and processes. 

Communicate effectively

While water-cooler conversations might spark serendipitous collaboration, AI evangelism must be more systematic and cross-company. How can you best leverage the network effect across your organization to not only spread the word but also drive the change? Some companies designate “catalysts”; others have created a “translator” role.  

At Toyota Motors Europe, translators act as a bridge between the business and the data scientists. They remove jargon (on both sides) to make concepts easily approachable and understandable. Similarly, Kmart Australia introduced a data translator role a few years back and assigned one to each operational area. The result was a 400% growth in new ideas within three months but also a 3x increase in benefit per data use case. These translators, embedded in the business units, could roll up their sleeves and work side by side with teams to implement outcomes. There is more to communication than just words. 

And as they say, a picture paints a thousand words. At Toyota Motors Europe, the data team created a visual representation of their data mesh to illustrate data sources and uses and the various “stations” in between, such as data governance requirements and approval processes. The idea was to make it simple and visual. Maps are effective means of communication. 

AI and data leaders must take a step back and map out the audiences they want to reach, the messages for each and the mechanisms of communication. A good place to start is by asking these questions:

• Who are the target audiences?

• What do they need to know? Why are you talking to them?

• What is their level of understanding?

• What form should the content take?

• What channels will you use to reach them?

• When should the content be delivered? 

Experiment to ideate and educate  

As word spreads, the curious will want to experiment with AI. Experimentation explores the art of the possible and generates ideas for ongoing projects. However, the value of experimentation isn’t only in the models built but in the experience itself. In the long run, will most companies build their own AI models? Likely not. Experimentation encourages hands-on learning. Toyota Motors Europe hosts hackathons regularly with representation from manufacturing, logistics, R&D and its geographical markets. This bottom-up approach allows teams to test ideas and learn about the data and technology available to build them. 

Educate broadly

Not all employees can participate in a hackathon. In fact, most employees are neither data nor AI experts — at least not yet. Successful change management requires comprehensive education to uplevel the organization; employees need it and want it. Many say that if they don’t get it, they will walk. According to a recent talent management study, 74% of millennial and Gen Z employees say they are likely to quit their jobs within the next year due to a lack of skills development opportunities. Fortunately, according to a LinkedIn study, organizations are concerned about employee retention. Providing learning opportunities was survey respondents’ no. 1 retention strategy. As AI penetrates enterprises, change management initiatives must educate employees across all roles and levels — “from shop floor to top floor.” 

At Toyota, training not only is just about young graduates but also extends to decision-makers at all levels including executives at the VP and EVP levels. AI workshops and hands-on labs start with questions such as “What is data?,” “What is AI?,” and most importantly, “What can I do with AI?” Real “hands on keyboard” experiences demonstrate the art of the possible, and that in turn brings more use cases for experimentation. 

At Alberta Health, their AI scribe captures information from patient engagements in the emergency room, freeing up doctors to deliver care and improve human interactions. It’s being used by a handful of emergency department physicians, who are reporting a 10%- 15% increase in the number of patients seen per hour. The idea for the scribe came from an emergency department physician, showing that it’s not the data team that identifies the use case. When you enable employees at scale, new ideas are born. And, where there is potential for concern about sensitive data and outcomes, more watchful eyes are available to keep their new AI colleagues in line. 

With the proliferation of AI agents performing a multitude of tasks, employees need to know what to expect. They must understand their own roles in defining, collaborating with and monitoring the output of their new “colleagues.” 

Adopt a literacy framework

In my research at Forrester, I developed a framework for thinking about audiences that starts by considering what you’re trying to achieve. The curriculum of ACES was designed to promote Awareness, Comprehension (or just better understanding) and Expertise and to Scale them across an organization with a feedback loop from the data (and now AI) experts to others. 

The idea is to not only focus on the experts but extend education to the whole organization. After all, everyone has a role to play in either capturing, protecting or using data. In fact, those who capture the data are often overlooked: the cashier, the field service technician or the user of the microwave in the breakroom, for example. And, in our new agentic world, we humans must understand our agentic colleagues.