Using Agile to Lead Artificial Intelligence Change

We are at the dawn of a new era in digital transformation – also called the 4^th industrial revolution. McKinsey Global Institute estimated that an additional $13 trillion could be added to global GDP by 2030 from today through digitization, automation, and artificial intelligence (AI), as these technologies continue to create major new opportunities and gains. The U.S. government is on the way to committing nearly $250 billion to promote emerging technologies including artificial intelligence, quantum computing, computer chips, and robotics.

Technology plays a role, but people and change leadership play a bigger role. An agile mindset and experimental approach lay a foundation for us to design our future in a way that technology works best for humanity (and not the other way around!). Like never before, AI machines now process information better than humans due to “big” data sets, open platforms, increased computational speed, and machine learning (ML). Permeating much of our daily lives, AI lurks in the background every time we browse our Facebook newsfeed, search on Google, get a song recommendation from Spotify or receive a same day delivery from Amazon.

AI is rapidly growing from machines doing specific tasks intelligently like self-driving cars, detecting fraud, or understanding human text and speech. Companies have much to gain by harnessing AI’s strategic advantage, yet are employees and leaders across the enterprise adopting digital/AI practices intelligently?

The 4^th industrial revolution is about aligning AI with core areas of the business; but it’s also about embracing important cultural and organizational shifts. And the odds are stacked! AI has a long development cycle and high failure rate- less than 2% of proof of concepts making their way to fruition. Of those that succeed, the long-term adoption of AI post-implementation is 60% at best.

With that in mind, let’s explore practical ways to leverage agile in the “people/employee side” of AI transformation, based on our first-hand experience with large organizations in both the private and public sector. Here are seven agile practices to effectively lead AI change for your organization:

Meet People Where They Are

In order to effectively build a bridge from current to future state, one must first assess where organizations and individuals are in their data science journey.

Some organizations are just starting to dabble with the possibility of AI. They may just want to “check the box” to comply with regulation or to get good PR press. Maybe they are starting to see the benefits of automation in some functions, but still rely heavily on external expertise.

On the other end of the spectrum, organizations like Google, Spotify and Amazon were born digital with AI in their DNA. With fully integrated data pipelines fueling their AI factory, business decisions are anchored in data and smart systems.

From an individual perspective, we can assess whether a person understands basic concepts in modern technologies, has the needed technical skills, and the capacity for change and adaptability. Wherever a company or individual is today is your only starting point for progression. Expecting behavioral change in rapid pace or all at once is simply unrealistic. Therefore, our approach should meet people where they are today and allow time for steady, gradual progress towards the desired future potential.

Develop Data Literacy

Often overlooked is the need to establish a common understanding of advanced technologies (e.g., AI, Augmented Reality (AR), Deep Learning (DL)) so that employees can exchange ideas and collaborate in the same language.

Digital fluency is needed, for example, around how algorithms use data inputs, a model and outputs to inform insights, decisions, and actions. Regardless of a person’s role and technical expertise, everyone can think like a data scientist by asking key questions: What and why data matters? What story does data tell us? How can everyday work and decisions be enhanced?

Additionally, it’s helpful to establish a shared understanding of key concepts such as computer vision, machine learning, supervised learning, unsupervised learning, features, data bias and privacy to name a few. Data literacy opens the realm for meaningful conversation and exploration of what’s possible.

Think Big and Small

Data science and AI yields fruitful results only when aligned to the strategy and mission of the organization. So, for example, if your mission is to improve human health, then you might apply Natural Language Processing (NLP) to find patterns in words and accelerate data-driven research in biomedical informatics. This is an example of a “use case” or scenario in which the AI solution would be valuable to a user.

Internalizing your organization’s north star vision and strategic intent (“thinking big”) enables the identification of practical use cases (“thinking small”). Use cases are then assessed for technical feasibility and business justification. Those with viability are turned into prototypes and lab experiments for further exploration and iterative development.

Tightly scoped projects where data is abundant and processes are manual are good candidates for starting small. Thinking with both breadth and depth increases your chance of success in the AI field where a low percentage of ideas actually make their way to production.

Build a Culture of Experimental Learning

Building a culture of experimentation and learning allows for “freedom within the framework” where people can safely test, learn, and adapt incrementally over time. Traditional business processes adhere to a pre-established set of goals, outcomes, and conditions. With experimentation, you have far more space and flexibility to shape your work in a more fluid and responsive way.

“Quick wins” gained through incremental steps/phases or rapid prototyping boosts stakeholder engagement and confidence and solve for business/project problems to determine next steps. When we create room for failure, we can learn in a way that success cannot teach.

The science of change facilitates moving beyond current skills and knowledge levels (agile style) by giving learners small chunks of information to digest, along with a tool or structure to apply learning. Bolstered by technology, contemporary “2.0 learning” integrates learning, collaboration, and knowledge management. Add immediate application to everyday work and you have a recipe for making the change stick. Just like the algorithms that are constantly running A/B tests and adjusting on the fly, we can regularly experiment with and fine tune behavioral change.

Empathetic Stakeholder Engagement

As companies introduce software bots and digital self-service, and as they transform assembly lines, we can help bring along impacted employees, leaders, and customers. In my work with clients, I hear resistance in comments like “I shut down as soon as you say Data Science” or “I don’t see how this relates to my job” or “I’m afraid of automation taking over my job”. There is one common denominator… Fear.

AI and the perceived threat of automation is not something to fear, but rather an opportunity to embrace. Turn fear into productive activity by empowering stakeholders early and often in the change process. Afterall, those closest to the problem (e.g., touching the system, interacting with the customer) often hold the key to unlock the smartest solutions.

Acknowledgement of people’s emotions and involving them in the discovery process will inherently get them bought in and actively engaged in change adoption. To inform your next steps, you could conduct surveys, interviews, and focus groups on a regular basis to hear your most impacted stakeholders’ thoughts and ideas. We must invite stakeholders across the board to imagine and reinvent for digital innovation.

Upskill and Augment Teams

According to McKinsey Global Institute, up to 375 million workers worldwide may need to switch occupations or learn new skills. We need the workforce – both technical and non-technical – to assist in the transition and to shape the AI future of work.

As deep learning and smart machines increase in application, it is more likely that current teams will be augmented rather than jobs eliminated. Imagine humans working alongside machines in a synergistic relationship. Humans bring creativity, problem-solving, and intuition; machines bring computing horsepower and safety to otherwise mundane, routine, and dangerous tasks.

Organizations can offer an array of learning and development opportunities tailored for individual skill gap development and different adult learning styles. These might include:

• a course catalog of mixed, self-paced, micro-learning, in-person and/or virtual courses and classes
• intensive boot camps for advanced technique and application
• a mentoring program that guides employees through a supported cohort experience
• workshops, workspaces and learning labs to reinforce social learning
• enterprise social networks, wikis, webcasts, webinars, and podcasts

Agile Leadership

Multiple research findings point to leadership as being the top determinant of organizational change success. Agile leadership is the ability to pivot quickly and easily while maintaining a constant, rapid motion. Leaders are instrumental in modeling this new behavior, building trust, accepting failure, and empowering growth potential in their people. Enlightened leaders have moved away from “predict-and-control” style towards a “sense-and-respond” approach. Cultivating agile qualities supports responsible development of AI solutions.

In recent times, we’ve seen democracies be demolished, increased depression and suicide, malicious applications, and unintended consequences resulting from machine-driven technologies (like social media). Left unchecked, algorithms can exacerbate societal and racial inequalities without humans even realizing it. Deep learning presents a “black box” solution difficult to understand, interpret and explain in an acceptable way. AI demands a new and responsible form of change leadership – with agility, emotional intelligence, open-source governance, and a focus on truth, equality, fairness, and morality.

Just like an algorithm, these agile practices are inputs that can be translated into meaningful outputs when mobilizing others towards an AI-powered future. Using agile principles, we can make a real difference in helping organizations to prepare for the tsunami of AI innovation ahead and reinvent their business models. Actually, it’s imperative that we take a lead role in the transition towards digital capability and responsible AI development. The quote “we design tech and tech, in turn, designs us” by Pamela Pavliscak shows the potential for AI technology to take a turn for the worst. However, we are at an inflection point where we can thoughtfully design and shape AI across government and business to create a higher quality of life. This requires all of us to dream big and act boldly; a “moonshot” sort of thinking to reinvent products, services, and organizations.

KEY POINTS

1. Meet People Where They Are
2. Develop Data Literacy
3. Think Big and Small
4. Build a Culture of Experimental Learning
5. Empathetic Stakeholder Engagement
6. Upskill and Augment Teams
7. Agile Leadership

Written by FMP Partners

Within the private and public sectors, Nicole Sroka leads wide-scale people adoption in the areas of organizational development, process, technology, culture, and environmental sustainability. Nicole is a creative mobilizer ready to connect, explore and solve for the complex challenges of our time. Nicole has partnered with FMP since 2019.

Dr. James G. Shanahan has spent the past 30 years developing and researching cutting-edge artificial intelligent systems splitting his time between industry and academia. He teaches at University of California at Berkeley, and has (co) founded several companies that leverage AI/machine learning/deep learning/computer vision.