AI is changing everything, right? For most of us, that change has been slower than anticipated. For some, glacial. The promise of AI is enormous, but delivering on that promise is hard, human work. 

It may sound counterintuitive, but the best way to face a complex challenge like incorporating AI into your operating model is to go back to basics and ask the simple questions of why, what, and when.

The hype, the hope, and the why

In April of 2023, OpenAI made its GPT-3.5 model publicly available, and the world changed overnight, right? Venture capitalists and tech juggernauts started investing billions; the hype cycle was cycling, and the brave new world was upon us. Or was it? 

At Slalom, we’ve spent years working with hundreds of organizations that all want to get the most out of AI. Though there are steps they all can take—from fostering adaptability and continuous learning in their workforces to ensuring leadership is also best positioned for the changes ahead—there is no simple panacea for everyone when it comes to AI. 

Organizations can therefore be forgiven for not knowing where to go with AI—who knows what the future holds? Why do anything, especially now?

In our experience, our clients’ catalyst to act now and (re)start their AI journey typically comes down to four themes that may sound familiar to you:

1. FOMO (fear of missing out): Competitors are already using AI to move faster, reduce costs, and deliver better client experiences. If you’re not using AI, you’ll fall behind, right?

2. More glow-up, less grow-up: Even in periods of sustained growth, companies are pressured to scale revenue while keeping hiring to a minimum. AI turns scaling challenges into opportunities, helping organizations achieve more without growing their footprint. Can your team glow big while growing only a little?

3. Lean in … to AI: In the opening keynote of the 2024 Microsoft Ignite conference, Microsoft CEO, Satya Nadella stated, “What Lean did for manufacturing, AI will do for knowledge workers.” In Lean, value means doing the things that customers will pay for, while waste is doing anything else. In our professional lives, we are often required to do things that are actually waste. Can we give these necessary but non-value-adding activities to the machines?

4. AI hangovers: AI has been the shiny object in the room for years, but many organizations have overindulged—too much, too quickly. The result? Underwhelming deployments that delivered little beyond inflated expectations. The lesson?

    

AI licenses do not an adaptive organization make.

In fact, they can often make things worse. The cure for an AI hangover? Detox: Stop binging, reset based on your real business goals, and scale from there.

Building an adaptive organization

When we talk about adaptive organizations and AI-enabled operating models, what it really comes down to is gaining the flexibility your organization needs to succeed with AI through the right combinations of humans, processes, and machines. Gone are the days of the “one step at a time” approach where you’d 1) develop a strategy, 2) design the operating model, and 3) deploy and manage change. These are now perpetually concurrent activities in an adaptive organization, and they take thoughtful planning, iterative experimentation, and a willingness to learn as you go. Perhaps the most daunting part: your job is never done. 

In our experience, however, we’ve found the following principles can help:

• Define value, and prioritize: “So what?” should be the first thought when testing AI. Align your efforts with areas where AI can deliver the most measurable impact—whether it’s boosting revenue, cutting costs, reducing risk, or improving operations.

• Start small, think big: Focus on small, practical wins that demonstrate value quickly, but always keep an eye on the bigger transformation. What’s one simple but ubiquitous process you could improve with AI right now?

• Test and learn: AI works best when it’s treated as an evolving tool, not a one-and-done solution. Develop a process for testing ideas, scaling the ones that work, and throwing the ones that don’t in the recycling bin (you might use them later).

Get specific on “value”

Just to be clear, this is not what we mean by being specific when defining the value you want to achieve from your AI program:

Specific is simple and to the point: “We value activities that …”

1. Grow revenue

2. Reduce costs

3. Mitigate risks

That’s it. If an AI use case doesn’t help do one or more of those things or isn’t a precursor to them, you need to ask yourself if you should pursue it. Granted, grace should be shown during experimentation as you fill your funnel of AI use case opportunities, but be judicious about what use cases take your finite resources. Say NO to low- or no-value things so you can say YES to the right ones. 

The essence of strategy is choosing what not to do.

Michael Porter
Harvard Business School

Source: What is Strategy, Harvard Business Review

Align AI efforts to strategy; what is your organization doing to realize its why? How does a given AI use case support that what? For an organization’s AI decision-makers, it’s imperative that they’re able to articulate what their strategy is and connect AI to value-adding processes. 

Scaling AI the right way

Deploying AI to build an adaptive, AI-enabled operating model is not a decision of “will I?” or “won’t I?” You will, because you have to. How your humans, processes, and machines come together in an effective manner is the biggest question you will face. There are many ways to be effective at these intersections; think of your AI deployment options as a spectrum:

• Assist: AI helps the human with simple tasks.

• Augment: AI enhances the human’s decision-making.

• Anticipate: AI predicts trends and risks and makes humans aware.

• Associate: AI provides recommendations for complex problems.

• Automate: AI takes over repetitive, low- and no-value activities.

• Agentize: AI operates independently of the human in specialized roles.

This spectrum isn’t always a maturity curve, but it can help frame your thinking early in your AI journey. Remember Start small, think big? Start with finding ways to assist and augment humans and their processes. Then, consider what it might mean to automate those processes. Posit what it might mean to agentize parts of humans’ work. Value correlates with how far toward the right on the spectrum you go, but so do cost and risk, and aren’t we trying to reduce them? Test and learn.

The hype will be realized: Now what?

There’s no escaping the fact that building an adaptive, AI-enabled organization is hard, messy, and requires leaders who are willing to experiment, learn from failure, and adapt as they go. History shows us that transformative technologies tend to realize their promise, often in ways and at a pace that are impossible to predict. AI is no different. The question isn’t if it will reshape the way we work, but how and when. While we can’t foresee every turn, one thing is certain: the organizations that take thoughtful, deliberate steps now will be the ones poised to lead the way.

At Slalom, we’ve been in the trenches with our clients, helping them navigate this journey. We’ve seen what works, what doesn’t, and what it takes to get it right. So, what’s your next step?