Strategies for successfully integrating an AI practice into your org

Welcome to Part 6 of our comprehensive 6-part series, Beyond Buzzwords: Finding your Purpose for AI , a collaboration by Noelle Saldana with Martina Hodges-Schell, product innovation and transformation expert

Integrating an AI practice successfully into your organization is a journey and an investment. The integration is prone to failure as organizations do not hire realistic minimum viable teams and do not often set this new organization up for success. Instead, they often simply install one person, often in a vacuum, and hope for the magic to happen. Don’t let this be you!

Usually, the focus is on producing results. The inconvenient truth is that AI is a complex problem-solving, creative, collaborative endeavour, and results aren’t linear.

Problem-solving is not a linear, predictable process. Instead, it offers a proven approach to navigating uncertainty and exploring possibilities productively. This illustration was inspired by Damien Newman’s design squiggle.

As you move forward with making AI an integral part of your organization, it’s critical to consider both who will be involved (the team and culture) and how they will operate (the strategy). Let’s look at each in more detail. 

Team & Culture: How big a team do you need?

To determine a realistic team size, you need to consider the different capabilities and expertise necessary for your AI team and the surface area and complexity of the scope. 

Ask yourself, truthfully, what’s the readiness of your organization? What do you need to invest in to be able to leverage a high-performing AI team? Realistically, how much time and effort will that take? 

You’ll need a technical leader who can prioritize the AI work to be able to make progress, and can evaluate what they need to say no to instead. 

This leader needs to be empowered to set up this team for success: 

• What does a lean, dedicated, collaborative, yet critical mass team look like? What skills do you need?

• How does this team need to work to deliver best practice value? How does that best integrate with other teams and ways of working?

• What current behaviors and habits hinder the success of this team? What needs to change culturally across the organization?

Strategy: How this new team operates

The AI team will work at odds with other teams’ priorities if they are not properly integrated into the organization’s strategy, prioritizations, and decision-making. It is key to have a collaborative way of working cross-functionally in general, but also very specifically for AI to have the desired impact. 

Consider: How do you need to adapt your ways of working, decision-making, and measuring and rewarding progress now that you are invested in AI? This is not as simple as adding your new AI person to all the meetings, but fully integrating new and existing work streams together. This is a lot of strategic work that will look different for every company. 

What does it mean to be an integrated AI org? We’ll wrap up this series with a final checklist that will help you address the hard questions and get the most out of your AI investment. 

• Are all goals aligned to strategic priorities and all teams incentivised to succeed by working together? 

• How are you updating strategic priorities to integrate this new AI work?

• How does the AI practice represent itself and influence strategy and execution in the organization?

• How do funding, prioritization, and decision-making enable the AI practice? 

By the way, if this topic and way of framing is new, you might also want to read more about transforming your operating model.

What’s next?

Anytime you’re exploring new possibilities, it’s critical to keep in mind that success isn’t about predictable and immediate results. it’s not about setting a new team up in a silo to come up with a magical solution on their own—your goal is to integrate and support any new function.

If all this sounds like hard work, that’s because it is! But remember, the alternative is investing time and resources in projects or teams that have minimal chances of success. 

Finally, we’d like to remind you that you don’t have to do it alone. We help support teams and individuals on their transformation journeys, so don’t hesitate to get in touch (here or here) if you’d like to discuss how we might work together.

This is part 6 in the series. Catch up on previous posts:

• Part 1: Introduction

• Part 2: AI hype is new, our reaction to it is not

• Part 3: Is the product you’re building a good candidate for AI?

• Part 4: You are inspired to use AI: Now what?

• Part 5: Your AI Investment = (AI Needs) + (the necessary team)

• Part 6: Strategies for successfully integrating an AI practice into your org

Your AI investment = (AI needs) + (the necessary team)

Welcome to Part 5 of our comprehensive 6-part series, Beyond Buzzwords: Finding your Purpose for AI , a collaboration by Noelle Saldana with Martina Hodges-Schell, product innovation and transformation expert

AI is not one-size fits all. Consider this spectrum of organization types with varying levels of AI integration, which starts with the least (AI-boosted) and progresses to the most (AI-led). Each stage requires different levels of investment, both from a technical and a personnel standpoint.

The Necessary Team 

For many organizations, the biggest challenge is addressing the gap between where they currently are and what they are trying to achieve and hire for. In general, organizations vastly underestimate the size of the team they need. 

Becoming an AI-led organization involves commitment to significant organizational transformation. If your goal is to have an internal AI practice that drives innovation for all your products, it’s unrealistic to expect you can get there simply by hiring only one or two people or having another function cover this area of expertise. This is especially true if you’re trying to cut costs by hiring junior people who don’t have the experience of leading a team or a function. 

How big are your AI needs? 

An organization’s current AI capability is directly related to its data maturity. (We recommend checking out this classic by Monica Rogati: The AI Hierarchy of Needs.) 

If you have no or low data and do not currently use data to perform any analysis, AI does not fit into your core product. You could, however, use AI to boost efficiency internally and find value from using AI tools.  

When you have a data strategy, you’re well-positioned to leverage data to create products that have AI capabilities. The extent to which you can do this depends on your data organization.

AI explorers are investing in a seasoned leader to develop their data strategy and capability. There is existing data within the organization, but there is likely foundational work still to be done. AI-powered organizations have had time to build upon their data foundations and developed essential data analytics capabilities. These organizations are ready to take the next step to integrating (more) AI into their products. 

AI-led organizations have either built their foundations on AI strategy or they have invested significantly in transforming their organization to use AI to differentiate their business. 

We have put together this simple table to summarize our recommendations for developing a level of AI capability based on an organization’s current data maturity and AI goals. 

AI Investment Cheat Sheet

	AI Boosted	AI Explorer	AI Powered	AI Led
Needs	AI does not fit into the core product. AI could boost efficiency.	AI poses viable opportunities for the business downstream Start with a data-driven product that has a feedback loop	Clear and valuable applications for AI in the business that aren’t being used yet	AI differentiates the business
Data Maturity	no/low Operational data only. Not using for any analysis or trends. Not centralised	Developing - laying the frameworks Product instrumentation, more diverse sets of data, some analysis, some centralisation	Developed Has a data and analytics strategy, centralised data, analysis/DS work being done	Industry leading Strong data strategy, DS/ML production pipelines, etc.
Approach for AI	Use AI tools	Hire 1-2 seasoned people (pioneers)	Hire teams and build out capability	Transform organization to become AI-led. Think significant remodel, not a quick paint job.

When you evaluate this chart, consider if your AI needs align with where you currently are with your data maturity and what you are currently able to invest in for hiring. If not, it may be time to realign internal expectations or revisit budget conversations. 

And, of course, you need to make sure your organization is ready for the transformation that will need to take place. More on that in the next post.

This is part 5 in the series. Read on:

• Part 1: Introduction

• Part 2: AI hype is new, our reaction to it is not

• Part 3: Is the product you’re building a good candidate for AI?

• Part 4: You are inspired to use AI: Now what?

• Part 5: Your AI Investment = (AI Needs) + (the necessary team)

• Part 6: Strategies for successfully integrating an AI practice into your org

You are inspired to use AI: Now what?

Welcome to Part 4 of our comprehensive 6-part series, Beyond Buzzwords: Finding your Purpose for AI , a collaboration by Noelle Saldana with Martina Hodges-Schell, product innovation and transformation expert

A framework for curating meaningful applications of AI 

Groundbreaking inventions are famously never the first idea someone has had. Think of penicillin, sticky notes, or Coca-Cola. These products’ inventors were searching for something else when they stumbled upon paradigm-shifting inventions. 

One sometimes finds what one is not looking for. When I woke up just after dawn on September 28, 1928, I certainly didn't plan to revolutionize all medicine by discovering the world's first antibiotic, or bacteria killer. But I suppose that was exactly what I did. - Alexander Fleming Wikipedia

The point we’re trying to make is not that great discoveries only happen by accident, but that there’s magic (and great gains) that happen in the creative and experimentation process. People should not plow ahead with the first idea or solution or hack, as they are likely missing the bigger opportunity.  

A pattern that’s often seen in tech, and AI especially, is companies adopting product trends for the sake of parity with the market when it is ultimately an ill-fitting or unsustainable choice. How many products today have an AI assistant that doesn’t actually help their customers solve their problems? 

It’s most powerful to come up with your own ideas that are suited for your product and your unfair advantage.  When we’re dealing with a brand-new technology like AI, you can apply the design process of “go wide, then decide.” Zendesk’s triple diamond offers a simple visualisation of this approach. 

Continuous design process, based on Zendesk’s triple diamond illustration.

 Going Wide 

Going “wide” is a critical time to let our imaginations run free to enable innovation. We can bring in inspiration from unexpected sources and give ourselves the chance to come up with quirky and nonlinear ideas. Taking this approach is what will help us look beyond the obvious and find those ideas that have the potential to truly propel our organisation forward. This is where your team might start to hack together something to explore the possibilities of AI. 

“How might we?” 

How might AI add value and an unfair advantage to your business? Consider questions like: 

• What are your strategic goals: Where are you playing and how are you winning? Where do you want AI to help?

• What is your company’s expertise and unfair advantage? What are you known for? What do customers expect from you? What opportunities would AI open for you? 

• What are you solving for? Which problems or pain points are you addressing? How is AI enhancing that? 

…Then Decide

After you have generated ample possibilities, you have to be smart and deliberate about how you focus on which ideas are the most valuable and realistic to pursue.  

First and foremost: Is there AI that exists that solves, or partially solves, the problems you want to solve? Do you have the data to support the AI? If either does not exist yet, are you willing to invest in generating it yourself? 

Once you limit your ideas to those that are AI and data ready, we recommend prioritising the evaluation of solutions where you can quickly and cheaply learn if you’re solving a valuable problem. 

A valuable problem typically is: 

• problem/solution

• valuable and usable for your (intended) customers (product/market)

• can scale that into a sustainable business model

• can be profitable 

Consider questions like: 

• Feasibility: Which ideas that you’ve generated look the most interesting? Can you come up with lightweight experiments to quickly determine if they’re worth pursuing?

• Ethics: What are the ethical implications of the solutions you’re considering?

• Positioning: Would this direction make sense to your customers? Is it something they would trust you to handle?

At the end of this process, you should have a short list of ideas that you can move forward with. We understand this can be uncomfortable. We place so much emphasis on productivity and execution in the business world. It is easy to fall into the trap of productionising an half-baked idea. Someone will create a simple prototype on-demand and use that to quickly get buy-in, but later suffer the consequences when that does not make a good product. 

You need to have a mindset of learning and iterating in a range of short and longer-term bets. 

When you run your experiments and pilots, what do you discover? How can you apply that to the next round and make it better?

This is a perfect opportunity to take a page from the design studio playbook, where you create opportunities for exploration, learning, and play (along with rigor and pragmatism).

This is part 4 in the series. Read on:

• Part 1: Introduction

• Part 2: AI hype is new, our reaction to it is not

• Part 3: Is the product you’re building a good candidate for AI?

• Part 4: You are inspired to use AI: Now what?

• Part 5: Your AI Investment = (AI Needs) + (the necessary team)

• Part 6: Strategies for successfully integrating an AI practice into your org

Is the product you’re building a good candidate for AI?

Welcome to Part 3 of our comprehensive 6-part series, Beyond Buzzwords: Finding your Purpose for AI , a collaboration by Noelle Saldana with Martina Hodges-Schell, product innovation and transformation expert

Once you are clear in your purpose for AI, a key question to ask at the start of your AI journey is whether the product you’re building is even a good candidate for AI. 

We’ll share a tough truth here: AI can’t fix fundamental problems with your product. It’s not a magic wand that makes everything better. AI is a broad technical landscape, and its implementation being more complex than it may seem at first (more on this later.) This is why it’s so important to take an honest look at where your product is today and identify any areas where it’s lacking before you try to move forward with a new AI project.

To get started, you need to ask yourself:

1. What added value will AI capabilities bring to your product?

2. What pain could AI alleviate for your users? 

3. Does your product roadmap include any of the following product foundations: scaling, addressing technical debt, platform migrations 

4. How is data instrumented on your product today? What existing data-driven feedback loops do you have for the product? 

5. Who owns data/AI in your organization?

To state the obvious: if there is no defined value for AI or any ownership of it in your org, this is a nonstarter. If your product has no data instrumentation or is saddled with deep technical debt, your energy is better spent addressing those first. 

Additionally, AI should not be a solution in search of a problem. AI should help accelerate solving pre-existing user pain points/opportunities. You’ll want to invest some time doing discovery to identify the problem or opportunity you’re trying to address, how to best solve this problem, whether your solution is usable and valuable, and whether it’s possible to create a sustainable business model. In the next post, we’ll review how to align these problems with AI solutions, if appropriate.

When it comes to figuring out how to best solve the identified problem, we suggest using the guardrails of creativity, pragmatism, and rigor. 

Products that are too far to any edge will be problematic for different reasons. For example, when you apply rigor and creativity without pragmatism, this is when you end with CEO pet projects that flop (because there was no practical application in the first place). Or, if you apply rigor and pragmatism with no creativity, this leads to products like government websites that people use because they have to, but it’s an unequivocally unpleasant experience.

Products with the most value are in the sweet spot where all three overlap; the user experience is delightful, it serves a purpose that encourages user engagement, and is adequately vetted and responsible.

Finally, we’d be remiss not to point out that creativity is not a linear process. If you want to foster true innovation, you need to be prepared for the messiness that it can involve and the detours teams may need to take before they arrive at a solution. We’ll be sharing more about how to create the ideal conditions for creativity and innovation in the next section. 

This is part 3 in the series. Read on:

• Part 1: Introduction

• Part 2: AI hype is new, our reaction to it is not

• Part 3: Is the product you’re building a good candidate for AI?

• Part 4: You are inspired to use AI: Now what?

• Part 5: Your AI Investment = (AI Needs) + (the necessary team)

• Part 6: Strategies for successfully integrating an AI practice into your org

AI hype is new, our reaction to it is not

Welcome to Part 2 of our comprehensive 6-part series, Beyond Buzzwords: Finding your Purpose for AI , a collaboration by Noelle Saldana with Martina Hodges-Schell, product innovation and transformation expert

When trying to understand AI in a broader business context, we think it’s helpful to refer to Gartner’s Hype Cycle of technology adoption. This framework reminds us that people and organizations tend to respond in predictable ways to new technology. 

When it comes to AI, we think we’re here (where you can see the asterisk on the graph below)—the market is getting saturated, and the noise is becoming deafening, and we may be at the peak of inflated expectations. It means that we are quickly heading towards the trough of disillusionment; we’re about to learn the hard way that AI is not infallible and not a quick fix solution to everything. 

Illustration of Gartner’s Hype Cycle: *We’re currently here, near the peak of inflated expectations

The current hype around AI reminds us of a similar moment in the early 2000s. When Apple launched the App Store in 2008, it unleashed app mania. Every company at the time felt the need to develop an app, and at its peak, the Apple Store had 2.2 million apps. But not every one of those apps was vital or even moderately successful. Companies spent a lot of effort and money on developing apps that didn’t solve a problem, didn’t deliver sufficient customer value, or were a poor product market fit. So for every killer app or new market created, there were dozens of companies that invested in creating a lot of waste. 

Today, similar excitement surrounds the promise and potential of AI. We are in the middle of a paradigm shift, which creates incredible momentum for innovation. However, we also need to be realistic about what it means to incorporate that innovation into our work. We can’t simply “add AI to the menu” of our offerings without doing some real work to understand what that means.

In the App hype era, many companies had unrealistic goals that they could be innovative by creating mobile apps for their products. Companies rushed to be a player in the app space, but ultimately most were not successful in being disruptive to their industries, let alone their own businesses. Similarly, we see a rush to use AI as a means to stay relevant in the current hype AI cycle. To “use AI” encompasses both integrating the technology into a product as well as using AI tools for productivity. 

Here are some common things we hear (that you should be on alert for):

• Let’s sprinkle some AI on our product. Can we add a ChatGPT chatbot?

• I don’t need a designer for my product, I just need Midjourney. 

• I don’t need a data scientist. My developers can use AI tools. 

There is a real risk in only leveraging someone else’s technology, e.g. calling an API and depending on the capabilities and limitations of a different company altogether. What happens when your access changes or seizes? What happens when the AI goes unchecked? Building major components of your business on the back of another company leaves you vulnerable to many external factors like mergers, acquisitions, leadership changes, or drastic changes in direction.

If you want to shape your own AI destiny, you’ll need to bring a human into the loop. Designers, data scientists, and practitioners of any kind don’t just supply the working product—they bring their expertise and critical thinking abilities. They are the ones who will be able to tell you when the tool isn’t working the way it should. If you remove them from the equation, you’re setting yourself up for failure. 

We challenge you to be strategic and responsible about AI in order to avoid Gartner’s trough of disillusionment. 

In the next parts of this series, we’ll discuss how to find AI opportunities in your products, how to approach exploring applications for AI, determining how much AI is necessary for your intended goals and how to successfully integrate an AI practice into your product development team. 

This is part 2 in the series. Read on:

• Part 1: Introduction

• Part 2: AI hype is new, our reaction to it is not

• Part 3: Is the product you’re building a good candidate for AI?

• Part 4: You are inspired to use AI: Now what?

• Part 5: Your AI Investment = (AI Needs) + (the necessary team)

• Part 6: Strategies for successfully integrating an AI practice into your org
  

Beyond Buzzwords: Finding your purpose for AI

Welcome to Part 1 of our comprehensive 6-part series, Beyond Buzzwords: Finding your Purpose for AI , a collaboration by Noelle Saldana with Martina Hodges-Schell, product innovation and transformation expert

Over the past year, AI has become the go-to topic in any social or professional gathering. AI has captured our collective imagination, and as a leader, you and your board are most likely eager to figure out how you will tap into the promise of AI at your organization. 

Before we jump into the details, we think it’s important to take a step back and ask a broader question: What exactly are you trying to accomplish with AI? Do you want to invest in this technology because it’s a way to help you evolve your products and achieve your strategic goals? Or is it more about adding the latest innovations to maintain your market share and status?

We think this distinction is similar to the difference between an artist and an art collector. The artist is the one who is covered in paint, spending hours in the studio, and potentially throwing out dozens of canvases that aren’t aligned with their vision. The art collector, on the other hand, just gets to buy the finished picture and hang it on the wall. 

It is important to define a purpose for AI that is a better fit for your organization; neither is inherently ‘right’ or ‘wrong’. If you’re not able to invest the time and deal with the messiness of being “the artist” for the sake of being the creator of the masterpieces, acknowledge that being an art collector is the better fit instead. It is essential to be thoughtful about the “art” that you end up collecting. 

To discern what you are trying to accomplish with AI and how much AI investment you need to succeed, we recommend outlining: 

• Potential of uses of AI internally  

• Product areas where AI could add value

• Who would own and drive AI in your org 

Ultimately, ‘how much AI’ that fits into an organization is not static and can be part of an ongoing journey. It depends on many factors, including where you are currently and how much investment you realistically can make in the near term. 

Spectrum of Organizational Approaches to AI 

Our goal with this series is to help you understand what you need to consider to develop an AI capability as well as how to avoid the common ways most organizational transformation fails. We’ll cover the above organizational approaches to AI in more detail in a later post. 

Here’s an overview of what we’ll be covering:

• Part 1: Introduction

• Part 2: AI hype is new, our reaction to it is not

• Part 3: Is the product you’re building a good candidate for AI?

• Part 4: You are inspired to use AI: Now what?

• Part 5: Your AI Investment = (AI Needs) + (the necessary team)

• Part 6: Strategies for successfully integrating an AI practice into your org