Artificial intelligence deployments are fraught with technical and tactical elements that have to be executed well in order to see a return on investment: The data must be accessible, cross-functional AI teams have to work together, and even after an AI pilot seems promising – it often needs to be integrated into legacy systems to be deployed successfully.
Above both the technical and the tactical issues lies a greater set of looming questions:
What is AI helping our department or firm to become?
How will it help us serve our customers, improve our operations – and ultimately win in the market?
At completely digitally and data-native firms – including much of Silicon Valley and big tech – these questions are often answered early on. Digitally native firms often have an understanding of AI’s role in their product, operations, and strategic differentiators early on. Firms like Microsoft have enough AI fluency among their leadership to adapt their business models and strategic plans to integrate AI into their vision.
For enterprise firms without data and AI in their DNA (i.e. the vast majority of firms), the issue of AI’s larger strategic role in the company is a massive impediment to turning AI into value.
The goal of this article is to help provide a framework for aligning what we call an AI Transformation Vision. While this resource was created with established enterprise firms in mind, the same framework can be applied to mid-sized firms or even to young companies who want to consider how they can wield AI to their own advantage.
We’ll be covering the following topics in order:
- The Value of an AI Transformation Vision
- The Components of an AI Transformation Vision
- How to Use an AI Transformation Vision
Throughout this article we’ll be referencing the AI Transformation Vision (ATV) template featured in the image below:
The Value of an AI Transformation Vision
Near-Term – Alignment: Most AI projects fail because of issues that can tie to executive AI fluency. Poor project selections, unrealistic ROI expectations, lack of strategic focus, and a lack of understanding of AI maturity all contribute to the 80-90% failure rate of AI projects. An ATV helps companies build the fluency and buy-in that they need to see real deployments and near-term results.
Near-Term – Avoiding Technical Debt: AI approached as a series of “plug-and-play” point solutions (without any fundamental improvement to AI maturity, data infrastructure, etc) leads to a kluge of spun-up data repositories and stand-alone algorithms that can add tremendous technical debt to an organization. As opposed to a hundred small data ecosystems, firms with a coherent vision can think about the capabilities they want to unlock, and build a shared “foundation” for new AI capabilities – unlocking not just one application, but many future applications.
Long-Term – Advantage: AI advantage won’t be gained by a single successful deployment, but by a transformation of how a business operates, and how it can win in the market thanks to AI capabilities. An AI Transformation Vision is the only way to enable genuine strategic transformation – as opposed to the scattered and disjointed AI investments that most enterprises engage in today. Started the journey towards a successful long-term vision, and investing in that direction, will immediately set your company apart.
An ATV allows firms to:
- Align teams and leadership towards a transformation roadmap that will lead the company to success in the market
- Determine what elements of AI maturity to build
- Determine which strategic anchors their projects can drive towards
- Determine which AI projects to pursue next in their AI journey
The Components of an AI Transformation Vision
Time horizon: The range of years used for this exercise. We generally recommend a 5-7 year window – but this same exercise can be used to help leadership teams think through longer or shorter time spans as well.
Aspirations: The specific goals that a company wants to achieve within the time horizon provided. These are the objectives or outcomes of our transformation – a statement of where we want our transformation to lead us.
This might involve the opening of new lines of business, a core transformation of who we serve and how, or simply growth, profit, or market-share related aims that AI and digital technologies might help to achieve.
Example(s):
- Having X% of US market share in X industry by X date
- Achieving X billions in revenue, with X% in recurring annual contracts, by X date
- Transitioning to X% of annual growth coming from X new product lines
Lines of business: Related product or service groups appealing to a particular buyer need.
Example(s):
- A brick-and-mortar retailer might imagine the expansion of their eCommerce business, and the growing importance of online customer activity data and recommendation engines.
- A super-regional bank might consider which of their product lines will become less competitive in an increasingly online and digital future. Some product lines may fade away entirely, while other services that still have a local advantage might be streamlined and improved with various back-office AI applications.
Map out the current state of your lines of business – and the ways those lines of business will change through the achievement of your AI transformation vision.
Consider meta-trends in your industry and how your firm will adapt. Consider specific AI capabilities that could further entrench your relative market advantage.
Questions to ask:
- How will Al enable new products and services, and change how we serve our customers differently in the future?
- Which lines of business will become less important in our Al-enabled future?
Related interviews:
- How to Rethink Banking in the Age of AI – with Ian Wilson, Former Head of AI at HSBC
- The Evolution of the Data Business in the Age of AI – with Shri Santhanam, Head of Advanced Analytics and AI at Experian
Core business processes: The operations that require the most resources, or have the biggest impact on the overall performance of the firm.
Example(s):
- An insurance firm might focus on improving the efficiency and effectiveness of their underwriting processes, particularly for their fastest-growing insurance segments.
- An online travel firm might focus on improving its ability to quickly match travelers to related travel amenities, including transportation, lodging, entertainment, and more – and might leverage AI to find pairings that suit the preferences of each unique user.
Map out the current state of your most important business processes – and the ways those processes will change through the achievement of your AI transformation vision.
Consider not only where efficiencies might be gained through automation and digitization, but also consider which processes might be eliminated entirely – or completely re-designed – based on changes in your lines of business.
Questions to ask:
- Which of our core business processes will become automated or less important?
- Which of our business processes will become more important as drivers of growth?
- How will Al fundamentally change how our functions operate for the better?
Data sources: Important data sources involves in serving customers and improving operations.
Example(s):
- A wealth management firm that sees relationship management as the key to maintaining growth in an increasingly commoditized space might place more emphasis on tracking client interactions (in-app, on their website, with wealth managers) in order to find patterns of communication and engagement that improve retention.
- An oil and gas firm might decide to place higher importance on collecting data related to efficient fuel transportation – allowing them to unlock data’s value to optimize routes, loads, and transportation types in order to keep costs to a minimum.
Map out the current state of your most important business processes – and the ways those processes will change through the achievement of your AI transformation vision.
Consider which data sources are most important to your business today and how they might be enhanced or made more accessible. Consider also the data sources that would empower your future product lines or provide a sustainable data “moat” for the long term.
Questions to ask:
- What are the proprietary data sources we want to start or keep collecting in order to differentiate our product, service, or operations?
- How will data fuel a differentiated advantage in our efficiency or customer experience?
Related interviews:
- Build a Data Moat Around Your AI Initiative – with Adam Oliner, Head of ML at Slack
- Assess Your Data to Find AI ROI Opportunity – with Adam Bonnifield, VP of AI at Airbus
- How to Raise Money for Your AI Startup – with Ben Narasin of Canvas Ventures
Customer segments: Distinct customer groups as defined by demographic/firmographic factors (age, gender, company size) or product lines (insurance customers, auto loan customers, etc).
Example(s):
- A regional bank considering its sustainable advantage with data about local customers may decide to focus more on the customer segments most likely to grow in a digital future.
- A niche retailer may use its burgeoning eCommerce and digital presence to target younger customer demographics, and to find growth with almost entirely new customer segments that their local stores are unable to reach (i.e. international customers).
Map out the current state of your customer segments – and the ways those segments will change through the achievement of your AI transformation vision.
Some customer segments may become vastly less important, or vastly more important. Entirely new markets may open up and become relevant through your ATV, creating new and important customer segments that don’t exist today.
Questions to ask:
- Which markets are poised to grow or shrink given market trends and Al impact?
- Which markets and segments will we best be able to support?
How to Use an AI Transformation Vision
Executing on an ATV involves change management, product ownership, processes for co-creating products with customers, and more that we aren’t able to cover in the limited scope of this article – but I hope that the overview presented thus far will provide you with a starting point and an accessible transformation framework.
A company’s future vision and strategy involve much more than just AI. An ATC is part of a larger digital transformation of the company broadly.
Plenty of the existing literature and work on developing a “digital vision” applies to AI or any other emerging technology. That said, AI has such an emphasis on data – and opens up so many new capability paths – that we believe it’s warranted for most enterprise firms to consider AI individually, in addition to being part of a larger digital ensemble.
- Firms with a robust digital transformation vision should usually have AI capabilities and possibilities rolled into that vision – as opposed to redrawing the strategy entirely.
- Firms without such a robust strategy in place should form their broader transformation vision with AI in mind – potentially by using the framework outlined here.
Bear in mind that an ATV – or digital transformation vision broadly – is a living, evolving document, and should be reviewed quarterly, or scrutinized in-depth every year at a minimum.
It’s unlikely that a company will have one of these, and it would take strong champions, C-suite enthusiasm, and a dire company-wide priority to make this vision happen. However, it should be seen as a north star direction to head in. The more components of this vision you can get to, the better.
This is the job of an AI catalyst, to educate leadership throughout the process of ideation, planning, and deployment, to get functional teams and executive teams closer and closer to an AI transformation vision – or at least towards a strong understanding of AI and a connection between AI capabilities and a company’s strategic goals.
