By Bryan Doerr, Executive Vice President of Product and Technology, Ensono Adapting two classic frameworks to today’s pace of change can help you craft a generative AI strategy built for agility and grounded in confidence. SECOND IN A MULTI-PART SERIES
Generative AI-based innovation can take many different forms. Internal processes can be refined and expanded; entirely new product and service offerings can be developed on top of existing applications; new client experiences can be created; insights from data can be multiplied. The possibilities are seemingly endless.
To address the vastness of generative AI’s potential without being overwhelmed or stymied by it, companies should think in terms of an innovation portfolio—one that encompasses many possibilities, but categorizes and prioritizes them based on a combination of risk levels, time horizons and potential impact. With this approach, companies can identify and capture opportunities for immediate gains through low-risk, short-term projects while also ensuring they can identify, assess and invest in long-term, high-risk initiatives. This allows for near-term improvements to ongoing business operations while positioning the company for future market success.
A diversified approach to innovation and investment isn’t a new concept for many companies. Innovation portfolio planning has long been a component of business strategy. But with the rate of technology change related to generative AI and corresponding market dynamics, the demands associated with assessing and prioritizing the innovation portfolio have changed. In the presence of so many amazing opportunities enabled by generative AI, deciding where best to focus investment and resources becomes harder and requires more analysis and deliberation.
Swift and Immediate (Horizon One)
Strategic and Balanced (Horizon Two)
Visionary and Forward-Thinking (Horizon Three)
Innovation analysis frameworks, used for identifying and validating innovation choices, help organize and navigate these difficult decisions, but these too need to adapt to the unique challenges of the current time. Innovation is expensive and too often betting on the wrong investment is risky, even when it’s backed by generative AI technology. Innovation frameworks—adapted for modern realities—help to mitigate this risk.
In a previous article, we explored how a broadened view of Mohanbir Sawhney’s Innovation Radar framework can help companies identify opportunities for generative AI-based innovation across the enterprise (“The power and limits of innovation frameworks in the era of AI,” Summer 2023). Now let’s consider how the adaptation and use of two other classic innovation frameworks, within the context of a portfolio approach, can assist in organizing and prioritizing these generative AI-based innovation opportunities.
The first framework to consider is the Three Horizons model. Originally conceptualized by McKinsey & Company1 and adapted by Geoffrey Moore,2 this is a strategic framework that enables businesses to categorize and evaluate their growth and innovation opportunities across three different time-based horizons:
Horizon One (0–12 months)
This horizon involves optimizing core business activities and offerings that are most tied to the company name, and provide the greatest profits and cash flow today. Think Apple iPhone or Google Ads. Activities include improving performance to maximize value. This is how the company determines its near-term success.
Horizon Two (12–36 months)
This horizon focuses on emerging opportunities that have the potential to generate substantial profits in the future but require considerable investment in the near term. The goal is to maintain today’s revenue growth and add new cash flow for tomorrow through strategic innovation. The challenge with Horizon Two innovation is that innovations can often demand resources from Horizon One activities, which are frequently withheld due to near-term pressures. Without needed resources, these innovations often stall and fail to transition into ongoing business operations.
Horizon Three (36–72 months)
This horizon is the most visionary of the three and involves exploring and investing in innovative ideas that could lead to the creation of new offer categories, disruption in the industry, entry into entirely new markets or radical improvements to existing business methods. Innovations in this horizon create future options for growth and efficiency.
1Coley, Steve: “Enduring Ideas: The three horizons of growth,” McKinsey Quarterly, December 2009. 2Moore, Geoffrey, “To Succeed in the Long Term, Focus on the Middle Term”, Harvard Business Review, July-August 2007.
Many companies fund Horizon Three innovations because it’s easy to argue that investment in the future is needed. The problem in Horizon Three is a lack of sustained focus because actual benefits are seen as too far into the future to be of value today.
Each of these horizons align to different strategic goals and success metrics, which require unique management practices and tools to achieve. Use of this framework allows companies to categorize, fund and appropriately measure the results of innovation activities. But while Three Horizons is a useful framework for categorizing innovations, the pace of generative AI evolution means the traditional timelines described above may simply not be applicable anymore. The model needs to be adapted to address this reality.
The needed adjustments relate to the pace of technological advancements and the timing at which innovations can move through each of the phases. Generative AI is evolving rapidly— literally month to month. This evolution is creating new opportunities and changing approaches to existing ideas at an astounding rate. In addition, tools and frameworks that embrace generative AI are emerging that enable quicker development, testing and deployment, allowing companies to shift innovative ideas from distant concepts to immediate or near-term actionable projects. Finally, the accessibility of generative AI is creating an explosion of ideas from many different sources, further driving the need to re-evaluate which activities are prioritized in various investment horizons.
The Three Horizons Framework is still a relevant framework for managing generative AI innovation, but the following changes should be considered:
These innovation opportunities will be more relevant and require much faster action. Consider reviewing the portfolio of activities and their associated success criteria more frequently, as a part of normal business operations, and train your organization to incorporate continual adaptation into “the way we do things”.
These innovation opportunities need time and funding to transition fully into the business, but this can’t be aligned with annual budget cycles—that’s too slow. Instead, businesses need to create a culture that pulls these innovations into production operations as opposed to legacy approaches that attempt to push them into relevance. And, because this increases risk, the culture must tolerate and learn from failure when a new innovation doesn’t work out as planned. Encouraging Horizon One sponsorship of Horizon Two innovations may help with this “pull” relationship.
This horizon is now 12–18 months, necessitating that these innovation opportunities are no longer the bastion of unfocused, lightly tracked activities. Innovations related to fundamentally new ways of doing business or with new offers or involving new markets, need to be managed with high levels of attention and management practices that ruthlessly prune innovations that won’t deliver benefits in this timeframe.
To achieve this way of managing the innovation portfolio, companies will need to reassess how they collect, evaluate and validate innovation opportunities for investment. This includes adjusting how they think about risk and potential, and even how often they conduct these types of assessment.
A more fluid, dynamic and agile assessment regimen across all three horizons is necessary. This is where our next framework—R-W-W—can help.
R-W-W stands for “real, win, worth it.” It’s a framework developed by Dominick M. Schrello, used to evaluate individual projects and innovation portfolios to uncover their business potential and risk exposure. Companies like 3M, GE and Honeywell have all adopted and adapted the R-W-W framework over the years.
R-W-W’s primary application is to help companies assess the fit of a new idea within their current market(s) and their current offer(s). The R-W-W screening process involves scoring a series of probing questions about the innovation concept, its market potential, the company’s capabilities or the innovation’s impact on a company’s capabilities, and the competitive risks, in order to identify issues that may hinder the innovation and manage risk. The questions attempt to help determine the following:
These questions aim to verify the existence of a genuine market need, and to assess the technical feasibility of the innovation. The goal is to determine whether or not the innovation project addresses a real problem or opportunity, and that the company can develop and deploy a viable solution to solve that problem.
These questions assess the company’s competitive advantage—or vulnerabilities—if they were to deploy the innovation. The goal is to analyze the company’s capabilities, the competitive landscape, and existing market dynamics to determine if the company can successfully launch and sustain the innovation.
Lastly, these questions evaluate the strategic alignment and financial viability of the innovation against the business’s short- and long-term goals. This includes analyzing the potential return on investment and the risks involved.
Step one in applying this framework is to create a risk matrix, such as the one articulated by George Day,3 to gain a clear picture of where various innovation projects fall on a spectrum of risk and feasibility. The risk matrix should employ a scoring system and a calibration of risk to estimate the probability of success or failure for each innovation project. A project’s position on this matrix is determined by how it scores on a range of screening variables.
To better analyze the risks in generative AI innovation within the R-W-W framework, consider the following additions to product and market risk assessments proposed by Day:
Generative AI can be used to create fake identities and falsify invoices and transactions, among other things. Therefore, it is important to ensure that the R-W-W framework includes product questions to identify the risk of fraud in new innovations.
Generative AI systems are reliant on the data upon which they have been trained, which could result in the AI system performing in an unexpected manner, including creating outputs that are not aligned with an organization’s own ethical principles. Therefore, it is important to ensure that the R-W-W framework includes an assessment of data availability and control to ensure reputational risks are managed.
Governments and regulatory bodies are increasingly focused on ensuring firms have adequate AI governance structures and controls in place. Failure to comply with these requirements could expose firms to regulatory fines. Therefore, it is important to ensure that the R-W-W framework includes market questions to identify the risk of regulatory non‑compliance.
A diversified approach to innovation and investment isn’t a new concept for many companies. But with the rate of technology change related to generative AI and corresponding market dynamics, the demands associated with assessing and prioritizing the innovation portfolio have changed.
3Day, George, “Is It Real? Can We Win? Is It Worth Doing?: Managing Risk and Reward in an Innovation Portfolio,” Harvard Business Review, December 2007.
Like the time variable in Three Horizons, the answers to the R-W-W screening questions, and hence the innovation’s place on the risk matrix, can change quickly thanks to rapid evolution in generative AI technology.
Competitors are regularly coming out with new offers that leap the entire market forward, requiring other generative AI companies to adjust their product roadmaps. At the same time, consumer demand, regulatory changes, ethics considerations and myriad other forces push and pull on these plans, with a cascading impact on the industry as a whole. Companies that build on generative AI technology—both internally and externally—need to adapt their screening cadences and processes to this market dynamism.
As the technology evolves, the lens through which innovations are assessed also changes. As companies roll out generative AI innovations, they gain new insights and knowledge that will change their assessment of future opportunities. And as new forces are introduced, risk factors may intensify or lessen. The best approach, therefore, is to not be rigid, but to adopt and adapt innovation frameworks that have been proven effective to the new generative AI landscape. R-W-W is an ideal framework for this, but the following adaptions on its use should be considered:
Rapid ideation and faster time to market for new products can democratize access to information, enabling contextual search and transforming information retrieval to be conversational, impacting both customers and employee experiences. It is important to ensure that the R-W-W framework is enhanced to accommodate the unique accessibility requirements of generative AI.
Generative AI can help businesses accelerate the development of new products and services. To ensure that the R-W-W framework is optimized for speed to market, consider incorporating agile methodologies into the innovation process. Agile methodologies prioritize speed and flexibility, allowing businesses to quickly iterate on new ideas and bring them to market faster.
Generative AI has the potential to disrupt entire industries. To cope with this disruptive nature, it is important to adopt a proactive approach to risk management. This includes not only identifying potential risks early on, but also developing contingency plans, and continuously monitoring the innovation portfolio to ensure that it remains aligned with the organization’s strategic goals.
Assessing and prioritizing generative AI innovations based on a time horizon scale alone is not sufficient. Nor is an annual planning and risk assessment of an innovation portfolio through the R-W-W lens. What’s needed is a combination of the two: A kind of three-dimensional chess game that accounts for risk, feasibility, market forces and the truncated timeframes on which all of those variables change.
By layering the categorization framework of Three Horizons with R-W-W—both before a project begins and frequently mid-flight—companies can rapidly shift and reorganize their investment portfolios on a continuous basis as new assessments occur and new opportunities enter the pipeline.
But to accomplish this, companies need to adopt a spirit of agile innovation at the core of their culture. Decision modeling, processes, tooling and workflows all must be adapted to encourage constant assessment and experimentation in these early phases of generative AI evolution. Leaders will need to foster and nurture the creation of AI-first organizations that identify opportunities proactively and assess them in new ways.
