There is a well-worn joke about a tourist lost in the countryside who asks a local for directions to the city. The reply comes: ‘Well, if I were you, I wouldn’t start from here.’ It’s the absurdity of the joke that makes it work. Where you are today is where you must start from. There’s wisdom in it too: where you have come from, and the road you’ve taken, can be just as crucial to the success of your onward journey.
A Microsoft study found that organizations already using AI were performing an average of 11.5% better than those that were not. It was published in 2019, which shows that while excitement around Generative AI (Gen AI) may be relatively new, the value AI more generally can add to business is well under way.
Gen AI represents the next wave of technology with the potential to unlock dramatic business efficiencies and outcomes, and to create new digital products, services and experiences. With AI, up until now, what we’ve had the opportunity to do is leverage data sets fundamentally to answer deterministic questions or to provide insight that could predict how something could play out. The significant difference with Gen AI is using the same data sets you would have used to predict, you can use to create new original content informed by the training data sets. That allows the creation of code, original content, new product design and processes.
In this sense, it is unlike any technology we have ever seen before and the temptation to rush headlong into its implementation, or at least to not get left behind, is understandable. According to a MarketsandMarkets report, the global Gen AI market is expected to grow from $111 million in 2019 to $4.5 billion by 2025, at a compound annual growth rate (CAGR) of 44.6% during the forecast period.
Future success depends on having the correct vision and foundations in place and ensuring that you evolve your organization’s capabilities to continually change at pace with the changes taking place around it.
As businesses stand on the brink of this next technological wave, to harness Gen AI’s power effectively, CEOs must pose three critical questions before embarking on the journey:
1. What is my organization’s AI vision?
At the heart of every strategic decision lies a choice. Having clarity on the business outcome you are pursuing is a critical component in determining what changes – organizational, technological or both – need to be introduced to get you there.
However great the temptation is to implement a new technology, it is crucial to understand what need you are addressing – is it to improve efficiency? Reduce costs? Improve the customer experience? Your AI vision will crystallize your organization’s direction of travel and ensure all decisions ladder up to support your desired outcomes.
Before setting a course to implement Gen AI, business leaders need a strong sense of where the role AI more broadly will play in their organization, where Gen AI (as well as other forms of AI) can create actual value for the business, and what the path is to implement to optimum effect. CEOs can determine where value resides by conducting an assessment of potential Gen AI applications against a customer journey.
For retailers, this might mean looking at top of funnel demand generation to conversion and checkout, to post purchase delivery and logistics. For automotive, it might be more specifically focused around telemetric data and unlocking its value. For consumer goods, the value is more likely to be found in supply chain and better short-term demand forecasting. Each industry will have a different set of drivers, in addition to regulation, making the hunt for value a tailored process.
2. What are our prioritized use cases?
Next you need to assess which activities performed in which parts of your business could benefit the most from the application of AI. In the buzz around Gen AI, it is also worth understanding what kind of AI is best suited for what task, the fundamental training data sets and the quality of those for what you’re trying to achieve and the costs.
In some areas you may be ripe for Gen AI, but in others more traditional predictive AI/Machine Learning may be more suitable. The cost implications are notable. The storage and compute required to drive a terabyte of data for a Gen AI solution could be between $1MM-$2MM per day, while a predictive AI/machine learning model would cost about $1K for the same usage.
By identifying and prioritizing your value pools, you will ensure that you are focused on the areas that can have the greatest impact for the business, are the most ready for AI implementation, and provide the greatest cost benefit. Gen AI excels in areas where more traditional machine learning methods may have struggled.
A great example of this is unstructured data. Machine learning is great with tabular data from transactional history to customer records used to predict cross sell, customer lifetime or customer churn propensity. Gen AI is more adept at making sense of unstructured data, such as insurance policy documents or knowledge bases, and providing an interface to interact with them. In these scenarios, Gen AI can shine a light on previously hard to analyze company information and build new sources of value.
This is an important stage of determining where you may want to make some big bets, where you may want to test and learn, and what to do and when. Prioritizing should take into account multiple factors from complexity, cost, readiness and value to name a few.
3. What foundations do we need to put in place to create efficiencies and accelerate return?
AI and Gen AI offer significant upsides to any business, but require thinking through the various kinds of investment that will accelerate your data & AI vision – from experimentation and developing GenAI playgrounds to the fundamentals of data collection, housing, maintenance and management and people, talent, training, and governance.
In AI, an approach that embraces constant iteration within the business will be key to success. A good place to start is an honest appraisal of your current state of readiness for adopting Gen AI. Having an assessment of where the gaps are in terms of skills, data, partners and technology will help to create a roadmap to prioritise investment decisions and scale progress from small experiments to enterprise wide adoption. Then it’s a case of organising around high value opportunities that can be rapidly iterated on and tested, building momentum and hands-on practical experience with these new technologies. You are seeking to build a data-led business in which your data & AI capabilities work in sync with your full set of digital capabilities, what Publicis Sapient calls ‘SPEED:’
- Strategy: developing and testing the hypothesis based on priority value pools
- Product: evolving at pace and scale
- Experience: enabling value for customers
- Engineering: delivering on the promise, at pace and at scale
- Data & AI: validating the hypotheses and uncovering insights for constant iteration
CEOs must be prepared to embrace iterative innovation in the realm of Gen AI. This concept encapsulates a culture of experimentation that hinges on testing hypotheses, learning from outcomes and adapting strategies. A mature experimentation mindset would mean embedding business stakeholders into the experimentation team, quick autonomous decisions, feedback loops integrated into the process, automated experiment scheduling, data cleansing and sequential testing techniques.