Digitalization requires new strategies
With the advancement and proliferation of Internet of Things (IoT), cloud computing and Artificial Intelligence (AI), the digital revolution has begun in every industry. Many conventional businesses embrace the exhilarating transformation and straddle into a new territory. For example, some banks and insurers have repositioned themselves as “technology” companies. However, many sectors are still lagging on the digital journey. What’s the secret winning recipe to the digital agenda? Let’s answer this trillion-dollar question by discussing the why, the what and the how.
Digitalization is inevitable. It’s simply a matter of how, not when. The amount of data produced everyday is mind-boggling. As of June 2019, the world is comprised of over 4.4 billion internet users. There are roughly 2.5 quintillion bytes of data created each day and the rate is accelerating with the growth of IoT. This year’s global spending on IoT is on track to reach $745 billion and will continue to grow by double digits through 2022, surpassing $1 trillion, according to International Data Corp. (IDC). Companies just can’t ignore or even fall behind the digitalization wagon.
Digitalization makes good business sense. Many successful AI use cases have proven that companies gain a multitude of benefits, such as unlocking new customer insights, improving user experience, gaining operational efficiency and creating new products and services. And through those various benefits, financial measures will improve as a natural by product. The first mover advantage and network effects could very well ensure competitive longevity and viability.
What is AI or Big Data? These buzz words can be vague and superficial. Beyond the technical explanation, in a practical sense, what is the digital transformation? There should be three stages: the Data stage, the Analytics stage and the Enterprise stage. We can refer to this as the DAE journey (see below).
The Data stage is the foundation, the most crucial first step. The reason is simply due to the unique nature of machine learning algorithms. Most algorithms start with a programmer inputting the algorithm. However, in machine learning, it’s the data that creates the model. We call this training the algorithm. It’s a self-learning and correcting cycle. The more data there is, the more sophisticated the algorithm becomes. Therefore, the type, quantity and quality of the data directly affects the validity of the algorithm predictions. Companies should make every effort towards building solid data reporting structure (for both structured and unstructured data) and data governance best practices. Breaking down silos among various standalone systems, such as ERP, CRM, manufacturing and supply chain, service, etc., and having a central data warehouse and good data integrity standards are crucial to the success of the AI capability. Having one source of truth may sound trivial but challenging to achieve.
The Analytics stage is all about building competency. Once the foundation is strong, in the analytics stage, the goal is to identify pilot projects to apply advanced analytics techniques to solve business pain points. These projects should start with internal facing “customers” and be manageable in size and complexity, so external help can be acquired to support. This will be an iterative process, for there may be several machine learning algorithms applicable to analyze the available data. The intent of the analytics stage is to create a safe learning environment for the organization to develop internal talent, and cultivate an evidence-based, data driven decision making culture. The outcome should be a robust internal competency on AI accompanied by a proven best practice governing the entire cycle of AI assisted business problem solving.
The Enterprise stage is pervasive data-driven decisions. With a solid foundation and competent AI capability, the focus now is about scaling up at the enterprise level. There should be systematic enterprise coordination of analytics activities. A central analytics organization structure should be implemented with direct line of sight to all other business departments or units. Data-driven decisions should be ubiquitous impacting key strategies such as go to market strategy, product and service offerings, pricing, supply chain management, product development, etc. Amazon is the best example as a company that’s excelled at the Enterprise stage.
The three stages of DAE must be treated as a linear journey and used as a gauge to assess the optimal starting point, since each company’s status quo on the digital transformation may be different. Some may have already taken a jump start, while others may need to take a step backward and reshape the foundation. However, real life examples have taught us that any attempt of skipping steps on the DAE journey or tackling two stages simultaneously will backfire.
The best strategy to conquer this quest is not to have a digital strategy. This sounds counter intuitive, but here are some proven approaches.
Just do it. When examining the various success stories on AI adoption, the similarity is that all the successful teams acted as entrepreneurs and overcame the learning curve by doing, not by debating on strategy or polishing AI business cases. Unlike conventional business cases, AI business cases can demand costly upfront investment without providing immediate paybacks. AI solutions require drastically different technology and problem-solving mentality. The dedicated focus on data cleansing and governing, spending the time to train the model and evaluate various algorithms is tremendous. In addition, AI business cases will require fundamental cultural change. Think of AI as a representation of a business decision model, not a process event. Therefore, applying conventional business planning process to AI will begin a paper exercise that will likely end in a stagnated paper exercise.
Initiate the top-down culture change. The real winning strategy of AI is to build the mindset of embracing changes and developing evidence-based decision-making habit. To break down the existing system silos, the best chance of success is a top-down driven behavioral change. Executives should directly engage and sponsor data initiatives and become role models in promoting advanced analytics applications throughout the organization. One of the key benefits of implementing AI is that it enables companies to be customer obsessed. As a result, new customer value propositions are likely discovered, which may push the organization into unexplored territory with new product and service offerings. This intimidating yet exciting paradigm shift could allow large conventional corporations reinvigorate itself. For example, China Communications Construction, a multinational engineering and construction company has embarked upon a new mission of “building a connected world,” by offering big data processing and cloud computing services.
Own the competency by building a dedicated team. A typical business challenge is normally met with three options: build, buy or outsource. However, AI is atypical, and the only right answer is building your own. Owning the AI governance and skills will be essential to the viability and competitiveness of the company, especially as digital revolutions continue to evolve, and technologies rapidly innovate. It’s certainly acceptable to acquire external help during the initial pilot and learning period. But the knowledge transfer must be planned from external vendors to the enterprise AI team as the team matures. A central AI organizational structure should be established with high visibility to the C-suite and equal exposure to all other business units and functions. AI team building should be comprehensive and diverse. Often the vital talent search is only fixated on data scientists. But it takes a whole army of people with industry domain knowledge, company specific business understanding, project management skills and most crucially expertise to translate the analytics results into business actionable insights. Both external hiring and internal training avenues need to be explored, but with more emphasis focused on training existing employees. Because there’s just not enough external supply of AI talent to go around. Role specific AI training program should be put in place for the broad organization, especially at the executive level, there should be adequate understanding of AI.
Start small and forget about the ROI. Since the AI transformation journey must start by doing, the ideal pilot projects should aim for fairly “soft” outcomes, such as process improvements or customer satisfaction. Don’t’ worry about the immediate ROI. The intent is to nurture a safe environment so teams can fail faster. Don’t rush through the analytics aspects. Test and debate about the various algorithms and outcomes. For example, predicting how likely a patient may get a billing code for a disease is not the same as predicting how likely this patient may get the disease. The initial pilot success is most critical, for it will rally the broader organization to strengthen support and gain further momentum. And this accelerated momentum will eventually help reach the tipping point.
Connecting the Dots
If data is the new oil, then AI is the best drilling technique to unlock the enormous wealth. However, the secret recipe is people first, machines second. Building a dedicated central in-house AI team, equipping the organization with the data-driven mindset, and establishing a solid foundation of data governance and aggregated data warehouse, are the best combination to win the digital revolution. The AI agenda must be executed with the entrepreneurial spirit, masterfully selecting the initial pilots to build momentum. Unfortunately, the danger in conquering AI is often complacency and self-deceit. Companies can invest heavily in acquiring the AI technology without systematically following the DAE journey. To which, it is important to remember just because a company can crunch numbers with AI technology doesn’t make it an AI company.
Written by: Jason Kang
Jason Kang is currently the global director of Strategic Marketing for Automation/IT at Siemens Healthcare Diagnostics Inc., where he’s responsible for developing the product marketing and digitalization strategy. As a Kellogg MBA class 17’, his healthcare career ranged from medical device consulting to managing commercial operations at GE Healthcare in the greater Houston area. His diverse med tech expertise spans both in-vivo and in-vitro technologies in multiple global markets. Email: Jason.firstname.lastname@example.org Linkedin: https://www.linkedin.com/in/jkang2017/