Customer-Centric AI: the Major Paradigm Shift in AI Governance (Part 3)
What actions to take to become a customer-centric AI company
In part 3 of this essay series, we will discuss the WHAT question: what actions to take to build a customer-centric AI company. I highly recommend reading part 1 and part 2 first, the WHY question and the HOW question, to better grasp the context.
Customer-centricity is not just a strategy, it’s a mindset. Mindset determines every action you take as it comprises a set of belief systems, values, and opinions. When customer-centricity is deeply embedded in your mindset, it influences decision-making processes, shapes company culture, and drives innovation. This means prioritizing the needs and experiences of customers at every level of the organization, fostering a culture of empathy, and continuously seeking ways to enhance customer satisfaction and loyalty - every team member is committed to the mission.
Specific actions you need to take towards this mission depend on your context as we discussed in previous parts - context decides everything in the case of AI. The following factors will most likely affect the way you apply customer-centricity to your organization:
Size of your organization
Structure of your organization, roles of the team members
Industry
Country (regulations, culture)
Business model (B2B or B2C, revenue models)
Your role: developer or deployer/provider
In this article we will discuss approaches that apply to all the cases but remember, this only suggests ways to get started, the real impact comes from continuously evolving and adapting these principles to suit the unique needs of your customers and the specific context of your organization.
We will start our discussion by separating 5 layers of AI Governance:
Legal Layer
Organizational Layer
Technological Layer
Societal Layer
Geopolitical Layer
These layers certainly have overlaps, they are not precisely separated but they are good guidance to see different aspects of AI development. The organizational layer is most important for every company, at the end of the day every other aspect turns into an organizational question: aspects that happen independently of your organization, you need to incorporate them and manage additionally. We will discuss actions that you should take, separately for executives and ML engineers, for each of the 5 layers of AI Governance - as we discussed in part 2, both bottom-up and top-down change is significant to make the shift towards customer centricity happen.
The WHAT Question Answered
All the actions you need to take within each of the 5 layers of AI Governance to build a customer-centric AI company:
Legal Layer
5-6 years ago, when the AI Governance field was in its seed (not even bud), one major question was circulating - to regulate or not to regulate [AI]. For most ML engineers, regulating AI was something unimaginable or even laughable, I myself experienced these laughs in my face as someone researching this question from the very early days. Fast forward within 2-3 years, some of the world’s leading countries found their answer and now we have the world’s first AI regulations heading towards enforcement. However, for me this question (to regulate or not to regulate) stays broadly open, moreover, I can somewhat relate to those laughs from ML engineers. Let me explain: (a) AI is developing faster than policy-makers can keep up and unless policy-making mechanisms change, it will always stay like this. (b) It will be hard or near impossible to audit the inner workings of complex AI systems due to explainability/interpretability obstacles; (c) AI systems can adapt and evolve over time, making it challenging for static regulations to keep pace; (d) Regulating against bias is complex due to the nuances in defining and identifying bias. Regulations have to regulate only major human rights issues, the rest has to be done by companies themselves. This is why the AI Governance field is far broader than regulations.
This means legal documents can be helpful for customer-centricity in a way that they are created with basic human rights in mind, hence they already fulfill fundamental customer-centricity requirements and serve as the initial guidance on your customer-centricity journey. However, true customer-centricity goes beyond legal compliance; it involves actively listening to customers' needs, continuously improving user experiences, and fostering a culture of empathy and responsiveness throughout the organization.
Five actions to take to use AI regulations to your advantage:
Organizational Layer
Change is hard but worth it. We are in a transitional period where everyone at this moment working in the AI field decides how AI will serve humanity in the (near or further) future. Once we embed these systems in our daily lives, they will shape our interactions, influence our decisions, and redefine societal norms. Hence every decision you make in your organization affects not only your direct customers but also broader society, city, country, or even global population. This means, you are taking responsibility upon yourself to decide on humanity’s future and you want to make sure that you fulfill this responsibility at the highest standard. This further means that it does not have to be external forces such as regulations to incentivize you to take action - this responsibility comes from within and you must build internal organizational mechanisms for effectively managing and governing the entire AI development life-cycle. However, if this responsibility is not internalized, then external forces will have an effect. Not just regulations though - if you remember part 1 of this article series, we discussed how much market tendencies have changed and how demanding and critical consumers have become. Unless you are truly customer-centric and put customers’ best interests at the front, you risk alienating your user base, undermining trust, and ultimately compromising the long-term success and reputation of your organization.
Five actions to take to build robust organizational governance for customer-centric AI products:
Technological Layer
You can back your decisions as much as technology allows it. For example, even if you wanted to implement complete algorithmic transparency and explainability in your organization, the “black box” problem would not provide much leverage. However, this fact does not give you excuses for not implementing certain organizational policies. For example, even if the model is not explainable, you must continuously monitor AI models in production to detect and mitigate any biases, errors, or drift that may occur over time, additionally check how your model performs in different contexts, societies, or countries. This way you can backtrack the model’s decision-making and notice inefficacies. After all, you are an AI company that innovates on AI technologies - it is in your best interest to innovate AI systems that adhere to your and your customers’ goals and; additionally to incorporate the latest achievements of others as currently there are huge efforts undertaken by big technology firms, non-profits, governments and research institutions to work out new AI Safety tools, why not use them?
Customer feedback loops are extremely helpful here: any issues, such as unexpected biases or errors, can be identified and addressed promptly. This proactive approach not only helps maintain the integrity of the AI system but also builds trust among users who see that their concerns are taken seriously and acted upon.
Here are five actions to take to implement best-in-class technological innovation and make your customers happy:
Societal Layer
Imagine, just within one decade, how Facebook and other communication apps changed the ways we interact with each other as a society. Not only we want to understand and streamline individual customer experiences but also how these individual experiences relate to society and the entire system (city, group, country, community, etc.). Especially as AI technologies are far more effective in shaping behavior and shifting our reality.
Uber’s AI is a good example of such societal change. Uber disrupted not only the taxi industry but also the entire economy with its smart trip planning and pricing algorithms: Traditionally, taxi fares were fixed or determined by negotiation. Uber's AI considers factors like real-time traffic, demand, and distance to dynamically adjust prices. This has led to lower fares for consumers, increased efficiency, and the rise of the gig economy - the driver-partner model created by Uber has become a blueprint for many other companies fundamentally changing how people approach employment across various sectors. It offered flexibility for workers but also raised questions about job security and worker benefits.
In this regard, your customer-centricity strategy must include not only individual customers but all the customers collectively, and any other groups directly or indirectly related to them - this is important for societal acceptance and building trust as AI is a brand-new technology. For example, if you were Uber, how would you convince society that the gig economy offers better (or equal) benefits than conventional employment and how would you compensate for workers’ benefits
Humans are social beings, everything one individual customer does is reflected back and forth in their surroundings. Customer-centricity means foreseeing and responding to these interactions so that consequences are manageable.
Here are five steps to incorporate societal considerations in your customer-centric AI strategy:
Geopolitical Layer
Geopolitics has traditionally been about states competing for territory. However, since the end of the Cold War, there have been significant changes. Now, cities, organizations, and corporations also compete for territories, but territories are now with the additional dimension, of cyberspace/virtual space. This new dimension introduces unconventional means of influence, such as leveraging digital platforms for global collaboration, using virtual spaces to foster innovation, and harnessing social media to build inclusive communities and drive positive social change. These tools enable organizations to reach and engage with a diverse audience, promoting cross-cultural understanding and cooperation. However, on the other side of the coin, bad actors use digital propaganda, cyberattacks, data manipulation, and the strategic use of social media to sway public opinion and behavior on a global scale.
I have extensively discussed this topic in the paper I co-authored, “Geopolitics of Smart Cities: Expression of Soft Power and New Order”.
And Substack article: “Anarchic State of AI Governance”
The two AI superpowers, the US and China, have diametrically opposed political systems, leading to divergent approaches to AI development and governance. However, other countries like India, UAE, and European nations are also emerging as significant AI players. These powers compete to set and spread their standards for AI globally starting from capturing powers at ITU (International Telecommunication Union) continuing with subsidizing the internal tech sector (to help local champions capture the global market) and even regulating (especially the EU with the so-called “Brussels Effect”).
The major customer-centricity question in the geopolitical dimension is how an organization can reconcile the political and cultural aspirations of different state actors, whether you’ll take sides, and whether you’ll become one of the players to strive for dominance. This has everything to do with what values and goals you set for yourself and how you interact with your customers in different parts of the earth. Navigating these complexities requires a nuanced understanding of local contexts and a commitment to maintaining ethical standards that respect diverse cultural perspectives, ensuring that your actions consistently align with your overarching mission to serve and respect all customers effectively.
Five actions to take to navigate new geopolitical realities:
Conclusion
At the center of all these layers of AI Governance is always an individual. Even if you are a B2B company, interacting primarily with other businesses, you are still engaging with individual people to make business deals. Your major task is to understand the individual you are serving with your products or services.
The major difference with product-centric companies is that product development is led by engineers who know the product. In contrast, in customer-centric companies, product development is led by those who understand both the product and the people for whom the product is created. When organizations value the product or technology more than the customers, they risk losing sight of the true needs and desires of their users. This can lead to products that, while technologically advanced, fail to resonate with or meet the practical needs of the people they are intended to serve. Prioritizing customer insights ensures that the development process remains aligned with user expectations, fostering stronger customer relationships and driving long-term success. Not even most successful tech corporations are safe from making this mistake of prioritizing product over customer:
In part 4 we will discuss why customer-centricity is the major paradigm shift in AI Governance. Stay tuned!
Ana Chubinidze is the founder/CEO of AdalanAI, building a novel approach to AI Governance.
email: ana.chubinidze@adalanai.com