People perspective: Culture and change towards AI-first

Adopting AI and creating value reliably and repeatably is not a purely technological challenge. Any AI initiative is crucially dependent on the people that guardrail and drive it. While AI as a general-purpose technology will impact sectors, those organizations where the workforce embraces its capabilities will be successful. This becomes all the truer when considering how good AI systems come to live: Through collaboration between stakeholders, business units, and practices.

There often is talk about the potential of AI to automate human labor, when in reality it enriches, supplements, or even augments human labor. While some domains are in reach for automation, today's AI is largely about helping with tasks that are perceived as specifically complex for humans. We see that organizations that are AI-first reduce operating costs, increase revenue, and give challenging, meaningful work to employees. Rallying the organization, building up the right talent, and speaking the same language when searching for valuable business problems is the focus of this perspective. Culture is king, even more so when adopting AI. This perspective comprises seven capabilities shown in the following table. Common stakeholders include the CIO, COO, CTO, cloud director, and cross-functional and enterprise-wide leaders. 

New: ML fluency

Building a shared language and mental model.

The boundaries and semantic scope of artificial intelligence and machine learning is not well specified. Both terms are also overloaded with varying mental models and emotional interpretations, which is why it's key to align internally on what stakeholders mean by it. Spread a largely aligned perspective on what these words mean and identify those stakeholders that are intrigued by it as your future internal AI champions.

Once that first layer of interpretation is spread across the organization, tackle the second, more technical one: AI projects and requirements differ in terminology and what importance is assigned to them. From the product management practice to the engineering and data science practice, align on what joint understanding is needed to work effectively. An effective way is to define interface words between different practices, for example, how can success be measured in ML versus how can it be measured in the business domain.

Implement these alignments through ML fluency and ML culture trainings, as they will help you get buy-in throughout your organization. It's likely that this understanding will become crucial in helping business owners adapt to the unique aspects of ML use cases and setting expectations with customers.

Lastly, consider how to best communicate AI outputs both in the organization and to customers. Consider that customers will have different mental models and terminology, so such as, letting an AI system gracefully fail and keeping trust is challenging. With the right language and fluency, you will not only be more efficient but also reduce the risk of building systems that don’t align with interests of your customers.

Workforce transformation

Attracting, enabling and managing AI-talent from user to builder.

Being able to attract, retain, and retrain talent that can push your AI strategy forward is one of the most crucial aspects of AI success. There are many roles that are necessary for AI success, some of which you can outsource while others can only have their impact as the in-house workforce. As a first step, your AI strategy leaders need to be tightly connected to your business and drive value from within. This role can seldom be handled by an outsourced firm.

Enable these leaders by hiring or developing the many roles that are needed for successful AI adoption:

Tightly align your hiring strategy with your overall AI strategy and ambition:

In addition to this internal workforce, bet early on the right AWS Partners to not fall prey to your AI agenda fizzling out. When talent is not present, broadcast your AI vision externally and start to run initiatives that will yield both results and inspire new talent. Recognize from the beginning that retaining talent in AI is difficult, as supply has historically been outstripped by demand. Another factor is that real-world AI differs significantly from the academic work that often drives talent into AI. Counter this factor by having opportunities for your AI experts to collaborate, present at conferences, and write whitepapers.

Attrition, however, is unavoidable. Be flexible and establish processes to hire talent with proper timing and to keep resources on deck to fill in when attrition occurs. The processes we reference in other parts of CAF-AI are crucial to helping make your business robust against attrition. Fuel your AI workforce through continuous re-training opportunities to learn new skills needed to perform well in the AI space. This approach has an added advantage of being able to have a person that has in-depth business knowledge as well as being able to run projects. Lastly, recognize that the headcount-to-value ratio in AI is lower than in other fields. A small team of strong practitioners typically outperforms larger teams as the work is less mechanical than intellectual.

Organizational alignment

Strengthening and relying on cross-organizational collaboration.

When AI becomes top-of-mind for organizations, providing an encapsulated and empowered separate unit that spreads and disseminates its value and knowledge across the organization is a typical first step. The AI center of excellence (COE) is a unit that can fill this role, where AI-focused teams are hired and evolved. Make sure that reporting lines in this organization align with those stakeholders that have ownership over the AI strategy in the organization and make sure that there are short paths to the C-suite. Do this to make sure decisions and changes can be made quickly when needed, and new teams can find their rhythm. At the same time, it's crucial to align the incentives of such a COE with your strategy, business, and most crucially your customers. A common mistake is to evolve AI units that do not deliver on business value.

Over time, your workforce transformation should enable your broader organization and other builders to effectively use the COE and existing AI services, as well as collaborate effectively. Be sure to prevent a not invented here syndrome, so the organization does not rebuild what is readily available in the cloud, provided it fulfills your business requirements. Make sure that your COE and talent develop an engineering mentality, recognize the cost of maintaining disparate systems, and establish an MLOps best practice that brings a DevOps mentality to the culture. While such units, other internal builders, and AI talent evolves, enable your data flywheel by establishing a data-driven product mentality. Permit businesses across the organization to not only share and govern data, but also establish a vivid ecosystem of data products. However, don’t build such data products for their own sake.

Culture evolution

Culture is king, even more so when adopting AI.

Developing an AI-first culture is a long and challenging process as it often requires breaking up old mental models. In typical cloud and software development, the cultural focus is on empowering builders to codify complex rules and systems. AI relies much more on a culture of searching for the right inputs that generate the desired output. To circumvent a culture that is centered around technology, embrace a mentality where builders, the business, and other stakeholders work backwards from business opportunities and customer needs to all the AI challenges.

Working backwards means pre-formulating the expected result of a change in your business environment and then asking what needs to happen to achieve that change. In a way, this is how AI systems are built: Defining the expected outputs, and then searching for inputs that contain a signal to enable that output.

With such a value-driven mindset in place, zoom in on the cornerstones of an AI-first culture:

Start expanding your AI-first culture with the following: