Beena Ammanath – Global Deloitte AI Institute Leader, founder of Humans For AI and author of “Trustworthy AI.”
In this age of humans working with machines, an effective leader with artificial intelligence (AI) requires a range of skills and activities.
AI is often seen as new, groundbreaking applications that offer an almost insurmountable competitive advantage in the marketplace. Sure, “moonshot” AI efforts can lead to powerful new tools, but the broader value of AI can be found in numerous implementations that augment multiple parts of the business. Not all of these (or even most of them) are new implementations. Things like robotic process automation, computer vision, natural language processing, and other types of AI have matured to the point that many of them are available as ready-to-use applications or offered as a service.
The AI acquisition question for business leaders ultimately boils down to: buy, build, or collaborate?
To begin with, understand that AI is not just one thing. It is the product of a constellation of hardware, software, data, other assistive technologies and highly trained human talent. The availability of these component elements determines what makes the most strategic sense for the company in terms of AI acquisition. Leadership is needed in answering the buy-build-partner question because the decision affects multiple areas of the business and there may be a lack of harmony between how stakeholders view the organization’s AI programs. A technologist may be inclined to build in-house. A marketing team may be looking for an easy-to-deploy off-the-self platform. The CFO might calculate that the best bottom line choice is AI-as-a-Service.
It takes leadership to balance these tendencies. All AI stakeholders must operate within a cohesive strategy and governance structure, and decisions are made in the C-suite. As you think about building a powerful AI ecosystem, consider the conditions under which the enterprise can best buy, build or collaborate.
Buying packaged solutions is a suitable way to start building the organization’s AI ecosystem. AI platforms related to back-office functions and automated customer engagement are reaching real maturity and there may not be a need to reinvent those wheels. Consider the organization’s desired outcomes, identify where AI could add value, and research the market for existing solutions.
This can be particularly beneficial for companies that are still at the beginning of their AI journey and may not yet have robust in-house capabilities and talent. Meanwhile, enterprises at all levels of AI maturity could look to co-develop AI applications, using kit-like platforms that can be further customized to suit the business.
Keep in mind that every technology (including AI) available for purchase is unlikely to offer distinctive capabilities as these tools are available to every organization.
Transforming a business to use AI has implications beyond the application. Over time, technologists are hired, staff are retrained, technology investments and replacements are made, and eventually the organization is ready to tackle bolder, more innovative use cases. It follows that organizations with more years of AI experience are more inclined to develop AI solutions in-house.
In this case, collaboration between business units and data science is essential. The business users are best placed to identify needs and opportunities for improvement and transformation. Meanwhile, the data scientists are best positioned to think about what is achievable with existing technology and the kinds of models that could be developed. Together, supported by the right technologies, processes, governance and talent, they can create applications that offer differentiating opportunities in the market.
Of course, not every AI pilot will succeed and make it to production. In addition, continuous innovation and management entail real operational costs. When deciding whether to build or not, weigh the expected results against the actual cost of making the tool that will deliver those results.
Many companies are innovating and exploring AI use cases, and partnering with these organizations can accelerate access to AI capabilities and applications. There are several ways to leverage the ecosystem of organizations developing AI.
A partnership can take the form of working with a service or product company that can develop a desired use case or offer AI-as-a-Service. This can be attractive in part because the ongoing innovation, ongoing management, and scale-on-demand can be managed by a third party. Partnerships can also include making a business investment in a promising startup or acquiring a startup directly to give you access to human capital and intellectual property. Or a company can also find enthusiastic partners in universities and academic institutions where AI is being researched and developed.
When exploring potential partnerships, consider the company’s AI strategy and goals, where the market aligns with the vision, and how collaboration can deliver a return on investment.
Interestingly, a recent questionnaire by Deloitte, where I am an executive director, found that many organizations today take an all-of-the-above approach to AI acquisition. The method of acquiring AI currently appears to be non-linear, with organizations buying early in their journey and building more AI-mature organizations. Instead, because every company struggles with its own needs, goals, and resources, the strategies they employ must necessarily be tailored to the enterprise.
Indeed, there is no one right approach to building a powerful AI ecosystem. There are only the best decisions for the company and leadership plays an essential role in this.