The AI mindset: preparing people is as important as preparing data

As companies work to understand the techniques and tools used for simulating cognitive functions in machines, they often overlook a critical aspect of preparedness

Businesses of every size, stripe, and sector are struggling to prepare for artificial intelligence. As companies work to understand the techniques and tools used for simulating cognitive functions in machines, they often overlook a critical aspect of preparedness.

Preparing people is as important as preparing data

While data is the requisite asset, people are the ones building, measuring, consuming, and determining the success of AI in enterprise and consumer settings. People are also the ones whose jobs will change; whose tedium will become automated; who will consume or reject the outcomes of AI; and who will feel its myriad impacts. This means that AI readiness requires businesses invest in the cultural and mental constructs of AI across leadership, employees, even users.

Prepare for the AI mindset

AI isn’t just about data scientists and machine learning algorithms; as we deploy software and machines to handle more and more cognitive tasks, it is about rethinking how we do things and why. This, ironically, requires a level of introspection on the part of the individuals designing, building, and using the technology, as well as the broader organization’s ultimate goals for the technology. Regardless of whether employees are engineers, agents, executives, or field staff, the idea is planting the seed of a new mindset, an AI mindset.

The following seeds or pillars for such a mindset were developed as the result of research interviews with dozens and dozens of people involved in building out AI programs in enterprise settings. They reflect three universal truths about AI, which serve as starting points for effectively building people’s understanding, engagement, and role in an organization’s AI journey.

1. AI diversified

AI must be built by people who understand the business and domain problem (not solely the technology). The gravity of some applications (e.g., credit scoring or medical treatment recommendations) demand diverse perspectives, multidisciplinary expertise, and workflows to monitor domain dynamics. It takes a village.

2. AI directional

AI is always learning and never complete; AI is probabilistic, is subject to errors, and will always require critical thinking on the part of people. Perfection is the enemy of progress and employees are instrumental in designing and shaping AI systems to balance the benefits, risks, and auditability of the application.

3. AI democratized

AI isn’t just for data scientists and developers; thanks to relatively easy-to-use cloud-based tools, harnessing the technology does not require advanced degrees in computer science. Organizations can further enable democratization through skills development, programs for learning and sharing ideas and best practices, and investing in tools with intuitive and customizable user interfaces.

Readying people for the AI mindset takes on different forms for different organizations but should span broader AI strategies as well as more vertical applications. From a strategic point of view, it’s essential to convey the intention and long-term vision for AI programs, including (but not limited to):

  • How and why AI transformation is different from digital transformation.
  • How user journey analyses must inform AI design and initiatives (internal, B2B, B2C).
  • How efforts must be driven from the bottom-up but empowered top-down.
  • Inherent limitations to narrow AI applications.
  • How the organization will support learning and sharing over time.

Simultaneously, it’s essential to ready the minds of those involved in specific AI applications. For example:

  • Use before and afters to demonstrate both the value and the risks of AI applied in this specific use case.
  • Build mental maps with examples people understand and can apply (e.g., a Netflix recommendation machine for your content marketers).
  • Train teams to think in a way that reflects their objectives for using the tool (e.g., think about what would be useful and appropriate if you had a tool to reduce redundant workflows, surface past occurrences, identify trends, predictions, etc.).
  • Deploy programs to augment employees’ professional development; skills and new roles to support and manage AI will come in many colors and aren’t just about math. Core skills for the augmented workforce include skills agility, learnability, creativity, and emotional intelligence, according to research by the World Economic Forum.
  • Train both teams and AI systems themselves to better communicate about the use of AI in the application. Companies should not obscure AI recommendations; empower agents—both human and AI-based!—with the language or designs needed for transparency.

Starting upstream with this kind of guidance can pay dividends later, too. For example, when developing dashboards or self-service portals for employees to extract insights or access trusted data, it’s essential that individuals have confidence in the tools, that the tools are designed to meet their efficiently, and that they understand the role their interactions have in the broader picture.

While thinking about such a mathematically complex technology through the lens of mental readiness may sound too squishy or too abstract, company after company interviewed reinforced the importance of designing AI for people. As the digital world grows evermore automated, organizations have a unique role—not just in extending AI’s practical and commercial applications, but in fostering mental constructs to define our relationship with it.

Copyright © 2018 IDG Communications, Inc.

How to choose a low-code development platform