We began the October APC meeting by sharing actions members have taken based on presentations and discussions at the June meeting in Cambridge, Mass. The themes of that meeting were managing change (three case studies and an experiential activity) and the workforce of the future (research and projections from Massachusetts technology entrepreneurs). Several APC members re-evaluated their change management practices – in one case reworking a change initiative that was doomed for failure, in others enhancing communication strategies and transparency at all stages of change projects. Many adopted the change tools presented at the meeting. Several members, impressed with Akamai’s Technical Academy, launched a similar program to train a diverse pool of candidates without technology backgrounds to fill technology roles needed by the organization.
The presenters at this October meeting covered various technologies – artificial intelligence (AI), application programming interfaces (APIs), analytics. In all cases, the focus was on leveraging the various technologies to capture the promised business value: speed to market, lower costs, increased revenue, new business models.
One key theme of the meeting that transcended the technologies is that new technologies and uses of them will continue to appear at an increasing rate, but keeping up with technology developments may be the least challenging part of the CIO’s role in the future. Researchers encouraged APC members to sharpen their perspectives and skills in a number of other areas essential for identifying opportunities to leverage technology to gain business value and execute to realize that value.
Perhaps the best example of taking this wider, more holistic view can be found in the Harvard Business Review (HBR) article “Building the AI-Powered Organization” that was sent to APC members before the meeting. The authors report that only 8% of firms in their survey of thousands of executives engage in core practices that support widespread adoption of AI and advanced analytics. The authors then identify practices that facilitate wider adoption. At the meeting, APC members assessed how well their organizations followed these practices, which cover multiple organizational aspects beyond the IT department, such as making the shift, setting up for success, organizing for scale, educating everyone, and reinforcing the change.
The following meeting highlights are organized by key aspects of companies covered by the presenters.
The HBR article “Building the AI-Powered Organization” emphasizes making the shift from experience-based, leader-driven decision making to data-driven decision making at the front line, where people trust the algorithms’ suggestions and feel empowered to make decisions. It also recommends setting up for successby budgeting as much for integration and adoption as for technology (if not more), and balancing feasibility, time investment, and value through a portfolio of initiatives with different time horizons.
Terri Griffith strongly encouraged transitioning from top-down to bottom-up job and task design in order to capture the greatest business benefits from AI and other technologies. She gave many examples of better results through broad participation, such as how algorithms created through crowdsourcing target lung cancer tumors for radiation therapy as well as oncologists and do so 75% to 96% faster. Research has demonstrated how people who design their work can increase their performance and employability. But without support and training, people design terrible jobs. Griffith described methods for helping knowledge workers craft tasks and jobs with good outcomes for the organization and individuals. She believes that the “doom and gloom” scenarios about automation are more likely if we only work from the top-down.
The HBR article “Building the AI-Powered Organization” also emphasizes organizing for scale by (1) developing hubs (including data governance, AI recruiting and training strategy, and work with third-party providers of data and AI services and software); (2) creating spokes (including tasks related to AI adoption such as end-user training, workflow redesign, incentive programs, performance management, and impact tracking); (3) developing a governing coalition of business, IT, and analytics leaders; and (4) deploying assignment-based interdisciplinary execution teams within the spokes.
Nigel Melville and Rajiv Kohli offered several structural recommendations from their research on how firms can effectively adopt and implement APIs for faster speed to market, lower costs, and increased revenue. In order to deal with the data integrity issues that arise with API implementation, they recommend multifunctional data governance teams that can align with standard frameworks such as AIC and COBIT. In order to deal with tensions between producers of APIs and those who use them, they recommend developing new roles and responsibilities such as an API product manager, enforcing service level agreements, and using analytics on data usage to inform sunset and extend decisions.
Phil Weinzimer also weighed in on governance structures needed to identify and prevent unnecessary risk while also ensuring that business outcomes are achieved on all technology projects. His research with 1,500 business executives concluded that lack of good governance was a major obstacle to achieving better customer experience, improved operational processes, and enhanced business models. He recommended moving from governance 1.0 (reactive) to governance 2.0 (proactive) and specified various components of governance 2.0, drilling down on each of the six key components.
Melville and Kohli discovered in their research that many people in the organizations they studied were not familiar with APIs and even less familiar with their value to their organizations. So an obvious recommendation is to educate employees on what APIs are and how they can benefit the organization and, where possible, how the specific employees can benefit. They identified the dilemma executives face when people are willing to help develop APIs that can benefit them directly, but may not feel incented to develop APIs for use by other business units. Melville and Kohli suggest demonstrating value through pilots. Of course that doesn’t quite get at the issue of developing APIs for other business units.
Lynne Markus raised the impending issue of retaining people with knowledge once analytics and AI augment their roles – and possibly take away those roles. She warned against a downward spiral in which experts become deskilled and leave; and what’s left of their jobs is taken up by non-experts, who are hired at lower costs. As automation advances, jobs are further deskilled.
The HBR article “Building the AI-Powered Organization” highlights the importance of making the shift from a rigid and risk-averse culture to an agile, experimental and adaptable one characterized by a test-and-learn mentality where mistakes are seen as a source of discovery.
And Markus reminded us that business partners are a source of risk. She cautioned that we should beware of a turning a blind eye when delegating activities to partners or other vendors who build and maintain algorithms for us.
 Get Your Digital Priorities Right (M. Lynne Markus); Leveraging Automation from the Bottom up (Terri Griffith); Effective Adoption and Implementation of APIs (Nigel Melville and Rajiv Kohli); Strategic IT Governance 2.0 (Phil Weinzimer)