Realize the potential of AI: Enablers

We are ready to say goodbye to 2018, the year we saw “the impact of Artificial Intelligent on humanity” debates taking place at every level. AI is already becoming a commodity and its potential positive (and negative) effect is not a hidden secret anymore. AI has seen widespread adoption by multiple industries. At the same time the rapid advancements, especially in image recognition and autonomous actions, surveillance, and concerns around bias in decision-making, triggered actions from governments, research institutes and regulatory institutions. Never before we have seen so many data privacy breaches, and testimonies by corporate CEOs. While there is enough press on negative effects of AI, there is no denying the fact that AI has transformed many early adopters, making it more urgent than ever for others to accelerate their digital transformations. Gartner says by 2020, AI will create more jobs than it eliminates.

In the coming year we expect to see much more scrutiny and audits on AI implementation from regulatory authorities, but the rate of adoption by companies will only go up. There is good news though. Unlike the early adopters who had to learn the hard way many elements of AI - Machine Learning, Natural Language Process etc., the companies will now find making the machines intelligent is surprisingly easy. There is ever-increasing plug-and-play solutions from the technology giants - Amazon’s SageMaker, open-source machine learning framework of Google’s TensorFlow and Microsoft’s Azure Machine Learning.

I have explained select use cases and foundational elements companies should possess to advance those uses cases in previous segments of this series. History has shown us that transformation is not easy. Not every company is successful in getting the vision and organizational alignment right. If you study early adopters who have successfully transformed themselves leveraging AI, you will notice certain distinct characteristics. They had C-Level support, they were ready to experiment, not afraid of failures, focused on growth rather than on savings, targeted core activities for AI adoption and open for investment in multiple enabling technologies, like IoT, Cloud etc.

There is certainly no single standardized approach to getting started on enabling an organization for AI led transformation. However, there are proven approaches and perspectives that can be adopted by the companies, who are planning for AI led transformation.

Leadership – set the ambition at executive level. Leaders should view AI as a strategic advantage and opportunity for competitive differentiation, potentially entailing new revenue streams and capabilities. Involvement at this level provide them with a clear understanding of the impact of AI will have on their financial and business models, and they will ensure stronger alignment between investments and corporate strategy. They will get firsthand information on related impacts and potential consequences.

Create blue print for organizational transformation. The blueprint should clearly communicate organizational strategy, KPIs and priorities to the entire organization, and must help define the roadmap to achieve the outcome. AI initiatives can happen throughout the organization and can be effected iteratively, but the blue print ensures that those initiatives stay in alignment with the corporate strategy.

Identify use cases. One should know what business problem AI is trying to solve. Looking at artificial intelligence as a magic ingredient that makes all things better will lead to costly mistakes. There should be collaborative sessions across business units, hackathons etc. to document real business problems that are candidates for AI intervention. Once identified, prioritize and focus on limited number of relevant use cases – sustainable success is possible by focusing on limited number use cases that add the most value, instead of trying to target as many use cases as possible.

Portfolio driven. The identified use cases will translate into a portfolio of projects, some could be tagged as short term - for those proven solutions are available through partners or can be developed in house with relatively less effort and some as long term requires substantial investment and partnership with external entities.

Build an ecosystem of technology partners – cloud, analytics, ML etc. AI is an ecosystem play. Building AI capability will require working relationship with wide array external partners and institutions. Companies should identify which aspects of AI implementation and monetization efforts they should own and which are best addressed through partners. This will also drive decision on data ownership, cybersecurity and privacy considerations.

Strong internal team with agile mindset. To completely exploit value from AI, companies need to build strong internal capabilities. It helps to establish self-organizing and managing dedicated cross functional teams. The team should be equipped with range of skills – business representatives, analytics, user experience (UX) design experts, data engineers, and DevOps, infrastructure, security and cloud architects. Organizations should consider adopting to scaled agile frame work (SAFe) or something similar that provides the agility at scale for program management.

Data governance strategy. Having a good data governance strategy is critical for AI led transformation efforts. Typically, data is distributed and in silos in most of the organizations. It might have several ownership. Data is the oil for digital transformation. Its fluidity makes it even harder to protect. Careful upfront consideration of who owns which data, what rights partners and other external entities involved in the program would have, what would be the data sharing mechanism without violating any privacy regulations, is the data going to be retained in-house or transported to cloud service provider for analytics, train the model etc. should be established to prevent issues that might potentially derail the program.

Cybersecurity measures should be interwoven throughout the program. AI related initiatives brings together diverse set of data, information sharing, infrastructure integration with external entities like cloud providers, and complex algorithms. Having cybersecurity architects as part of the cross-functional agile teams is much more critical than ever before.

Training and reskilling. The workforce needs to be reskilled to exploit AI rather than compete with it. Humans need some time to adjust with the new way of working along with machines. Some AI initiatives might change business processes and potentially might automate certain activities. Employee concerns should be addressed upfront, as human-machine collaboration is expected for successful AI led transformation of the company. Some workers will have to retained, while others will have to be redeployed within the company, leadership within the organization have to play this vital role in aiding these transitions.

Image Source: PwC 2019 AI Predictions

Studies by reputed firms and industry analysts show, adoption of AI will continue to accelerate, enabling companies to anticipate demand, optimize business operations and R&D, reduce production costs at improved quality, and improve user experience and revenue. 

The mentioned enablers in this final part of my AI series were instrumental in driving value from AI at scale in those successful transformations. Companies who want to ride the AI wave should adopt these learning and quickly establish these enablers.