AI is Nascent, New, Hard… and Inevitable

At a recent FORA Roundtable in NYC, produced in association with Genpact and featuring their Chief Digital Officer, Sanjay Srivastava, HFS Research led a discussion with 20 enterprise leaders diving into the what, why and how of AI for business operations. The participants each came with a unique perspective on the challenges and opportunities of AI for their business, and there was a breadth of maturity between those in the “getting started” and the “getting good” phases of AI experimentation and adoption.  It is clear that we at the beginning of a long cycle of learning and understanding how AI will impact business operations – but as the impact of AI will be inevitable, enterprises must start having these conversations, or risk falling far behind. 

The ”why” and “what” of AI have largely been defined… now it’s getting to the hard task of “how”

There is no longer any question about why to bother experimenting with AI. It seems most people feel that the risk of not investigating and experimenting outweighs the risk of trying and failing, and have accepted AI as an inevitable part of business moving forward.  A greater understanding of what AI really is also seems to have also developed.  The most important clarity that most organizations have gained over the last several months is that AI is not some monolithic thing or a singular technology.  Instead, we’ve come to understand AI as a toolkit, or “a bucket of stuff” that enterprises can use to make their operations more intelligent; building blocks that include various elements of foundational AI moving across a spectrum toward more packaged solutions. 

There’s also an emergence of integrated automation (exhibit 1), which demonstrates the power of AI as combined with the other important change agents of RPA and smart analytics.  Many companies lamented that they’ve started to make progress with RPA but not yet dabbled with AI, despite the opportunity to combine forces using the elements of these tools to make operations not just more efficient and effective, but more intelligent and intuitive. But, the distinction between RPA as a script-focused point solution and AI as a business solutions focused tool seems to have gained credence.

Exhibit 1: The Future of AI lies in integrated automation

How will enterprises adopt and industrialize AI? By facing critical challenges head on

• Have change management at the heart of the discussion. Change management is one of the biggest challenges to tackle as cultures shift and change as a result of AI.  According to a recent HFS survey, the majority of AI leaders say that culture is holding them back from effectively implementing AI. Fears about backlash and perceived negative impacts of AI are contributing to AI Fatigue and muddying the waters.  Our roundtable discussion highlighted the importance of a transparent and candid AI strategy, and plenty of change management investment to support it. 

• Put data quality in the spotlight. Data is fundamental to AI– getting better, cleaner data and knowing which data is needed – is critical for any AI program to work well.  Ultimately the quality of data will be the foundation of any AI project.

• Consider ethical implications in-depth. One of the lesser- explored challenges for AI impact is how ethical considerations get factored in, in particular relating to data ownership and being conscious to potential algorithmic bias.   Each of the roundtable participants recognized a need to explore these impacts in much more depth and embed them in AI discussions.    

• Rethink talent strategy and the role of humans. One clear challenge for AI and talent is needing to find individuals to bridge the gap between tech and business.  Despite the pervasive refrain about reskilling, some organizations have yet to embark on any reskilling, for lack of really knowing what skills need to be learned. Some characteristics that have been helpful, such as a natural curiosity and comfort with data, are more cultural or personality traits than skills. 

The diversity and inclusion element is very important to any AI discussion; AI is not just only for the tech savvy cool kids. Many roles will be impacted and we all need to learn (or unlearn old habits) to come to terms with a new reality.  The machine intelligence paradox dictates that the more businesses become reliant on AI, the more important true human skills become, and some human-in-the-loop component must be baked into AI design. 

• Find the right business cases and applications. The how is not just about how to implement AI but how to justify it.  Designing all AI use cases to put human users at the center will not only help to find the right processes but also to reduce the effects of AI fatigue. Companies also need to resist the urge to go too deep and instead go mid-depth to solve business problems.

• Leverage service providers on the journey toward industrializing AI realistically. There’s an important role that service providers can play in helping enterprises to effectively adopt AI and to approach integrated automation more holistically.  Service providers helping clients navigate the AI landscape admittedly tend to spend too much time trying to dazzle clients and prospects with all the cool new possibilities, rather than taking a more grounded approach to applying technology to business problems and making AI digestible.

The Bottom Line:  Making AI relevant to business operations requires a fundamental refresh and rethinking of critical processes and enterprise goals.

No company has found a silver bullet answer for how to approach AI, and the journey looks different for every organization.   We are already seeing use-cases of effective AI in business operations as a result of these integrated solutions, but many questions still loom along the inevitable AI journey.  The impacts of AI are happening — whether we like it or not – and will be significant.  Now is the time to prepare by having thoughtful discussions to rethink the potential opportunities that AI and its fellow change agents represent.