An approach to assess and implement artificial intelligence

10 July 2019 Consultancy.eu

With the popularity of Artificial Intelligence growing fast, organisations all over the world are exploring use cases to unlock the technology’s promise. Getting AI right is however proving a challenge for many organisations, with over half of initiatives said to be underperforming against objectives. According to a newly launched AI implementation methodology by Deloitte, agility, transparency and a holistic business strategy are key to making the most of the potential.

Recent years have seen global technology take huge strides forward, particularly in the world of Artificial Intelligence (AI). Tech giants like Alphabet and Amazon are constantly cooking up innovative new applications in their labs which are already changing the world of work as we know it. Beyond the now seemingly simple automated mechanisms which can save time and resources on menial work, algorithms are also being birthed which can teach themselves winning game strategies, recognise human emotions or even mimic a human conversation.

While it is hard to overstate the potential of AI, however, that does not mean it is impossible to do so. Any conversation about AI must still factor in the many AI initiatives which get stuck in proof of concepts and pilots, making a lot of companies struggle with the question where to start and how to scale. At the same time, the rate of technological development has often outpaced regulation and public scrutiny, meaning an ethics deficit hangs over the production of AI. Businesses looking to leverage the miraculous technology subsequently find its implementation easier said than done.

An approach to assess and implement artificial intelligence

Naser Bakhshi, a Senior Manager in Artificial Intelligence at Deloitte, explained, “A lack of a sound vision and right prioritisation causes a lot of AI projects to stall. Many initiatives start with a technology that sounds cool, without thinking how it really can make an impact on the organisational goals. They should start with forming a vision on AI that is aligned with the company’s strategy, rather than just letting the one that shouts the loudest experiment freely.”

Tapping into demand for AI implementation solutions from its clients, then, Deloitte has developed a methodology to facilitate the discussion on ‘where to play’ and ‘how to win’ with AI. According to Bakshi, this sees the Big Four firm take a company’s strategy and long term vision as a starting point, before exploring the goals and aspirations and where AI can actually make an impact. This is determined by avoiding perceiving an AI initiative in isolation, but rather using customer-centricity as a strategic theme, considering ultimately how the changes will benefit consumers and attract more of them.

One example Bakhshi provided regards one of Deloitte’s clients acting in a highly competitive market, where margins were under pressure. Using the methodology, Deloitte recommended that they should focus on automating their processes, leveraging AI and Robotic Process Automation. In comparison, a highly specialised firm depending on high-end expert knowledge would benefit most “from an AI-powered expert system that can process many unstructured documents.”

Ultimately, Deloitte’s AI Value Assessment (AIVA) follows a structured three step approach to test if generated ideas are desirable, feasible and viable for execution. Bakhshi surmised that it boils down to three simple questions. “‘Do we want this?’, ‘can we build it?’ and ‘does it make sense?’,” which by answering, companies can develop a thorough understanding of the exact cases that could benefit from AI and the related value.

Agility and transparency

Deloitte has also embedded key elements of agile methodology in its AI approach to ensure value based, relevant prioritisation of tangible deliverables. Deloitte’s Asset Light approach leverages its partnerships with big tech firms like Amazon, Google or IBM, as well as working closely with various niche players in the field of AI to provide solutions to companies which help minimise the risk of losing money on long term commitments (e.g. licenses, hardware), while exploring different AI platforms and technologies before making a final decision for a preferred technology which can be a (private) cloud, on-prem or an hybrid model.

“A lack of a sound vision and right prioritisation causes a lot of AI projects to stall…. Making sure AI makes an impact on the organisational goals is paramount.”

Bakhshi elaborated, “Deloitte’s Asset Light approach allows for the client to utilise Deloitte’s assets as long as they need to postpone making large investments in tooling, hardware or people, until they are certain about the solution… Each phase ends with a clear cut-off point and a go/no-go decision for the next phase, allowing for the client to assess if they want to continue development. This way you can not only scale fast, but also fail fast if the idea turns out to be not as successful as originally thought. This is part of the job when working on extreme innovations.”

Hardware, software, corporate culture and scaling mean that companies have plenty to take into account during their transformation, then. That is not all, however, and firms should be careful this does not lead to an oversight on many businesses’ parts when it comes to the ethics of their AI solutions. According to Stefan van Duin, who is an expert in developing AI solutions, it is therefore easy to understand just why the public is therefore so anxious about AI.

“Getting the right inspiration is an important prerequisite in these discussions”, Van Duin, a Partner in Analytics and Cognitive at Deloitte expanded. “Often AI stays very abstract. People may have expectations that the current technology just can’t deliver, or they think too much in small incremental steps… The more we are going to apply AI in business and society, the more it will impact people in their daily lives – potentially even in life or death decisions like diagnosing illnesses, or the choices a self-driving car makes in complex traffic situations. This calls for high levels of transparency and responsibility.”

Van Duin added that Deloitte is committed to transparency and responsibility in AI, and that the technology’s use must always be explainable to employees and customers. This presents a challenge, though, as AI is not transparent by nature, and has a habit of blurring the lines of accountability for decision making.

The Deloitte Partner concluded, “So, the question is: How can we make AI as transparent as possible? How can we explain how an AI-based decision was made, what that decision was based on, and why it was taken the way it was taken? Transparent AI makes our underlying values explicit, and encourages companies to take responsibility for AI-based decisions. Responsible AI is AI that has all the ethical considerations in place and is aligned with the core principles of the company.”