Some organizations have already identified the benefits that can be gained from Artificial Intelligence and Data Science, bringing in talented resources to enable them to build AI models and solutions. But more often than not, the business doesn’t understand the capabilities and huge potential of AI well enough, nor the investments that are required in tooling and people to make the benefits sustainable and long-term. This gap in understanding often leads to tactical exercises and proofs-of-concept that end up going nowhere.
In this Article, Andrew Burgess (Strategic Adviser on AI – Founder and CEO of Greenhouse Intelligence) explains how the data scientists and the business executives can work together to maximize the value of AI.
How do you talk AI to a business person?
1. Explain AI in simple terms
Business people usually don’t want to hear about algorithms or python libraries and it is highly unlikely that they will understand these aspects, but they will still want to understand what AI does and what it means for their business. It is important to explain the capabilities of AI in simple terms and focus on what AI can do, rather on how it works. To explain AI, Andrew uses a framework of 8 AI capabilities that are split into three main areas: capturing information, explaining what is happening, and finally explaining why it is happening.
2. Focus on the positives…
AI can bring huge value to an organization and it is important that the business understands the full range of benefits available. These include customer satisfaction, risk mitigation, cost reduction, loss mitigation, revenue leakage mitigation, and revenue generation. These terms will definitely resonate with the business people especially when they align with their strategic goals.
3. …without getting carried away…
Don’t over promise and under deliver. Understanding the business’s AI maturity is key in defining where the starting point is, and will help build a successful AI journey. A great way of assessing a business’s AI maturity is to talk to the different business lines and understand where they want to be in terms of implementing an AI strategy: their ‘AI ambition’. The gap between the current situation and their ambition is the scope for AI. Setting the expectations for what AI can do will avoid the teams getting over-excitement and help achieve realistic goals.
4. …and without ignoring the risks
There are a few risks specific to AI and it’s important to keep them in mind. Problems like Black Box models, bias, AI’s naivety, over-excitement… can all happen. The business people have to understand that there are risks associated with an AI journey, but that there are also solutions to overcome them.
5. Get a strategy
Getting an AI strategy in place is important to define what the business challenges are within the business, as well as the objectives, and how AI can help deliver them. It is the tool that brings the business and the data teams together. When creating an AI strategy, it is crucial to identify the opportunities for AI that align with the business’s objectives and/or can solve their specific challenges.
6. Think big but act small
Having a strategy and a vision is key, but to ensure you get there, it is important to start with things that are very practical and achievable in a short period of time. Try and identify the activities that are going to give the most value and are easiest to implement. Start with proofs-of-concept from the highest priority opportunities in the list, but also work on data cleansing, change management, and education.
7. Look at all the options
Many business people become paralyzed when they look at the plethora of AI vendors out there, but vendors are only one option, and should only be used when the requirement exactly matches their standard capabilities. At the other end of the scale, where you have a unique data set and/or a major project, then you may need to code from scratch, but most opportunities will likely involve the use and configuration of data science platforms.
8. Keep measuring and checking
How do you get continuous improvement and optimization on all the activities you are doing? As your program is starting to build momentum, make sure that you’ve got regular meetings with the program team to monitor progress. Demonstrate a clear return on investment and have simple dashboards that display what the benefits are.
By following these eight points, data scientists, analysts, and developers can engage effectively with the business to deliver AI projects that are valuable and beneficial to the whole organization.
Written by: Eve-Anne Tréhin
Originally posted on: H2O.ai Blog