Technological revolutions have become a norm in this era of innovation. Artificial Intelligence offers organizations a competitive advantage. There is also considerable pressure on organizations to go the AI route for fear of losing an edge to competitors.
88 percent of respondents indicated that their company already has, or has plans to implement Artificial Intelligence and Machine Learning technologies within their organization. While, 95 percent indicated it would either complement or make it easier to do their job rather than reducing or making their role harder.
However, with AI technologies only arriving in the workplace now, it is still not clear whether they bring productivity benefits or not, or whether they will have any impact on an organization's revenues.
That is not to say that they won't, or that they don't have the potential to be beneficial after all, any additional business planning or strategies can only be a good thing but rather that organizations need to think about AI before they rush into it.
Following are the ten questions that organizations need to answer:
Do You Really Need AI to Solve This Problem?
Some automation and analytics are simple enough that they can be solved with much simpler procedural code rather than building and maintaining an AI model. Enterprises need to look at what they are trying to do and decide if AI is worth the investment.
How Will AI Improve Your Customer Engagement?
Dr. Rob Walker, Vice President for Decision Management and Analytics at Peg systems, pointed out that that while the hype around AI makes it seems like it can solve all your problems right now, the businesses that drive the fastest and best results will be those that focus in on tackling specific issues that will have the greatest impact on customer experience. Businesses should identify the low-hanging fruit, higher-impact opportunities to transform their brand for optimal customer engagement.
Where AI is proven to make an immediate and quite tangible difference is in customer relevance. Look for channels in which AI can help determine the right message at the right time for any given customer. If cost reduction is an important driver, think about AI-powered chatbots to reduce mundane customer service tasks.
What is the Organization's Business Case?
If AI is deployed simply as a "gimmick" or experiment without solving a business problem, then it can be short-lived and is often abandoned quickly, as leadership will not see any return on investment and people will simply stop using it and the entire technology will be dismissed as "not working for us."
Do You Have the Necessary Data?
This is really the big one, he cautions. Using AI involves being able to train a model on data. If you don't have the data collected then AI will not be as effective. An example would be if want to automatically route documents based on how people have routed similar documents in the past. In this case previously collected data would train an AI to do this. Unless you have the text of the document and who the person routed it to based on this text, it is going to be hard to train a model to do this. So, companies planning on using AI in the coming years need to be thinking about data collection now.
Organizations also need to establish a baseline to measure performance for AI. "The first rule of using AI is a simple one AI is only as good as your data and goals allow it to be,said Katrin Ribant, CSO and co-founder of New York City-based Datorama. Goals are equally as critical. This is because without a robust key performance indicator (KPI) framework, you will quickly find yourself lost in data and complex widgets. Without understanding performance targets, it's impossible to truly understand how to properly optimize your actions to achieve your desired result.
Do You Trust the Data Sources That AI Will Use?
Klaviyo is a Boston based builder of an e-commerce marketing automation platform. Agata Celmerowski, their VP of Marketing, believes one of the key questions that organizations need to answer is whether their data and date sources are suitable for AI. In the ecommerce industry, marketers and brands like need to be thinking about the data they're collecting, how they're storing it, and how they can use it to create a better experience for their customers. Without that focus, they'll be limited in what they can do with AI. It's important to view data as a strategic business advantage and use it to make personalized content.
She adds that 2018 is the year that using accurate, rich, and meaningful data to enhance every customer touch point will stop being optional. From there, the natural progression will be applying increasingly sophisticated techniques to analyze that data and determine the best ways to accelerate growth. There's a lot of hype in Mar-tech space today around things like artificial intelligence, but so far it's been a series of empty promises. Brands that are leveraging data science today are largely forced to do the heavy lifting themselves.
Is Your Data Architecture Suitable?
According to Ribant, while data is important, it is not enough. Organizations need an effective data architecture. They need to build a robust data strategy and ensure they have the right architecture in place. This seems like it would be a given, but in my experience, this is becoming a bigger problem as companies hold onto data just in case, and look to activate it later. With the rise of data lakes, which allows for data to be stored for a relatively modest investment, companies are starting to learn about the hidden cost associated with these data assets.
Can Existing Data Management Systems Support AI?
Can the existing data management systems hold up under the new load of artificial intelligence? AI systems use data as fuel and if it's incomplete, tainted, inaccurate, or biased, then low octane levels will lead to less effective AI. That said, don't wait for the data to be perfect current AI is perfectly capable of determining what data works and what is too unreliable to use.
What Are the Consequences of Getting It Wrong?
Sometimes, what is called AI is really statistics and correlation finding. So sometimes, like humans, it will get this wrong and the amount it will get wrong can vary based on data quality. So, do we want to implement AI in a process with a lot of variability that may have a lower accuracy rate if the times it gets it wrong can have major consequences? If, like with self-driving cars, there is a human there to take over when the AI gets it wrong then that may make the situation workable, but if the accuracy is so low that a human is needed to take over more often than not, then that AI may not be worth implementing.
What Are the Risks?
AI comes in two distinct flavors transparent and opaque, and both have very different uses, applications, and impacts for businesses and users in general, shared Walker. In some instances, businesses will need to be able to explain exactly how they reach certain algorithmic-based decisions. In these cases, impacted organizations will need to employ a transparent form of AI that can provide that ability for the logic to be explained for any given decision.
What impact will this have on workers in organization?
Nepf is a professional services firm based in Los Angeles that helps organizations with their digital transformation projects. Andrea Schnepf is the manager of the company. She pointed out that where AI fails, it is generally because of the people involved in a given project. People don't like any change and with AI there are perceptions that range from fear of losing jobs to ineffectiveness, she said.
Where AI may replace part of a function of an employee's role, impacted employees will typically and naturally respond negatively and with resistance. Sustained adoption of AI requires employees that are brought along the journey from the start. Effective change programs would involve education and targeted interventions to involve employees and their line managers.
Will AI Integrate With Your Current Stack?
AI should not be deployed as a standalone technology, but as an integrated solution as part of a broader process where people, process and AI work together efficiently and maximize productivity, results and outcomes.