Artificial Intelligence (AI) is clearly a must-have when it comes to being competitive in today’s markets. But implementing this technology has been challenging, even for some of the world’s top companies. There are issues with data, finding the right talent and creating models that generate sufficient ROI.
As a result, many AI projects fail. According to IDC, only abut 35% of organizations succeed in getting models into production successfully.
“While we see AI technologies performing a swath of incredible feats such as Google Translate, AlphaGo, and solving a rubik’s cube, it can be hard to tell which business problems AI is apt to solve,” said Ankur Goyal, who is the CEO at Impira. “This has led to a lot of confusion—and a vendor community that has taken advantage of it by labeling things as AI when they aren’t. It’s very reminiscent of early last decade when cloud technologies took off and we had a lot of cloud washing going on. We had vendors marketing themselves as cloud players when their offerings were vaporware. Similarly, we are going through a period of AI washing now.”
So then, if your company is thinking of implementing AI, what is the best way to start? How can you help boost the odds of success and avoid the pitfalls?
Here’s a look at some strategies:
Beware Of The Hype
AI is not magic. It will not solve all your company’s problems. Rather, you need to take a realistic approach to the technology.
“Unlike traditional data analytics, machine learning (ML) models that power AI are not always going to offer clear-cut answers,” said Santiago Giraldo, who is the Senior Product Marketing Manager of Data Engineering at Cloudera. “Implementing AI into the business requires experimentation and an understanding that not every experiment is going to drive ROI. When an AI project is successful, it is often built on top of many failed data science experiments. Taking a portfolio approach to ML and AI enables greater longevity in projects and the ability to build on the successes more effectively in the future.”
Interestingly enough, there are often situations when the technology is really just overkill!
“Often times businesses take on AI projects not realizing that it might have been cheaper to continue a process manually instead of investing large amounts of time and money into building a system that doesn’t save the company time or money,” said Gus Walker, who is the Senior Director of Product Management at Veritone.
You don’t want to spend time and money on a project and then realize there are legal or compliance restrictions. This could easily mean having to abandon the effort.
“First, customer data should not be used without permission,” said Debu Chatterjee, who is the senior director of platform AI engineering at ServiceNow. “Secondly, bias from data should be mitigated. Any model which is a black box and cannot be tested through APIs for bias should be avoided. The risk of bias is present in nearly any AI model, even in an algorithmic decision, regardless of whether the algorithm was learned from data or written by humans.”
Identify the Problem To Be Solved
In the early phases of an AI project, there should be lots of brainstorming. This should also involve a cross-section of people in the organization, which will help with buy-in. The goal is to identify a business problem to be solved.
“For many companies, the problem is that they start with a need for technology, and not with an actual business need,” said Colin Priest, who is the VP of AI Strategy at DataRobot. “It reminds me of this famous quote from Steve Jobs, ‘You’ve got to start with the customer experience and work backwards to the technology. You can’t start with the technology and try to figure out where you’re going to sell it.’”
The problem to be solved should also be specific–that is, something that can be measured–and narrow. Don’t boil the ocean.
“It is the small steps that count,” said Mike Brooks, who is the Senior Director of APM Consulting at Aspen Technology. “Do not make the mistake of trying to make AI work for everything, all at once. After analyzing value for each AI initiative, the real benefits come when it solves a very specific goal.”
While it is important to estimate the ROI of a project, there is often too little attention paid to the cost side of the equation. But this can lead to disappointment. After all, it is never fun to be over budget on a corporate initiative.
“Companies looking to implement an AI project should start by looking at the cost of the operation and doing an analysis on how that cost structure compares to best practices,” said Jerry Kurtz, who is the EVP and Head of I&D at Capgemini North America. “The cost of storing and transforming data is typically 70% of the budget, and only brings 10% of the value. Being able to leverage AI to solve business problems is only 30% of the cost, and brings 90% of the value. If an organization can reduce data costs and improve data quality, they’ll have more budget to put toward leveraging AI to solve those business problems, like improving productivity and efficiency.”
Implementing AI can be wrenching for an organization. Employees may be skeptical of the technology and could fear for their jobs.
This is why there needs to be focus on getting buy-in, which means having clear communication of the benefits. It should also involve a commitment from the C-Suite. Consider that—according to a recent survey from O’Reilly—the biggest bottleneck for AI is an unsupportive culture.
“For AI to succeed you must have the buy-in of your workforce and the right employee upskilling programs,” said Anand Rao, who is the Global AI Lead at PwC. “You can’t simply offer AI training courses to employees; you need to go further and offer both immediate opportunities and incentives to apply what they’ve learned. Furthermore, business stakeholders and end-users—not just the tech staff—need to be included from the beginning of any project. If they’re not brought in at the start, your organization risks building a solution that does not work for the people who will be using it.”