Part 6 in the “RPA and AI in Finance” series, which examines the role that robotic process automation and artificial intelligence can play in finance operations.
Are you starting to consider using artificial intelligence (AI) to solve some of your business pains?
That’s probably a clever idea, but know that these projects must be carefully planned. You also need an approach to building in-house competencies, as opposed to getting outside help. In Part 5 of this series, we described how to solve business pains with AI as opposed to just starting random projects simply because you must have AI.
This week we’ll go into more detail about what needs to be in place before you start on your AI journey. Specifically, for the finance function, we’ll discuss what you can and can’t do yourself.
It’s time to look at step two in our five-step model for succeeding with AI:
Think in terms of specific projects instead of general knowledge
Most people these days have general knowledge about AI. At least, we can start to imagine what it can do for us, and that’s why we feel we must have it. However, general knowledge won’t get you far as you start to be specific about what business pains to solve and how to go about it.
The challenge: We often meet business leaders who think they need general in-house AI-competencies: “We need to build an AI resource center with skilled data scientists.” Now, this might be a promising idea for some organizations. Amazon would not be where it is today, for example, without focusing on data science as an in-house resource. However, for most companies, that’s probably not the way to go.
Consider this analogy: Many businesses have a fleet of company cars. However, we’ve never seen a company that insists on having an in-house car manufacturing facility, or even its own lease agency. They buy from car makers or lease from agencies.
Similarly, you don’t need to build your own AI solutions (unless you are the new Amazon, of course).
The fix: Start by solving specific AI-suitable business pains by using external experts or available business-centric solutions. Cash in the benefits, business-wise and in terms of improved understanding of AI. Then solve the next business pain and cash in. Continue until it’s evident that you will benefit from building an internal competency center.
You need to prove that it works at a large scale before it makes sense to build up competencies internally. Once you’ve proven the value, there’s also money to invest in building competencies if it makes sense.
Off-the-shelf products can get you far
In the finance function, we’ve observed companies that simply took off-the-shelf solutions from the Internet or from vendors and applied them to their operations. One example saw a finance function use one of these solutions to replace all human forecasting in its commercial finance function. Instead of humans doing the forecasting, and corporate arguing whether it would hold true or not, countries and markets now had to argue why the AI-produced forecast would not hold true. In most cases, the AI forecast proved much more accurate than the human forecast.
This is not surprising, as we know human forecasting is biased. We’ve seen this done elsewhere, too, using in-house competencies. We’ve also seen the first chatbots in finance able to answer simple queries rather than doing manual lookup.
AI is coming, and it doesn’t have to be complicated. But don’t start by building a huge in-house center of excellence. Start with specific projects and prove that it works—then we’ll talk.
This article originally appeared on LinkedIn and is republished by permission.