Part 8 in the “RPA and AI in Finance“ series, which examines the role that robotic process automation and artificial intelligence can play in finance operations.
Have you ever worked on a project where consultants were involved? Who worked with the solution: you or the consultants? It was probably you. So why is it that so often when we bring in consultants, we don’t have enough of our own people involved? This is not to say you shouldn’t bring in consultants – often they bring competencies you don’t have – but more to point out that you need to own the solution.
That also goes for projects involving artificial intelligence (AI). You’re the customer of the solution, and if it doesn’t work for you, you’re the one with the problems. Make sure you’re working with the right partners on such a project, otherwise, it spells trouble ahead! This challenge is what we’re addressing in Step 4 of our model for how to work successfully with AI.
How to choose the right partner? Most companies have processes for choosing vendors to work on projects. Typically, they issue an RFP (request for proposal), and vendors then submit their suggestions. But what do you do from there?
4. Reference-check your business partners – the right way
Challenge: There is an enormous difference between academic knowledge about AI and authentic experience. We would say we know how to plow a field with a plow – from a theoretical perspective. But put most of us on a tractor, and we would fail disgracefully. In the same way: It’s easy to study AI and get a degree – even a PhD. But applying AI on a real-life case, and real-life smelly and dirty data, is a whole other ballgame. So just because your business partner is a big name, it’s no guarantee that you will succeed in your quest.
Fix: Use the “Thomas Schultz” 10-step AI-test on a napkin as a sanity check on the claims of your business partner. And ask for real-life, operationally implemented cases and references that are providing the client with true business value. (Oh – and check up on those references as well.)
Most partners can show you the fancy slides, but they have few successes to show for it in real life. That’s not because they’re incompetent or it’s easy to work with AI. No, it’s because it’s hard and requires a lot from all parties involved in the project. Just because you bring on a partner on your AI project doesn’t relieve you of your responsibilities. No, you must take ownership of the project, run the meetings, understand the solution design, and implement the product. You. Not the partner. Are we clear now?
What does this mean for finance?
It means two things:
1. You must have project management capabilities.
2. You know something about AI and how it could potentially impact the solution (see the Thomas Schultz 10-step process for testing an AI solution).
If you’re clueless about what you’re trying to achieve, you will fail. You’d be surprised how often companies are rather clueless about the purpose of what they’re doing. That won’t happen to you, though, and certainly not in your next AI project. Now you know how to select a good business partner, and you’re ready to embark on your project! Have you figured out what that project should be yet? If you have, let us know by emailing us: Anders Liu-Lindberg and Thomas Schultz. If not, go back to Step 1 in the model, because we’re confident you’re not without business pains for AI to solve.
This article originally appeared on LinkedIn and is republished by permission.