Does Artificial Intelligence have a place in business processes? I have some business rules in my BPM system – does that count as Artificial Intelligence? What exactly is Artificial Intelligence? These are some of the questions that come up when a business process owner thinks about Artificial Intelligence. While there are some well agreed upon definitions about what Artificial Intelligence is, when it comes down to the practicality of business processes, the definition is contextual and a little subjective.
Process Automation: Types of Logic
When we talk about Process Automation, there are few different types of logic that can be used (this is not an exhaustive list):
- Decisions in a process
- Predictive models
- Neural Nets
- Genetic Algorithms
- Genetic Programs
The list above could be broken down into deterministic and non-deterministic. A decision or a predictive model can be classified as deterministic – this means that if you know all the input, a human being can use pen and paper to figure out the output. Neural Nets, Genetic Algorithms and Genetic Programs are considered non-deterministic – if you know all the inputs, a human being cannot figure out the output.
Most large companies at this point have implemented decision logic into their business processes and many are well on their way to using predictive models. The thought leaders are looking to differentiate themselves by using the non-deterministic variety. This is a little scary for many of the large organizations where even a fraction of a percent here or there could mean millions of dollars – it is a little scary to let an algorithm that you cannot predict make decisions on your behalf. There is also a deeper reason why people are scared of the non-deterministic (what some might call Artificial Intelligence) algorithms – they try to replicate (and replace?) human decision making.
Whether Artificial Intelligence makes sense for you from a business risk or a philosophical level is something that only you can answer. However, without using Artificial Intelligence, you are certainly limiting how far process automation can go. A lot of Process Automation stops because a human is needed to make the next decision. With Artificial Intelligence, a machine can make the decision for the human and there could be more automation.
Valuable BPM-specific information like this should be shared, no? Share this with your friends and peers. Follow Princeton Blue on Facebook, LinkedIn, Twitter & Google+ to get focused BPM-specific updates.