Unless you have been living under a rock for the past few years, you must have heard of the hype around AI (artificial intelligence) and ML (machine learning). These are among the most popular terms within the technology industry these days. Companies claim to apply AI in their product development, people say they have ML expertise. Hype may come with reality, but AI/ML has no shortage of challenges. 

AI is not as easy to sell as it may sound on the media. Most buyers do not understand it and are not prepared for the work and uncertainty involved. Some don’t even understand it enough to know what to ask for. You have to take into consideration that there are trust issues, since it is hard to evaluate if someone can do it. 

The presentation statistics on AI is a bit unclear. Many organizations feel compelled to say they are using AI or plan to do so, just for marketing purposes. Most who think they are using AI are in fact just using mock AI, for example, a rules-based chatbot. We can see many companies start trying to do AI, and fail and then just add a few old fashioned methods and continue pretending they are doing advanced work. When talking about a new technology like this, in a 100,000 employee enterprise, surely someone somewhere is using everything. Very few firms, including enterprises, are actually using AI as a core part of their business. 

Just like predictive modeling 30 years ago, those that are advanced can have a key competitive advantage using AI. We can expect to see a lot of disruption, but these would be from a small number of leaders. You can expect that some businesses who use true AI will survive beyond a few years. For example, in one industry that we are working on, a service that costs $50 in work from a team of advanced domain specialists can be replicated for about 50 cents with machine learning. Once that is in place, all the non-AI players will be completely incapable of competing. 

Most firms do not hire others to do AI. At East Agile, AI/ML work remains internal and/or unpaid. It reminds us of the time we did mobile 10 years ago. But we can clearly see more upcoming inquiries about this. A lot of our clients are currently working on AI, but they prefer to do it in-house. The expectation is that it's cheap, fast and easy, and those who do not understand it will probably outsource it.

The decision to spend a hundred thousand dollars on engineering or half a million on mining data is definitely not an easy one. Those who get it generally already have a team and do not need to outsource it, for others it’s core to their business IP so it is not something they outsource quickly. But this will evolve in the very near future and outsourcing will be possible for more mature AI teams. For followers, outsourcing is hard because there are often no immediately visible results or a proof of concept. It is often impossible for non-experts to differentiate total scams from really good people who are making progress. Standard app development is much easier to observe both in quality standards as well as progress. 

All in all, it is important to understand that most challenges of AI projects come down to human factors. The reality is that we all need to find a way to make these relationships work as enterprises and outsourcing companies will need each other. 

Questions? Comments? Concerns? Contact us for more information. We’ll quickly get back to you with the information you need.