A quick guide to validate .AI value

Finally, you got the grip on the jargon of the industrial IoT and (like me) did learn fast, after proving some technology concepts, that "plug & play" with scalability is harsher than expected the next tsunami has arrived,... artificial intelligence (AI). However, many of us firstly expect it applies only to other industries and the hype will pass by, now it becomes clear that most of the future products will have some kind of embedded intelligence or ML toolset to differentiate from other commodities.

The good news is that the AI/ML tools are rapidly becoming more development-friendly and eventually turn into commodities. Just like today, there are tons of No > Low Code IoT development platforms, thousands if not millions of standard IoT temperature sensors the AI/ML will end up at the same stage...someday!

But what if your business does not wait for you to be ready, your key competitor is the first mover, new ventures start to eat the cash cow that today is your comfort zone?

The answer is obviously simple and equally complex when you are driving the map of establishing (or perhaps for the first-time) a eco partner community in a new world loaded by the.AI hype, black boxes, and startups and many other uncertainties. To help out a little bit I have a quick guide that you can use to evaluate an .AI solution;

The first ask would be pretty straight forward by why the solution is and will continue to be superior to alternatives. A solution does not naturally become more valuable just because it uses AI. It could make it even worse due to the complexity by none explainable AI or other data solutions with the same outcome that do not use AI. There must be some reason why AI is advantageous and there are some basics to be evaluated;

• Speed can be an advantage when time is of the essence like Assets that are mission-critical applications in the O&G sector. In healthcare where stroke detection technology, every minute equates to a loss of millions of brain cells, so an algorithm analyzing an image faster than a human makes a meaningful difference.
• The cost can also be a driver. For example, EDGE / embedded AI analytics that resolves significantly latency and throughput of high volume time series asset data. The reduction of the data close to the Asset eliminates sending and storing useless machine data in the cloud and enables real-time distributed applications limiting the need of complex systems and turns into significant benefits.
• Accuracy matters, especially for tasks that are monotone and boring to humans, labor-intensive workflows like preventive Asset vibration and routine planned asset inspections. In case asset power analytics is making use of AI pre-label data and classified ML, thus delivering high-quality and reliable predictable conditions and material degradation long before a possible failure would happen. Al could detect anomalies in seconds, in a powerful way humans can’t.

In order to determine continued viability we must also consider defensibility:

• Technology defensibility isn’t guaranteed just because a company uses AI. It is perhaps most important to keep a competitive advantage, especially with the speed of AI is (incl. open source) developments these days. To understand defensibility it’s first good to understand the methodology that the technical team is employing. The more insightful, process-driven, and perpetually-improving the approach, the better. Another proxy for defensibility is the caliber of the AI team. If it's packed with people who studied or researched the cutting edge of what's possible is just not good enough. In case there are subject matter experts and the team members have your industry business understanding it will be a good sign.
• Data defensibility relates not only to the unique access to data, but also to whether that data will improve the product over time. Open source is great but if a company is using nothing but open source data, others using similar algorithms it could end up into similar services or worse a new commodity. Yet, if a company has proprietary data that improves the accuracy and speed of an algorithm, that is a long-term sustainable advantage.

Bottom line is that like me on the picture, back in 2017 on a real-world test site, we are all together on the journey of digital transformation to deliver business outcomes. Artificial Intelligence has some big shoes to fill, especially since “intelligence” is a moving target. AI is both much less and much more than we think it is. AI is ready to deliver solutions to important real-world problems with remarkable speed, cost savings, and accuracy. And for your information, the complexity of cables and multiple boxes and desktop on the picture is today the size of a single cigar box and... plug and play to deploy :-)