EXSY:IT - Expert System S.p.A.

[Jurgis]

Spekulatius found this company and asked me what I thought. I wrote some thoughts, but would like to see if CoBF crowd has additional insights. I know there are at least couple people who work in AI field.

 

This is not a buy recommendation, my thoughts are mostly neutral to negative, so if you're looking for a "buy this" topic, you can skip the rest.  8)

 

Links to stock and company:

https://www.marketwatch.com/investing/stock/exsy?countrycode=it

https://expertsystem.com/investor-relations/

 

The company sells a number of NLP and knowledge graph based text data processing products.

 

Their technology seems to be real. Whether it’s great/mediocre/crappy, it is difficult to say. There are no benchmarks or comparisons. (This is overall mostly true for ML/AI companies: unless you know what’s in the system, you don’t really know how good it is. And even if you know what’s in the system, a good team would get way better results than bad team.). Their technology has old roots – the company was established in 1989 – so there is a risk that they are behind the leading edge. It’s not obvious that they are, but it could be. Some things that they mention (like knowledge graph) were popular ~10 years ago and are not very popular now, but they still could have their niches, so it’s not a given that their tech is old. If I had to bet, I’d bet ~50/50% that the tech is old.

 

Their tech and advantage descriptions are abstract and full of buzzwords. What’s worse, they are full of old and abstract buzzwords.

 

One of their customer videos where they show Zurich insurance guy, he mentions being impressed with OCR. OCR is a problem that was solved 20 years ago. If these guys were selling OCR, that’s not impressive at all. Actually they are selling other things, but I’m surprised their customer would mention OCR.

 

Now, they are saying some things that are current, like automated customer support in banks. The issue really is how good they are, which is tough to tell. These are hard, complex problems. It’s easy to wave hands and say that ExpertSystem provides automated customer question answering system. It’s hard to make such system good.

 

I searched for reviews, but found very little:

https://www.gartner.com/reviews/market/insight-engines/vendor/expert-system/product/cogito-intelligence-platform

https://www.predictiveanalyticstoday.com/expert-system/

https://www.trustradius.com/products/expert-system-cogito-intelligence-platform/reviews

 

They don’t specify the size of their deals. I understand why, but it makes it difficult to figure out if the deals are substantial or just tiny exploratory ones.

 

Another concern is that they basically do consulting-like work for their customers. I.e. that they really don’t sell off-the-shelf products, but more like software packages that have to be integrated into customer software for whatever task is. And if the amount of integration work is significant, then their margins can be low/non-existent. They say that 39% of revenues is professional services. I wonder how much of that is “integration” work.

 

They are losing money. OCF is positive, but they seem to spend way above OCF for acquisitions (? "purchase of intangible assets" ). The stock is trading at ~125MEuro with 2019 sales at ~40MEuro. Their 2019 growth was ~10%. IMO that’s not impressive during one of the best years for tech spending/growth.

 

MOIglobal has their white paper: https://moiglobal.com/wp-content/uploads/eis19-massimo-fuggetta-white-paper.pdf

So presumably someone presented Expert System at MOI global in 2018 or 2019.

 

I'm interested what other people think about this. If their tech is good, they could grow substantially and then the valuation might be cheapish. From what I looked at though, I'm not buying.

[Green King]

hard pass. sounds like a consulting business that trying to sell itself as tech. it is more likly they have nothing than it is tech. they would have pivoit to An SAAS product and be bigger and worth a lot more by now. given the time it has got to develope.