IBM Watson Machine Learning Review

Good automation capabilities and saves companies money but there's not real market comparison to tests accuracy


What is our primary use case?

Most of the use cases have been around taking in a lot of human-created data by providers. For example, electronic medical records including the notes, the information, and the clinical aspect of care. What they're typically trying to do is adjust a lot of that type of information. 

Sometimes there is member feedback or member sentiment that's captured. The idea is to try to quickly assess and analyze that information and then prioritize it. The goal is to put the data in buckets for particular use around quality improvement and around identifying risks earlier, and things of that nature.

How has it helped my organization?

The solution has allowed us to remove a lot of manual labor and cut costs around getting people to perform certain tasks.

What is most valuable?

The most valuable aspect of the solution's the cost and human labor savings. 

The ability to automate a lot of the work helps save organizations time and money. It also accelerates the whole digitization process. It's tremendously valuable.

What needs improvement?

I haven't dealt with the solution as significantly in the last probably two or three years. That said, my last deeper dive into that was around the need for the product within the organization. I'm sure it's gotten better and better as the program has gotten better, however, early on, they relied heavily on building out these massive reference tables. That was a ton of the work that had to be done

Honestly, I haven't seen any comparative report that has run the same data through two different artificial intelligence or machine learning capabilities to get something out of it. I would love to see that.

The top three industries should go head to head to feed the same data and then get evaluated for accuracy. Something like that would be tremendously interesting. I don't know if something like that exists.

For how long have I used the solution?

I have a long history with the solution. I've used it for 15 years - well before IBM purchased it and changed the name.

What do I think about the stability of the solution?

I would have to assume it's getting better due to the fact that it's being more widely adopted. More companies are paying the vendors to use it. I would have to assume it's gotten better. 

What do I think about the scalability of the solution?

My understanding of the solution is that it is very scalable. If a company needs to expand it, they can do so easily.

How are customer service and technical support?

I have never dealt with technical support. I can't speak to their level of responsiveness or knowledgeability at this time.

How was the initial setup?

It's been many, many years since the implementation of the product. 

I don't remember it being a complex operation. What I remember was that they were trying to do it kind of as a SaaS solution where they used open connections and set up APIs and pushed data through a cloud-based model. As long as the sender had the ability to communicate in that way, it was pretty simple.

There are other healthcare companies, both on the payer and provider side that are still not there from a technology perspective. So they had to send batch files and lots of things like that. They couldn't use electronic CCDs for clinical records. Due to this, there was a limitation. It may be more complicated when the sender can't match the technology that the receiver would be looking for them to utilize in terms of delivery.

What's my experience with pricing, setup cost, and licensing?

I can't recall the exact pricing. It's my understanding that the gist of the pricing was based on the volume of work. I don't remember any real detail on the pricing. I can't really say if it more expensive than other options or not.

What other advice do I have?

I'm not a consultant or a reseller. I've only dealt with this solution in the past as a user, for the most part.

When done right, this automation can be very helpful. There are companies that were pioneers in this space. For example,  Autonomy, which was purchased by Hewlett Packard. However, in that case, it was an epic failure. They've been in a lawsuit for the last decade.

I've been more focused on other parts of healthcare information. So honestly I'd say in the last probably two, two and a half years, I haven't really dug into the analytic capability.

I would advise other organizations to proceed with caution and do a lot of homework and maybe pilot the solution before diving in.

Overall, I would rate the solution at a seven out of ten.

**Disclosure: I am a real user, and this review is based on my own experience and opinions.
Find out what your peers are saying about IBM, PyTorch, TensorFlow and others in AI Development Platforms. Updated: September 2021.
535,544 professionals have used our research since 2012.
Add a Comment
ITCS user
Guest