The only thing we have trouble with is integrating IBM ODM with the cloud. The product is on-prem, and we need to migrate the rules to the cloud. It's a hectic process. 

There is some promise of how decisions could take advantage of machine learning and artificial intelligence (AI). That has been slow to develop. It is still not clear where the market will take it, but that is something that I am looking forward to down the road.  Today we augment some ODM decisions with AI and with analytics, but I expect that in the next several years we will see much more growth in this area.

ODM has recently released support for Decision Modeling and Notation (DMN) models that can be authored and executed right within the product.  This too is an area that is in its early stages and I expect will mature quickly to an enterprise level.

Configuring it on the Tomcat site and on the client's cloud environment, we faced a lot of issues. After configuring it on the Tomcat side, getting it on the Unix server was an issue.

There are two issues that I am facing right now. The first is that the errors I get from time to time are not easy to debug or easy to understand. They are very vague because if a XOM file is missing or there is a deserialization problem, on the client's side I only get a 500 Internal Server Error. To learn where the problem is, I have to go on the Rule Execution Server and test it myself. The deserialization issue is very vague. The error messages should be more straightforward and easy to understand.

I would also like to see the installation part on a single disk only, instead of on three disks. In our local environment we are installing it on Windows, and for the client side we are installing it on a Linux server. For a new user or a new developer or a fresher, it was tricky to understand which disk should be installed first, even if they're labeled Disk 1, Disk 2, Disk 3. Configuration is kind of tricky for a new person. Right now we have to use Installation Manager. It should be like installing any other software.

Finally, it should also start supporting the NoSQL DB. Currently, It does not support NoSQL. I don't know if the newer version supports it, but as far as I know it doesn't. It would make the job easier. Right now the client had to buy a MySQL license, so they are maintaining a separate database for ODM. It should be kept on a single system. NoSQL DBs are better and it can be done.

The solution's licensing cost needs improvement.

Merging. If merging could be just like SVN, that would help. It should be a lot simpler. That's the only thing I don't like about ODM, merging from our local computer up to the Decision Server. I feel like it could be a lot easier.

Also, extracting specific rules. We've had to do a lot of custom work on that. The testing, we've played a little bit with it, but we have our own testing methods. If it could be as simple as we have for our custom work, that would be great.

The initial set of features that I'd like to see are, for an IT guy like me to start building a rule, we end up using a lot of IT tools. But I think there should be more involvement by business. Business should be able to create those rules from scratch. I think Blueworks does it great. I think there should be a facility for business to really create rules from scratch. The only part for IT should be to make sure the platform is stable, the technology platform, but everything else, the business should be able to create those rules.

It is business-user friendly, but it still needs to be from scratch, from ODM.

Designing orchestration.

I would like to have integration of the user interfaces, and they are putting them together in the next version of the solution. There used to be two separate interfaces. Now, they are merging them into one.

An area for improvement is that the documentation for ODM is not very clear. It has improved. There are very few open-source videos online to learn from. But that is a challenge for a lot of companies. 

Get to the cloud.

ODM could be improved with better integration with other platforms. In the next release, IBM should include integrated analysis tools, summary metrics, and a friendlier way to integrate with dashboards or columns.

One area for improvement is master data integration. That should be more fluid. The others are hierarchical drop-down lists, and hierarchical master data. We have a few hundred others that I have on the list.

What I'm really interested in, what I'd really like to see with this technology, is artificial intelligence, RPA. How can that interface with ODM?

I don't see a need for additional features. I like it just the way it is.