What needs improvement with SAS Viya Platform?

One thing I expect more from SAS Viya Platform, specifically on the report side, is how the calculations work. It is not that strong in that sense. It was very difficult to build the required calculations directly on the reporting side, and we sometimes struggled. When customizing reports, adding filters, or creating and placing charts in a proper way, it required a lot of effort and time, and many settings were not readily available or visible to users or developers so they could easily build them. We struggled a lot there. SAS Viya Platform is very good overall. It is quite comparable to other BI tools we have in the market. Where we struggle and need improvement is on report calculations and how we can make them more feasible for different kinds of calculations. If we can improve there, along with the setting of the UI, placing elements exactly where we want, and utilizing the entire canvas, that would be valuable. I found it difficult to customize or prepare reports on the reporting side with SAS Viya Platform. Additionally, nowadays there are many AI capabilities that could be built in. I would expect that if those could be integrated into the existing SAS Viya Platform, it would be a really good tool to use. I have not used much of the AI side of SAS Viya Platform, but I would like to use it. If many things could be added, such as automated calculations, automated charts, or generating many insights and narratives, it would make SAS Viya Platform a great tool. Two main improvements would transform this into a great tool: improving the UI and calculation capabilities, and merging AI capabilities.

Based on my implementation, I see a few clear areas where SAS Viya Platform could be improved. The first is a PROC EXP entity analytical gap; my biggest pain point currently is that SAS 9.4 has no direct equivalent in Viya. For social networking and graph analysis, which is critical for detecting connected fraud rings, I am forced to keep SAS 9.4 running. Native graph analytical capability in Viya could significantly improve my team's performance. The second is upgrade complexity; the upgrade process is very involved. In my case, I cannot jump directly to the target release; I need to go to the intermediate release due to dependencies such as version compatibility on the deployment node. A smoother and more automated upgrade path would reduce the risk and efforts. Lastly, certificate and infrastructure management is an issue; my upgrade was partly triggered by an expiring certificate on the Viya environment. Better built-in certificate lifecycle visibility and alerting would help. On the integration side, the PROC HTTP-based indexing is fragile. Currently, I push alerts from SAS BI into SAS Visual Investigator using PROC HTTP indexing jobs, which is low-level and any payload structure change can break the integration. A more robust and configurable connector between BI and VI would reduce maintenance load. I would also state that Kafka-ESP integration needs more visibility, and there is no tight loop between VI and upstream data systems. When an investigator closes a case or marks a claim as confirmed fraud in VI, that feedback does not automatically flow back into the scoring models. A closed-loop integration where adjudicated outcomes feed back into the model would significantly improve detection accuracy over time. Regarding the user interface, Visual Investigator customization is limited; tailoring VI alert layouts, case workflows, and fields to match my specific insurance domain would have been achievable with more flexible no-code customization. I find that Visual Analytics and VI are very much siloed, so investigators work in VI while business reporting happens in Visual Analytics. They still feel like separate products rather than a unified suite.

I think SAS Viya Platform could be improved as it is a little bit limited; for instance, you have to purchase specific features such as charts and quality control, so it is not a full analytics product. You have to buy separate parts and it feels fragmented. In comparison, other packages I have used are more complete and less divided; SAS is quite divided, particularly with quality control, which used to be included, but now requires separate packages for control charts and so on.

There are some areas of improvement I would mention regarding SAS Viya Platform. New users may find it complex initially, as its learning curve can be steep. Understanding CAS and different components is very challenging for a new user. The cost is also quite high; SAS Viya Platform via SaaS can be expensive compared to some open-source or other cloud alternatives. There are also some limitations in Visual Analytics. While it is easy to use, it could be more flexible compared to Power BI or Tableau. These are the main improvement areas where SAS Viya Platform should focus. I would add some additional pointers about needed improvements in SAS Viya Platform. Regarding customer support, it is good, but the response time for complex issues can sometimes be slower than expected. The documentation and learning resources are not very beginner-friendly, making it challenging for beginners to navigate. Regarding integration, it could be smoother with Databricks, Snowflake, and Power BI, but it is more complex than it should be. Since updates are automatic, users are sometimes not fully aware of changes or new features that are being updated, as compared to Power BI, which publishes updates on its website. SAS Viya Platform should work on these aspects.

I cannot think of any improvements for SAS Viya Platform from the top of my head, although users might have suggestions. I would add that some users need more training in SAS and SAS Viya Platform. An AI assist, such as a live assist, would be helpful for them.