I have put a number of ideas on the ideas.splunk.com site for feature requests for the Splunk SOAR product. I posted one of them about three years ago, which finally got implemented in the latest release that just got announced, so the time to implement new features and things like that is a little bit concerning. I tend to post my ideas there so that other people in the community can see the features or ideas. They can then upvote them and make comments on them. I thought that is what the site is for.

Creating playbooks using the solution’s playbook editor, for me, is very cumbersome. There have been instances where I have said to myself that I just don't want to use this editor. I might just use a code block and write my own code within it. I've tried using the editor for some of our playbooks, but I find it's cumbersome. It's easy to drag things in the GUI, but for the actual coding part and joining those bits in a full code, it's not as good as I would like. They have tried to make it as simple as possible, but its functionality is not up to the mark.

The functionality in the playbook editor is 80 percent there, but that 20 percent is still lacking. They could make it more efficient.

Splunk's support for integration is subpar and has room for improvement. 
Splunk should make more effort to keep up with the latest developments in the external world, so that their applications, integrations, and enrichment apps are up to date. Additionally, the documentation and support should be improved, as the experience, their users have had in the past has been unsatisfactory. We were very disappointed that our queries were left unresolved for six months, as it was a time for response rather than solutions. Additionally, several tickets were lodged with Splunk, yet the issue persisted for half a year.

I would request that SOAR add a feature that allows the extraction of documentation from playbooks. This would enable developers to quickly understand the features and use cases associated with a playbook, so they can modify or interact with it. This would eliminate the need for someone to manually explain each playbook in detail. 

I would suggest making the app customizable and deployable in an easy and straightforward manner. This would save time and effort compared to the current process.

Sometimes we flag events based on conditions in the app or service that is sending us the feed, and we focused on a couple. We get some normal events, but we also see some security issues occasionally in the same feed. I don't know if they injected this or if this was the first time we saw it. There was another type that was security-related, but we didn't know about it before. 

We have playbooks written to extract these events and put them into the workflow since it wasn't structured as expected. It was a miss for us. We couldn't figure out why it broke or what actually happened there. It was something in this feed with legitimate and security events, so we tried to understand the names and what we would call them.

It was a unique time. That goes back to an inability to detect these kinds of events. API documentation is typically a weak spot. Many vendors focus on the product first and save the API information for the very last. 

Splunk's integration isn't bad. However, it comes down to which APIs are available. For example, I would like to automate file extraction, and a particular vendor seems to have an API that should do that, but I can't. You're at the mercy of the vendors. While APIs probably leverage more than ever, it's still like pulling teeth to get some vendors to support it correctly. Nevertheless, it's highly beneficial when it works.

Depending on the playbook, it can sometimes get a little crazy and overwhelming, but I think it's generally okay.

What we have seen is if the workflow gets halted or if we want to halt a workflow, it cannot be resumed. We have to trigger the entire plan from step one. That is a bit annoying. If something is wrong, we can't just resume stuff. We'd like it to be possible to pause things without having to start from square one. 

Reporting could be better. We are getting reports, yet not in the way we want. Whatever fails, for example, we want all those errors, the logs, in an attachment, which can be sent easily over an email just by the click of a button. Right now, we cannot send over an email. We have to pull everything, and we have to download it.

The application does not work properly and does not pass the log-based configuration. I feel that some kind of review should happen in the application. This review should validate things so that we can get the right information.

Splunk does not tell us where the IP address is associated with.

The number of playbooks on offer should be increased. 

Some of the training materials are on a basic level. They don't feel like they're really in-depth. I would like to have more advanced and in-depth training.

The UI can be more customizable for the clients.

The algorithm and machine learning have room for improvement and can be more user-friendly.

The integration with the phone system, price, UI, and performance have room for improvement.

Improvements are needed in automation options as customization is limited, which may make complex use cases challenging despite the solution being able to meet basic requirements.

Currently, the tool only allows categorization into two categories, malicious and non-malicious, which has been identified as a limitation by security analysts in various group brainstorming sessions. The ability to create custom categories for emails can benefit security analysts.

There is a lot of room for improvement with the UI. 

I would like to have more integrations with cloud technologies and functionalities such as AI within Splunk SOAR.

The pricing could be a bit more reasonable. It would be great if it were feasible for smaller organizations. 

SOAR is probably the most unreliable product Splunk has and that's because most of it is content driven from what you put into it. There are certain parts of it that have a little bit of difficulty at volume too. It's always changing. There is new stuff coming out for it that's going to make it a little bit better, but it does have some drawbacks.

It's specifically geared for SOC and not broader automation. The artifact filtering that's forced on everything inside the platform is pretty awful. It's for a subset of active playbooks which, out of the two hundred that we own, I think three or four of them are active, but we have to play with that setting for each one of them. 

Every block should also have that option specifically because if you're not doing the artifact filtering on the front end, it's not good. 

We've had lots of processes that have been victim to filtering not working appropriately at scale. It's hard to actually track down and trace because we can't reproduce the issues that we see in our testing environment or in production. That was two minor versions ago. It might have changed, it might not have, but we don't have a lot of trust in that feature.

UI elements like interacting with our analysts are near impossible. Finding stuff on the actual dashboard is really impossible most of the time. One example is that the timeline takes up three-quarters of the screen, but not a single person uses it because you have to individually set the container, the artifacts, and the actions to a specific attribute field that's really difficult to correlate to the actual events you put into it. The artifacts are really weird too because they're not traditional forensic artifacts. You shouldn't be able to change the value of an actual artifact. It was in that capacity but we also use it for that purpose in the platform. 

We've run into a few minor issues. Some of the playbook writing is a bit complicated. We've had a few hiccups with the source control. We'd really like to use GitHub deployment keys for a dedicated account. We haven't been able to do that. I think those are some of the major ones. 

There is a general learning curve as far as playbook writing goes. 

The visibility of the solution’s playbook viewer depends on the right you assign to the analyst. SOAR has the flexibility to distinguish between the roles of analyst and owner. If the analyst's role is to just work on a ticket, they cannot view the playbook design platform. That is limited to the owner. That can be both a good and bad thing.

A major problem I have faced in SOAR's rights distribution is roles and responsibilities. Suppose I am initially granted user rights or analyst rights, but later on, I also get admin rights. SOAR is unable to amend the limitations of my role. I raised a support ticket with Splunk about this. They said it's a bug in their 5.3.5 version. To fix this, I had to reinstall the entire platform from scratch, just to amend the rights and responsibilities of one role. This bug was not fixed.

Also, the latest GUI is terrible. The previous one was better.

Another point is that while using Splunk SOAR in an investigation is not difficult, there are some complex parameters. We have SOAR case management, but the licensing is going to put a big hole in your pocket. Also, there is an issue with investigation node addition. When you are doing node additions you cannot grant the entire environment to have SOAR visibility into the incident. So when you integrate it with an ITSM tool, like ServiceNow or Jira for ticketing purposes, there is a challenge. When you do nodes for investigation on a regular basis, sometimes it does not update our ServiceNow platform, which is terrible. It is a redundant activity for an analyst to update that in the case management as well as in the ITSM tool. Although SOAR provides integration, the functionality of investigation and nodes is terrible when it comes to integration.

An additional area for improvement is custom function creation. It's terrible. A newbie cannot create custom functions right away. They would require a solid understanding first.

Also, the reporting is really awful. If I want to do a report for a customized time period, such as the last three days or the last four days, or from the 10th to the 12th of June, that is not available in SOAR at all. That kind of feature is available in Cortex XSOAR. Reporting is a real challenge.

The Splunk SOAR case management feature lacks some of the functionalities like the possibility to fully customize the fields for the tickets/events and create custom statuses. 

Splunk SOAR can improve IoT/OT security-related case studies or your use cases. Their integration with identity and access management (IAM) solutions is a bit shaky. They don't have good integration with a lot of IAM solutions. They do have good capability in terms of user access management internally, but even with privileged user access, they have a good module. However, if they have to integrate with solutions, such as CyberArk or IBM IAM solutions they are lacking, the visibility of user access is not that much.

Splunk SOAR should improve its ease of upgrade, which is a pain point for us right now. Each upgrade to the version requires expertise and time commitment. Then, we usually have to troubleshoot it with support.

The cost of Splunk SOAR has room for improvement.

There are only problems if the customer is not ready with emergency plans or standard procedures if something breaks. There is some homework to be done before you can really properly use Splunk SOAR.

Resolution times could be faster in terms of support.

It could be easier to implement. 

The initial setup could be simplified.

In my opinion, the focus should be on improving its simplicity, specifically the interface, and configuration.

I don't have much experience with that. I'm not sure as I don't have much technical knowledge about SOAR in general. I have a little bit of experience with SOAR. I can't speak to any shortcomings right now. 

The scalability could be better.

It's an expensive solution. 

In the beginning, we couldn't find any specific documents for every function. It wasn't easy to navigate to what we needed. However, lately, it has improved and we are able to find Splunk documents for all the functionalities of Phantom. 

It would be helpful, on the other hand, if there were videos regarding each functionality. That would make it even easier to work with Phantom. We are able to find some documentation in written form, and that's fine. If it is in a video format, then it would be better due to the fact that, in some environments, we find some other issues or something and it would be nice to have a visualization of the process.

The solution is a bit more expensive than other offerings.

I'd recommend that the solution add some new apps, or some average services, like bots or G-Suite. We may already have G Suite in Phantom. Bots, like any common VPN service, would be great, however.

Phantom was only recently acquired by Splunk so it is not fully integrated yet. Our area of concern is that Splunk Phantom works with the other Splunk products. At this point, there are certain things that are not fully operational across the rest of the product line.

The extension of the product to allow for better integration with other data sources is something that needs attention. We want to see improvements made to the APIs such that we can connect to many different systems and data sources.

The search capability could be improved by way of better indexing and also integration with third-party solutions such as Elasticsearch.

I would like to see escalation management and integration with communication tools like Slack.

I would like to have more capability around analytics.

There needs to be a better facility for documenting and storing issues, as well as being able to find those issues. Splunk does a good job of that, so I think that it will be done.

We haven't had too much experience on the solution.

The solution is relatively new in the market.

It would be ideal if we could automate processes even more.

The interface is great, however, they could still keep refining it to make it even more user friendly.

We've had trouble implementing the solution with Microsoft products. There seems to be an integration gap. 

The pricing of the product could be more reasonable.

I haven't used it fully, but based on my usage, I could not find simulation tools and features. It currently lacks simulation features, which are important for me for creating a playbook.

It is also very expensive for my region.