What is our primary use case?
My main use case for Devin is that it helps me with scripting, troubleshooting, and understanding technical tasks. In an L2 support environment, I often use it to work with scripts or investigate issues I have seen before, and having an AI assistant saves a lot of time. It is also useful for explaining code and suggesting different ways to approach the problem.
A specific example of when Devin helped me with a scripting or troubleshooting task is when there was a problem with code that was not working properly. My company does not allow other AIs, but Devin is integrated with Claude, and we were unable to find that bug inside our IDE, so we used Devin, and it was really helpful for us.
Devin helped me pinpoint and resolve the bug in that situation with code suggestions because nowadays AI is suggesting code in a way that mirrors human capability. AI is even writing and suggesting where the bug actually is and how to resolve it.
My main use case and how I use Devin day-to-day includes reducing the time I spend on repetitive research, scripting, and initial troubleshooting. Instead of starting from scratch, I use the same code, adapt it using Devin, and get a result.
What is most valuable?
The best features Devin offers include multiple AI models, ranging from ChatGPT to Claude to Claude 4.5, with every version available. Obviously, this requires many tokens, which we already have for our organization.
Out of those AI models, Claude is the one I find myself using the most, and it really makes my workflow smoother because it not only provides suggestions but even fixes problems and helps us resolve them. Devin has impacted my organization positively by addressing previously common bugs that required a lot of people working on a single bug for extended periods and late nights. After implementing Devin, we share code on GitHub and resolve issues right away. The suggestions make us think and encourage us to approach problems differently and more clearly.
I can share specific metrics. Previously, we had around fifty people for a simple task, but now even if there are people on leave, we can work with ten or twenty people.
What needs improvement?
Devin can be improved since the results are not always perfect on the first attempt. For more complex tasks, it could add new Claude features such as Fable.
The interface of Devin is good. Overall, it is a productive tool for IT professionals, and it does not replace technical knowledge; it just makes everyday troubleshooting and scripting much faster.
I chose nine instead of ten because it needs to integrate and upgrade many things, such as adding Fable and reducing its frequent updates, because my company does not allow everyday updates. My company requires every update to go through enterprise governance.
For how long have I used the solution?
I have been using Devin for approximately a year, before it was called Winsurf.
What do I think about the stability of the solution?
Devin is stable, but it also has a lot of updates frequently.
What do I think about the scalability of the solution?
The accuracy and reliability of output from Devin's AI is currently based on different models, so I cannot give a definitive answer. The accuracy also depends on the different models used. If we use the highest model, it consumes a lot of tokens, so they should try to reduce token consumption. However, it is currently performing well.
How are customer service and support?
Customer support has not been necessary because we already know how to address issues ourselves.
Which solution did I use previously and why did I switch?
We used to use Visual Studio Code, which did not have any AI capabilities, and that is why we made the switch.
What was our ROI?
I cannot speak to investment, but I can discuss how it reduces people's workload, which should give a return on investment for upper management such as a Vice President. It saves time for me as a tech lead.
Which other solutions did I evaluate?
Before choosing Devin, I did not evaluate other options because this is the first AI tool we are using.
What other advice do I have?
My advice to others looking into using Devin is that many people are already using it, but they can go ahead and use it as well because it is really valuable for good coding and saving time.
Regarding Devin's AI capabilities, I think its governance and security are progressing on a correct path.
Devin is deployed in my organization on a private cloud because we have a very secure environment. We use Azure as our cloud provider.
I would rate this product nine out of ten.
Which deployment model are you using for this solution?
Private Cloud
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?
