Guest: Divanny Lamas (LinkedIn)
Company: Transposit (Twitter)
Show: Let’s Talk

Transposit’s annual State of DevOps Automation and AI Report uncovered an interesting paradox: a lot of companies are investing in process automation, but they’re still seeing an increase in the number of incidents and rising costs.

In this episode of TFiR: Let’s Talk recorded at the KubeCon in Chicago, Transposit CEO Divanny Lamas talks about the current trends and pain points in incident management and how the company’s AI-powered solution is helping companies deal with these challenges.

Current trends in the market:

• Many companies have some level of incident management in place, but the tools they use focus notification and alerting.
• The cost of an incident for organizations is about hundreds of thousands of dollars per hour and it brings everyone in a firefighting mode, instead of building technology or a product.
• Managers can’t hire enough SRE, DevOps engineers, and platform engineering team members to cover what they need to do.
• The current AI technology is nowhere near the point of replacing anyone. Instead, it is a force multiplier; it lets teams do more with less. It allows them to stay capable of handling all the load that’s on them.
• People that are in cloud environments are overwhelmed with changes, tooling, and the number of systems to coordinate between. AI can help reduce stress, but it’s still a very long way from being able to take an entire person’s job and do it.
• There’s awareness and acknowledgment that AI is going to play a really big role in managing incidents, but in terms of maturity, it’s still in the very early stages.
• There is a lot of skepticism around the use of generative AI in incident management. It’s just too early for people to fully automate some of these workflows and fully trust the outcome.

How Transposit helps companies: 

• People hate writing incident updates because they know that everyone is going to read them. It’s stressful to think about creating an update that summarizes all the things that just happened. AI takes a lot of that friction, mental load, and toil away. The goal is to reduce the number of parallel jobs that a person has to do during an incident and let them focus on solving the actual problem.
• It will integrate with your existing tooling to leverage the investments you’ve already made, not to come in and make people rebuild every part of the process. Organizations typically will integrate it into their Jira environment, into their observability tooling, or into whatever they’re using for CI/CD.
• Transposit’s approach is to keep humans in the loop and augment (not replace) them. When there is an AI-driven suggestion or insight, people can respond and choose to run or choose not to run. For example, if we are creating a summary of messages that are happening in the alert, there is a prompt and an edit button so people can come in and modify.
• The goal is for our AI to feel more like a valuable teammate versus a system that just tells you what to do. It’s about that collaborative approach.

On incident management training and culture: 

• It’s very common in Incident Management circles for people to talk about the challenges of training. AI is an intuitive way to work with the solutions that brings down that training overhead.
• Instead of asking someone to provide an update every 30 minutes, you can have a friendly chatbot that says, “I noticed that you haven’t made an update in a while” or “This particular thing that you’ve done seems like it might be deserving of a communication to the wider team.”
• It becomes much more collaborative and ongoing on-the-job training, which really helps with that cultural adoption issue.

What’s next: 

• An unlimited free tier with unlimited users, because Transposit wants to make sure that everyone could see the magic of AI and Incident Management. It’s available on their website and anyone can sign on with Slack. The package includes basic AI.
• The enterprise tier with more sophisticated AI plugins, including change integration, ability to bring in status messages from other systems, and automatically checking those things for you.

This summary was written by Camille Gregory.