Bloomfilter is looking to meld process mining and process intelligence with software development via a partnership with Celonis. The goal: make developing applications at scale more efficient.
The company will power the Celonis Software Development Lifecycle Management (SDLC) application. Celonis for SDLC powered by Bloomfilter will align business priorities with technology initiatives, measure financial impact of resources and map and analyze processes and recommend ways to be more efficient.
At Constellation Research's Connected Enterprise, I caught up with Erik Severinghaus, Co-founder and Co-CEO of Bloomfilter, to talk shop. Here are some of the takeaways.
The intersection of software development and process mining. Severinghaus said the combination of software development and process mining is designed to build software that is "more efficient, more predictable, more observable and more nimble."
While software development is scaling with generative AI code, enterprises need to focus on the underlying processes. Bloomfilter, which is built on the Celonis platform, pulls in data from tools including Atlassian's Jira, GitHub from Microsoft and GitLab to show how software development is being done.
Severinghaus said:
"We pull all the information about how the software development process is working, and then we use process mining technology to help a customer understand, where there's waste, inefficiency, rework and steps that are skipped or not in compliance with the company's process. These issues could lead to production outages and security vulnerabilities."
What process mining reveals. Process mining typically has a few surprises for enterprises and the biggest one in software development tends to be rework. "You connect to these systems and pull out the way the process moves and we find things like 40% to 45% of tasks require rework," said Severinghaus. "When you look at the business impact of all of that rework, it becomes really apparent why 78% of software is late, over budget or doesn't ship at all."
Bloomfilter monitors software development processes, software process adherence, allocation of costs and the value being generated.
Why process matters with generative AI. "Enterprises are increasingly using automated code systems and there are a lot of interesting places where process interacts with AI," said Severinghaus. "First we couldn't do this without AI, but when you think about the software development life cycle there are about 15 different tools being stitched together. Customers are using AI to build software and there's been less and less governance."
Severinghaus said process mining is a way to analyze software development and pull out what's important. One of the most important things to do is to assess risk. "We are using generative AI to stitch the process together to govern what's happening in an AI-driven SDLC," he said.
The C-level buyer. Severinghaus said the buyer of Bloomfilter varies since every company is a software company today. For software companies, the primary buyer of Bloomfilter is the CTO or chief product officer. In more mainline companies, CIOs tend to be the main buyers with the CFO and COO playing a big role. "They want observability and they want to see what's happening within this black box of software development translated into words they can understand," said Severinghaus.
What's next? Bloomfilter's partnership with Celonis gives it the ability to plug into more systems including Salesforce and ServiceNow to deliver process intelligence for SDLC. Severinghaus said governance will also become vital as AI agents are including in the software development process. "If 2024 was the year of throwing AI against the wall and see what use case sticks, 2025 is very much going to be the year of how we govern and get business value out of it," said Severinghaus.
