Bloomfilter

Bloomfilter builds a measurement layer over the software delivery lifecycle. It connects to the systems engineering organizations already run — issue trackers, source control, code review, CI, deployment and AI coding agents — and reads the metadata those systems generate as a byproduct of daily work, without requiring new workflows or manual tracking. The collected events are stitched into a continuous map of the delivery process, tracing work items end to end, measuring flow at each stage (strategize, specify, plan, build, verify, review, ship, support) and binding each agent session to the specific work it touched.

The resulting dataset is intended to let organizations attribute agent activity to concrete outcomes: the commit an agent wrote, the pull request it entered, whether the change shipped cleanly or returned from QA, and whether time saved during coding reappeared later as rework. Bloomfilter benchmarks performance, process adherence, risk and cost against a pre-AI baseline and exposes the results through its own dashboards, through the reporting tools customers already use, and through plain-language chat from assistants or Slack, with each number traceable back to its source record.

Earlier positioning described Bloomfilter as a process intelligence platform for software development, applying process mining to time-series data from a customer's stack to model a team's specific development process and then identify improvements, giving stakeholders a shared view of what is being built, when it will land and what it costs.

Founding story

Andrew Wolfe, described as founder and CEO and later as co-founder and co-CEO, previously worked at Explorys, Tableau and OnShift and built the company around the problem of managing technology products at scale. Erik Severinghaus is the other co-founder and co-CEO.

Business model

B2B software sold to engineering and technology organizations, delivered as a platform that integrates with customers' existing development toolchains via connectors.

Traction

Connectors are described as running in customer production environments. At its 2023 seed announcement the company reported general availability of its platform, and its earlier investor wrote of a growing pipeline of customers and users.

Latest developments

The company's current product messaging centers on connecting delivery systems, attributing every AI agent's work, and benchmarking performance, adherence, risk and cost against a pre-AI baseline, with dashboards and natural-language querying.