News

Baz Announces Planner and an Extended Seed Round Co-Led by Battery Ventures and boldstart ventures

At AI Engineer World's Fair, Baz announced Baz Planner - a gateway that eliminates entire classes of bugs and vulnerabilities in code planning - alongside $9M in new funding, bringing total raised to $17M.

Agentic coding company Baz announces a planner that lets engineering teams eliminate entire classes of bugs and vulnerabilities at superhuman scale, and an extended funding round co-led by Battery Ventures and boldstart ventures.

SAN FRANCISCO - June 29, 2026 - Today at AI Engineer World's Fair, Baz, the leading agentic coding platform company, announced Baz Planner. The new product is a gateway that sits between developers and the codebase, automatically routing every idea through dynamic loops that instantly detect, root-cause, patch, and validate vulnerabilities and bugs - proactively rewriting coding plans to eliminate entire classes of issues before they reach production.

plan ∙ DATA-2187

baz planner extension

Claude Code · Baz Enterprise

~/orders-pipeline

>Plan DATA-2187: backfill late-arriving events in the orders pipeline. Don’t write code yet.

●Routing the plan through Baz to check it against the warehouse + dbt project before I propose anything.

●Baz(plan · data-loss & cost scan)└ Done (21 tool uses · 39.2k tokens · 31s)

●Proposed plan — late-event backfill in orders pipeline (DATA-2187)

1.Stamp event_ingested_at (processing time) onto every row in stg_orders.

2.Drive the fct_orders incremental model off event_ingested_at with a 3-day lookback.

3.Add an idempotent backfill for the last 14 days, guarded against --full-refresh in prod.

Baz Filtering the incremental model on event_time silently drops late events — they land after the watermark. Switched the predicate to event_ingested_at so no rows are lost, and the lookback re-merges late arrivals idempotently.

−51% tokens vs. an unplanned session · 0 silent data-loss paths left open

Approve this plan before any pipeline code is written?

Baz was founded by the team that scaled Palo Alto Networks' Cloud AppSec business to lead code-to-cloud security. The same four principles that reshaped cloud development are now needed for AI-generated code: observable, explainable, predictable, and reproducible. The explosive adoption of AI coding has fueled this need in every enterprise.

Baz's flagship product for AI code review already helps frontier development teams govern and secure AI code with purposeful agents for enforcing coding standards across product, design, architecture, security, and SRE. Baz has more than 100 leading AI, infrastructure, and cybersecurity customers around the world since launching last year, and its AI Code Review product has been ranked #1 on the precision-weighted Code Review Bench.

While cyber-capable models can autonomously scan millions of lines of code and discover subtle flaws, Baz Planner rigorously scrutinizes every ad-hoc change against current and prospective architecture, automatically flagging them before any code is saved. Baz Planner evaluates every model suggestion against a strict risk matrix, blocks unsafe paths, enforces defined boundaries, and only allows progression toward production after rigorous risk mitigation and approval. With Baz Planner, teams have reported up to a 65% reduction in downstream rework, measured by the frequency of revert and hotfix PRs following a merge.

"Our customers are building the backbone of the new AI stack, and they pushed us to go beyond review and intervene as early as the planning stage, where bugs and vulnerabilities are cheapest to eliminate. Baz exists because they refuse to accept that AI-generated code means blindly accepting risk," said Guy Eisenkot, co-founder and CEO of Baz.

Baz's agentic coding platform helps engineering teams ship code faster with agents operating like experienced engineers: in the context of behavior, requirements, APIs, architecture, production systems, and security. Instead of focusing on style and correctness, Baz analyzes how new code affects runtime to catch bugs, silent regressions, breaking changes, and security flaws.

Purpose-built agents for engineering work

Baz's purpose-built coding agents run on top of its platform and can be dropped into any engineering workflow:

Spec Reviewer Agent validates code against product requirements, designs, and expected behavior.
Advanced Security Agent reasons across auth and network boundaries, infrastructure, pipelines, and finally application code to uncover vulnerabilities.
SRE Agent correlates repository changes with production telemetry to identify reliability, performance, and observability risks, then proposes fixes.
Fixer Agent automatically applies and validates safe code changes in an isolated runtime, turning review feedback into tested commits.

Baz agents collect requirements, designs, docs, execution logs, user comments, and organizational knowledge into one place. Baz agents can also be programmed to enforce internal standards, engineering practices, security policies, and domain-specific requirements, helping teams scale consistent, high-quality reviews across every change.

Baz raises an additional $9 million in seed funding, bringing total raised to $17 million

Today Baz also announced $9 million in new funding as part of an extended seed round co-led by existing investors Battery Ventures and boldstart ventures; new investors AFG Partners and Disruptive VC also participated. The new funding will be used to support research across coding agents for engineering work.

"Guy and the Baz team built the code-to-cloud security playbook at Palo Alto Networks, and they are now applying that same rigor to the most pressing challenge in AI-native engineering. As development teams deploy fleets of coding agents, Baz is becoming the super harness that coordinates them - from spec-driven development and UI review to security, quality, reliability, and planning - ensuring AI-generated code ships safely and at scale," said Barak Schoster, Partner at Battery Ventures.

"We're thrilled to double down on Baz. Every other tool reviews code after it's written. Baz's Planner intervenes at the planning stage and kills entire classes of bugs and vulnerabilities before the code is ever authored. That's institutional engineering judgment, codified, and it's the layer the AI stack has been missing. This team ships with incredible velocity, and they're just getting started," said Ed Sim, Founder and General Partner at boldstart ventures.

About Baz

Baz builds agents that operate codebases. Baz creates precision coding agents that review code, find bugs, enforce standards, fix issues, and help teams ship high-quality software faster in the age of AI. Baz is backed by Battery Ventures, boldstart, AFG Partners, and Disruptive VC. For more information, go to https://baz.co/.