Brainfish founders build Ballet to turn plain-English workflows into reviewable code

Why it matters

Ballet tests whether generated, reviewable code can clear integration backlogs for proprietary systems while giving engineers control over what reaches production.

Daniel Kimber and Ajain Vivek (@ajainvivek), the founders of Brainfish, are offering Ballet in early access to design partners with multi-system workflows stuck in engineering backlogs. Users describe an outcome in plain English, and Ballet generates reviewable, version-controlled integration code for internal databases and third-party software.

The product takes Brainfish into a different operating problem. Since its 2022 founding, Brainfish has organized product knowledge so AI support agents can answer customer questions. Ballet moves from answers to actions: routing a lead, detecting account surges, changing an order or triggering a renewal play across systems that were never designed to work together. (Ballet)

Ballet's pitch centers on operations teams that manually move data between systems or wait for engineering resources to build integrations. Its early-access program targets RevOps, Support Ops and product-led growth teams that can identify the desired outcome and the relevant data but lack the engineering capacity to build and maintain the workflow.

From customer support to execution

The shift follows the founders' own path through product and support work. Kimber studied Information Systems at the University of New South Wales while working at hotel-software provider SiteMinder, where Forbes Australia reported that he led the migration of 36,000 customers to a modular platform. He later held product leadership roles at Willow and ordering-and-payments provider me&u. Vivek, Brainfish's CTO, has worked across software architecture, infrastructure and automation. (Forbes Australia)

Before Brainfish, Kimber and Vivek spent a year building an earlier venture that failed. A Brainfish account of that period says they went six months without speaking to customers and misunderstood the buyer behind the problem they were trying to solve. The founders reset at the end of 2022 and built Brainfish around customer-experience problems they had encountered in their own careers. (Brainfish)

That history helps explain Ballet's narrow opening focus. The founders are starting with revenue operations, marketing, sales operations and support operations, where employees understand the desired outcome and the data involved but often lack the engineering capacity to build and maintain an integration.

Ballet's examples include enriching and deduplicating inbound leads before they reach Salesforce; combining product telemetry with Clay and ZoomInfo to alert sales representatives in Salesforce and Slack; and propagating pricing or SKU changes from Salesforce CPQ into Stripe, an internal billing service and Freshdesk. These are Ballet's proposed use cases rather than independently documented customer deployments. (Ballet)

Code instead of a visual workflow

Ballet says it generates reviewable, version-controlled code against an API when an integration is needed. The product separates deterministic steps, where the same input should produce the same result, from agentic steps that require judgment. Ballet is presented as Brainfish's execution layer for multi-system workflows, generating code that engineers can inspect and approve. (Ballet)

That design addresses a central obstacle for AI automation inside businesses. Technical review becomes more consequential when an automated workflow can edit a live CRM record, change a price or send data into a billing system. Ballet's claims about consistency, auditability and maintenance have not been independently benchmarked.

Ballet's site currently solicits design partners through a "Get early access" flow. Its signup page previously stated "Coming June 2026," but the available material does not establish whether that date marked a launch, an early-access opening or a target that slipped. The current offering is hands-on: Ballet proposes a 30-minute working session on a customer's real systems, then produces code for the customer's engineers to review. (Ballet)

A crowded automation market

Ballet enters a category filled with mature automation platforms and newer AI builders. Make's Maia creates automation scenarios through natural-language conversations. Zapier MCP gives AI tools access to more than 9,000 app connections and over 30,000 actions, according to Zapier. Pipedream combines prebuilt actions with custom Node.js, Python, Go and Bash code. n8n similarly mixes visual workflows, code and AI components.

Ballet's wager is that connector catalogs and visual builders break down when a workflow touches proprietary software or grows into a large graph that only its original author understands. Ballet says it can write against any API and leave behind inspectable code instead of another visual workflow that an operations employee must maintain. (Ballet)

The hard part begins after the first successful run. APIs change, internal schemas drift, authentication expires and edge cases accumulate. Ballet says it will build, run and keep workflows working. Its ability to maintain generated integrations over time will determine whether the product removes engineering work or simply moves that work into a new system. (Ballet)

Brainfish's second product bet

Ballet is a product from the team behind Brainfish. Its site identifies Brainfish Pty Ltd as the legal entity and describes Ballet as part of Brainfish. Brainfish describes itself as the knowledge layer and Ballet as the execution layer. A Brainfish answer can trigger a Ballet workflow, while Ballet can also operate independently against a customer's existing stack. Ballet says it is covered by Brainfish's SOC 2 Type II program, with details available through the Brainfish Trust Center. (Ballet)

Brainfish says it was founded in 2022 and has answered more than 10 million customer questions. In 2025, the company announced an A$10 million pre-Series A led by Prosus Ventures, with participation from Macdoch Ventures and Peak XV's Surge. Kimber said the money would fund Brainfish's San Francisco expansion and engineering growth. (Brainfish)

Forbes Australia previously reported a $3.85 million seed round led by Surge. The available disclosures do not clarify whether that financing is included in the later A$10 million figure.

Ballet puts the founders' support-automation experience behind a broader thesis: commercial employees closest to a workflow should be able to specify what it needs to accomplish, while software handles the integration work and gives engineers a reviewable artifact. Kimber and Vivek learned in their first failed venture that proximity to a problem does not guarantee an understanding of its buyer. Ballet's design-partner program gives them a controlled way to test whether operations teams will entrust generated code with production workflows.