For the Remainder of the Work

New goal-driven automations, enterprise app integrations, and messaging assistants speed up the 75 p.c of labor that coding brokers can’t contact

Zencoder launched Zenflow Work, a significant enlargement of its AI orchestration platform that accelerates every thing across the code: the planning, the coordination, the reporting, and the communication that devour most of an engineer’s day. After which it goes additional, bringing the identical structured, goal-driven AI workflows to product, advertising, gross sales, buyer success, finance, and HR groups.

AI coding brokers have modified how software program will get constructed. Engineers use Claude Code, Codex, Cursor, and others to jot down, take a look at, and assessment code sooner than ever. However ask any engineer how they really spend their day, and coding is perhaps 1 / 4 of it. The remainder is chasing info, sitting in conferences, critiques, emails, paperwork, prep work.

“Coding brokers made engineers 10x sooner at writing code. However there’s truly extra routine work outdoors of coding than there may be in coding, with individuals spending three quarters of their time doing that,” mentioned Andrew Filev, CEO and Founding father of Zencoder. “Zenflow Work is for the opposite 75 p.c.”

The Different 75 p.c: What Coding Brokers Don’t Contact

Zenflow Work introduces goal-driven automations that deal with the work surrounding code: multi-step workflows that execute throughout Jira, Linear, Notion, Gmail, Google Docs, and extra, and conclude routinely as soon as success standards are met.

• Standup briefs: Each morning, the agent searches Jira for points up to date within the final 24 hours, teams them by what shipped, what’s in progress, and what’s blocked, and writes a 5-bullet abstract. It’s ready earlier than you sit down.
• Launch notes: The agent reads PRs merged to primary this week, categorizes by function, repair, and enchancment, and writes user-facing launch notes in Google Docs. What used to take half a day takes 10 minutes of assessment.
• Stakeholder updates: The agent compiles accomplished points from Jira, drafts a weekly replace e mail in Gmail, and saves it as a draft to your assessment. The e-mail that will get dropped first whenever you’re busy now writes itself.
• Assembly prep: The agent checks Google Calendar for tomorrow’s product assessment, finds all Jira and Linear points talked about within the invite, pulls their standing, reads the Notion spec, and writes a prep doc. 45 minutes of tab-switching turns into a single job.
• Dash retro information: The agent pulls accomplished and rolled-over points, calculates completion price by group member, identifies re-estimated objects, and writes findings to Notion. Retros grounded in actual information, not reminiscence.

Additionally Learn: AiThority Interview with Glenn Jocher, Founder & CEO, Ultralytics

Not Simply Engineers: Each Group Has Their Personal 75 p.c

Gross sales reps spend extra time on proposals and follow-ups than promoting. Entrepreneurs spend extra time coordinating than creating. Finance groups spend extra time chasing receipts than analyzing spend. Zenflow Work handles the repetitive multi-step work throughout all of them.

• Gross sales: A prospect emails necessities, the agent reads the thread, drafts a proposal in Google Docs, asks “they talked about SOC 2 compliance, ought to I add a safety part?”, and refines primarily based in your reply. A 3-hour writing session turns into a 20-minute assessment.
• Advertising: The agent reads PRs merged to primary, writes a advertising transient for every user-facing change, and creates a marketing campaign concepts web page in Notion. Entrepreneurs know what shipped inside days, not weeks.
• Finance: The agent scans Gmail for subscription receipts, cross-references towards the accepted vendor record, and flags duplicates, unapproved instruments, and something unused for 30+ days. Most corporations have 20–40 p.c SaaS waste. This finds it.
• HR: As assessment cycle approaches, the agent pulls six months of Jira and Linear tickets for every worker, reads their 1:1 notes in Notion, and writes a structured prep doc. Managers stroll into efficiency conversations with information, not impressions.

Your Agent, in Your Messaging App

Customers can begin duties, get real-time updates, and chat with their Zenflow agent instantly from Telegram or Slack, with WhatsApp assist coming later. The identical messaging-first interplay that made OpenClaw a phenomenon, delivered by means of structured workflows and a curated set of integrations with a tighter safety perimeter.

“Zenflow’s orchestration strategy is already confirmed on the coding aspect: in inside benchmarks, clever routing throughout pipeline levels made coding 2.7x cheaper per resolved job, at equal or higher high quality. Zenflow Work brings that very same multi-model engine to enterprise workflows.”

Deep Enterprise App Integrations

Zenflow brokers join on to the instruments groups use each day, studying, writing, and coordinating throughout platforms:

• Venture Administration: Jira, Linear, Notion
• Communication: Gmail, Slack, Telegram
• Productiveness: Google Docs, Slides, Sheets, Drive, Google Calendar
• Distant MCP connections with OAuth for Amplitude, Sentry, and extra third-party companies

The discharge additionally provides built-in file administration (preview and edit PPTX, PDF, XLSX, DOCX) and a browser tab that lets brokers view web sites, examine parts, and seize screenshots as a part of their workflows. Extra integrations are on the best way.

Multi-Mannequin Orchestration: 2.7x Cheaper Per Resolved Process

Zenflow’s model-agnostic structure can also be an financial benefit. Inside analysis on a tough subset of SWE-bench Professional exhibits that routing every workflow stage to the proper mannequin, relatively than working a single frontier mannequin end-to-end, cuts value per resolved job by about 60 p.c whereas matching or enhancing output high quality. Holding the planner mounted to Claude Opus and swapping in Gemini Flash because the implementer made the pipeline about 2.7x cheaper than working Opus finish to finish. The orchestration layer manages this routing routinely.