No Developers, 32 Automations: How Buzzvil's Business Team Turned No-Code AI Workflow Automation Into a Team System

Buzzvil, a Korean ad-tech company, has a BD Strategy & Operations (BDSO) team that grows revenue for media partners. Nobody on the team writes code, and there are no developers. Yet it now runs 32 AI workflow automations. The interesting part isn't what they built — it's how personal shortcuts became a shared system.

Start with one annoying task

Roll out a tool company-wide and a few early adopters use it before it fades. BDSO started from the other end: each person picked one task they kept doing by hand.

• Cut repetitive time: an hour a day of email reply prep became AI summaries and draft replies, with a 10-minute target.
• Remove error-prone steps: tax invoice deadlines checked by eye became staged alerts.
• Stop waiting on others: changing 10–20 ad units one by one, in a developer's queue, became a single command — eliminating 90%+ of manual steps.
• Simulate decisions in numbers: settlement terms for new media partners are modeled before signing, since one clause can swing several million won a month.

From personal tool to team tool

When revenue suddenly drops, an internal tool called AutoBDSO pulls scattered data and messenger context together and returns a first-pass analysis in seven minutes.

What broke at 32

• Most automations ran on personal PCs and stopped when their owners were away.
• People built overlapping tools.
• When authentication broke, empty reports went out anyway.

The sum of individual skills wasn't yet team capability.

The leader's job: systematize

The team lead's role shifted from building the most automations to connecting them: creating space to propose and share ideas, moving automations to a shared environment on common data, assigning owners and adding review and improvement processes. That took about six months.

A playbook for marketing and sales teams

1. Ask each person to eliminate one painful repetitive task.
2. Track results in numbers — minutes saved, steps removed, errors avoided.
3. Before you pass five to ten automations, set up shared infrastructure, owners and checks.

For more from Buzzvil, read Internal AI Agents Need Data Structure Before Models. On why enterprise AI change lags, see Publicis Calls the AI Marketing Opportunity a 'Two-Speed Boat'.