Sendoso Built Hundreds of Workflows. Then No One Owned Them.

Sendoso’s co-CEO described an awkward result of automating sales work: after roughly six months, the company had hundreds of workflows that people couldn’t easily manage. He said no one owned them. For a small marketing team, the transferable question is immediate. Who checks a recurring AI task after its creator moves on?

Kris Rudeegraap told the story in an August 19 panel released as a Marketing Millennials podcast on September 25. Sendoso started by automating repetitive sales-development work, including writing emails, checking data, removing duplicates and uploading lists. Rudeegraap says the team became smaller and got more results. The episode doesn’t provide the headcount, measurement period or outcome data needed to test that claim.

The second phase is more revealing. Encouraged by the first results, Sendoso automated more work. Rudeegraap said the workflow count reached the hundreds. About six months later, he described the result this way:

“No one cared about it. No one owned it.”

That is a management failure with a technical trigger. An automation can keep running long after the person who built it stops watching. If it drafts a wrong email, pulls a stale list or repeats a broken step, the organization needs someone who notices and has authority to fix it. More workflows create more places for responsibility to go missing.

Sendoso’s answer: make ownership visible

Rudeegraap now describes Sendoso’s AI agents as role-based digital colleagues that live in its org chart. In the podcast discussion, he says they have access to tools, reporting lines and performance reviews. The company’s proposed structure gives each agent a human manager.

The metaphor of an employee has limits. Software doesn’t have judgment, incentives or a career. What the org chart does supply is a place to put the name of the human who answers when the output is wrong. A job title for the bot matters much less than the manager attached to it.

This is distinct from the running cost of one agent, which we covered earlier. That story asked how much human work it takes to keep a tool useful. Atlassian’s AI team showed how agent output can also increase the work of writing instructions. Sendoso’s account adds the question a growing team faces: can you find all the tools at work, and can you name the owner of each one?

The staffing claim itself isn’t new. Sendoso leaders discussed their sales-development redesign in March. The September podcast adds detail on what happened as automations multiplied. It remains the company’s account, with no independent inventory of the workflows or audit of its results.

Marketing inherits the same problem

The first Sendoso jobs were sales-development tasks, not a marketing department’s campaign operations. But the failure transfers cleanly. A marketer can build an agent to refresh lead lists, rewrite webinar follow-ups or flag campaign changes. Three months later, those tasks may run every week while nobody remembers the rules they use or who approved the last change.

Jessica Vose, Kana’s head of marketing and another guest on the panel, has given a more grounded starting point in her own writing:

“Clean up your data. Document your workflows.”

She also calls for removing unnecessary approvals and training the team on the new tools. Her company sells agentic marketing software, so that advice comes from a vendor with a stake in adoption. The advice still names the work that an org chart alone can’t do. Someone has to decide which data the agent may use, what a good result looks like, and when a person takes over.

For a five-person marketing team, that doesn’t require a new management layer. It requires a short inventory. List every recurring AI workflow, the person who owns its output, the data it touches and the last time someone checked it against a real result. If a workflow has no owner, the next decision is whether it should keep running.

The Sendoso story has an uncomfortable sequence: automation reduces repetitive work, success invites more automation, and the new work of supervision becomes hard to see. The company says it made that work visible by assigning managers to agents. A marketing team can make the same move before it reaches hundreds of workflows. Name the owner when the first one goes live.