Hermes audits ad spend and proposes creative variants
Eric Siu describes a Slack ad-strategy workflow: Hermes pulls spend-performance data for a 78-post audit, explains waste and prepares three creative variations across two ad formats.
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Audit these 78 ad posts using spend-performance evidence. Explain apparent waste, then propose three creative variations across two formats in this Slack thread. Keep proposed campaign changes separate from execution for our team to review.
Reach people exploring AI agents
See placements →What happened
Eric Siu describes a Slack ad-strategy workflow: Hermes pulls spend-performance data for a 78-post audit, explains waste and prepares three creative variations across two ad formats.
Reported by Eric Siu (@ericosiu)How the agent did it
- Request the 78-post ad audit through the shared Slack conversation.
- Pull spend-performance data for the audit.
- Ask the analytics agent to explain apparent spending waste.
Where a person stepped in
People initiate and extend the Slack conversation. The account does not establish campaign launches, autonomous ad edits, approval policy or measured return improvement.
Eric Siu describes his team using Hermes inside Slack for ad strategy. The account identifies a bounded request—a seventy-eight-post advertising audit—and a shared conversation in which another team member asks for refreshed creative ideas.
The agent pulls spend-performance data. An analytics agent explains where the team is wasting money, while a creative agent produces three variations across two ad formats. The useful sequence connects performance retrieval with a concrete creative output instead of asking for unrelated copy ideas without campaign context.
The public account does not name the losing ads, reveal spend figures or show that the proposed variations were launched. It also does not establish a measured improvement in return on advertising spend. Team members’ reported sense of working faster is subjective, so it is not treated as an independently measured productivity result.
Slack is the shared handoff surface: people can extend an audit conversation with a creative request, and specialised agents respond within that work context. The source does not describe a demonstrated approval policy or autonomous campaign edits; neither is invented here.
The original public X embed confirms the author, timestamp and matching opening about Hermes, Slack and the seventy-eight-post audit. Nous’s attributed first-person reprint supplies the analytics explanation and creative-variation details. This supports an executed team workflow while leaving campaign outcomes and independent quality verification unresolved.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: X · Eric Siu (@ericosiu)
Frequently asked questions
What makes the audit specific?
Siu identifies a 78-post audit with spend-performance retrieval, followed by three creative variations across two ad formats.
Were the variants launched?
The public account does not establish that the proposed variants were put into live campaigns.
Was the claimed speed improvement measured?
No. The team’s sense of being faster is a subjective report, not an independently measured result.
More details
Remaining steps
- Extend the conversation with creative-refresh ideas across two formats.
- Inspect the three proposed variations before further campaign decisions.
About this account
Hermes retrieves spend data and specialised analytics and creative agents explain waste and generate three variations across two formats within the shared conversation.
Firsthand account by Eric Siu (@ericosiu); original publication date verified against Official X public embed created_at 2026-05-29T22:48:13Z; matching author and opening cross-checked with Nous attributed reprint. Outcomes are the author’s reports, not independently reproduced. Suggested prompt written by Agents In Use; not the author’s original prompt. Detailed workflow claims are supported by the attributed first-person reprint at https://hermes-agent.nousresearch.com/docs/user-stories; the opened official X public embed contains only the beginning of the long original post.
Credit: Eric Siu (@ericosiu)
Source published 29 May 2026
Source checked 14 Oct 2026
Tools and topics
Hermes Agent · Slack · Ad spend-performance data · Analytics and creative agents
78-post ad audit · Performance-led creative refresh · Two-format ad variants
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.