AI-built email workflows became operational inside Vetano when the hiring platform needed to revise two welcome messages and protect its sending infrastructure from poor addresses. The result extended beyond faster editing. It created a controlled system for account notices, event outreach, reply handling and human approval.
Chris Fairley founded Vetano in early 2025 after a difficult job search convinced him that a strong resume no longer provided sufficient evidence of a candidate’s abilities. Resumes receive about seven seconds of attention, according to Ladders’ eye-tracking research. Vetano takes a different approach: applicants demonstrate their abilities on video, with identity verification applied to the footage and the references they submit. The company now has three employees and has begun attracting its first paying customers.
The operational lesson extends beyond hiring software. Vetano’s system shows how a small company can automate substantial outreach while keeping important decisions in human hands.
- Vetano expanded an account-message rebuild into event outreach and lead handling.
- Claude worked through Zapier’s Next-Gen Zaps rather than changing the existing automation directly.
- Chris reviewed test messages and approved workflow changes before they went live.
- The event campaign used 37 mailboxes across 13 domains, with daily bounce-rate monitoring.
- A second agent, Viktor, managed campaign operations and address checks, while Chris retained final approval over replies.
The welcome workflow was the control room
Vetano’s original process sent every new signup to Slack. A Zap then determined whether the account belonged to an applicant or a business and delivered the appropriate welcome email. When product changes altered both messages, Fairley estimated that careful manual editing would require two or three hours.
His first attempt to have Claude make those changes failed because Claude could not interact with the existing Zaps. A Zapier message then introduced Next-Gen Zaps and its connection with Claude. After Fairley pasted the prompt into it, the revised workflow was ready within minutes.
The rebuilt process went beyond changing two sentences. It skipped test accounts, prevented duplicate welcomes and checked addresses with ZeroBounce. It could continue sending if ZeroBounce was unavailable, use a dedicated Vetano address and include temporary information such as an expo note that later removed itself. Before a change went live, the Zap sent a test email to Fairley’s inbox for approval.
That last point matters. The workflow did not merely automate faster; it made failure behavior explicit. For a small team, that distinction is more valuable than speed alone because an automated mistake can spread quickly across every new account.
From onboarding to outbound campaign
The harder test was Vetano’s Platinum sponsorship of the Small Business Expo in New York that October. Before the show, Fairley reached out to nearly 9,000 registered businesses. Organizers supplied a new registration export every Monday, including about 85 new businesses per day.
Fairley initially considered sending from his own inbox, but email providers limit how much a single address can send. Exceeding those limits can damage delivery for everyone on the list, so Vetano prepared a fleet in Instantly: 37 mailboxes spread over 13 domains, which received three weeks of preparation before use.
Viktor, an AI agent operating alongside Claude, monitored the fleet daily and flagged any mailbox whose bounce rate crossed 3%. A weekly reconciliation paired each incoming registration file with Vetano’s master spreadsheet in Google Sheets. Claude assigned businesses a score from T0 to T4, excluded shared role inboxes and prepared operating instructions for Viktor.
Addresses were checked with ZeroBounce before sending. When an address contained an obvious error, such as a misspelled Gmail domain, Viktor corrected it instead of dropping it. ZeroBounce kept 2,874 dead or risky addresses off the campaign list. Between September 29 and October 14, the system dispatched 19,569 messages; 73 failed, producing a reported rate of 0.37%.
The sequence illustrates why automation must address list preparation as carefully as sending. A large volume alone does not create a successful campaign. The system also needs deduplication, audience rules, address verification, domain corrections and thresholds that allow a weak mailbox to be paused before it affects the wider fleet.
Why the human approval points remain
The most useful part of Vetano’s setup is where automation stops.
The welcome workflow sends a test email to Fairley’s inbox. He reads it and approves the change. For the strongest prospects, Claude drafts a personal email, Zapier moves the draft into Gmail, and Fairley reviews and sends it. Replies are posted to Slack with the sender’s title, industry, tier and phone number. Claude drafts an answer in the channel, but Fairley sends it.
When Fairley marks a lead as interested in Instantly, a workflow Claude built adds that person to HubSpot. An opt-out is placed on a blocklist the same day. Every link is tagged, giving Fairley visibility into which emails drive traffic to the Vetano page.
This is not a fully autonomous outreach machine. It is an assisted pipeline with explicit gates. That distinction is essential for leaders considering similar systems. Automation handles comparison, deduplication, routing, drafting, tagging and checks. A person decides what goes live, which high-fit prospects receive personal attention and how a reply is answered.
The trade-off is clear: Vetano gains speed and consistency without surrendering control of its reputation or contact list. The cost is a system that must be maintained, monitored and updated as the product, audience and campaign change.
A decision checklist for project teams
Teams building a similar system should work backward from approval and failure handling:
- Define the final human approval point before connecting any sending tool.
- Separate the applicant and company paths, then test every audience-routing rule.
- Decide whether a campaign should continue, pause or seek manual review when a verification service fails.
- Exclude test accounts and prevent anyone from receiving a welcome message twice.
- Remove shared role inboxes and reconcile each new import with a master list.
- Correct clear address errors while blocking dead or risky destinations.
- Prepare sending mailboxes before beginning a high-volume campaign.
- Monitor bounce rates daily and define when a mailbox should be paused.
- Tag links and route replies with enough context for a human response.
- Send opt-outs onto a blocklist immediately.
What to watch next
Vetano’s system offers a practical lesson for agile project management: AI-built email workflows are most useful when they improve controlled handoffs, not when they remove every human decision.
The commercial test is whether the first paying customers support the idea of hiring without conventional resumes. The operational test is whether Vetano can preserve bounce rates, response quality and human review as volume grows.
That possibility is why the automation exists now rather than after a surge. For Vetano and similar teams, the winners will not be those that simply send the most emails. They will be those that know when to automate, when to pause and when a person should still have the final say.
Key takeaways
- Put approval and failure handling into an AI workflow before optimizing for speed.
- Combine deduplication, domain checks and bounce monitoring rather than relying on one control.
- Keep high-value messages and replies under human review even when drafting is automated.
Frequently asked questions
What did Vetano automate first?
It initially rebuilt two account welcome emails for applicants and businesses. It later extended the automation to expo outreach, list preparation and reply handling.
Did Claude send campaign emails without human approval?
No. Claude prepared campaign work and drafted messages and replies, while Fairley reviewed and sent personal emails and responses.
How did Vetano protect its email mailboxes?
It used 37 mailboxes across 13 domains, spent three weeks preparing them, monitored bounce rates daily and flagged mailboxes crossing a 3% threshold.
What was the reported bounce rate?
The campaign dispatched 19,569 emails from September 29 through October 14. Seventy-three failed, resulting in a reported rate of 0.37%.