First builds in one afternoon
2 October 2026: In one afternoon, the CEO installed the tools on his own laptop and ran his first builds. Independent practice and a return session were the agreed next steps at that point.
Evidence & practical questions
Two agencies. Building independently.
One Dublin afternoon is documented. Later feedback from two anonymous specialist agencies reports independent builds and daily CEO use.
Feedback from two CEOs · Both agencies anonymous
What the two CEOs reported.
Both specialist agencies had two to four participants, two full-day sessions and agreed Q&A between sessions. The shared outcomes below were reported in the owner’s follow-up conversations with both CEOs.
A second anonymous specialist agency
The second agency also worked with tools suited to its own technology environment across two full-day sessions and independent practice.
“It is not what the technology can do. It is what I can think of that brings me value, and the technology does the rest.”
These are follow-up accounts from the CEOs, not measured performance results.
How we work responsibly
Recruitment decisions stay with recruiters.
We teach the practical controls alongside the build process. Human review is essential, but it does not replace the agency’s data-protection and regulatory responsibilities.
Approved tools and data
Start with suitable permitted material. Inspect what an agent can read, what it sends to external services and which permissions are enabled.
Plan, test, verify
Ask the agent for its plan before allowing actions. Test with representative examples and inspect source information before sharing drafts.
No automated candidate decisions
AI can help recruiters design sharper searches and filters across permitted data sources. Recruiters inspect results and make every decision. No automated candidate ranking, rejection or outreach.
Complex integrations or production use may need technical, security or legal review beyond the learning builds.
Questions owners ask
What agency owners ask.
Our recruiters aren’t technical. Can they really build this?
No coding background is needed. We look for curious people who know the desk and will practise between sessions. They learn to brief AI coding agents in plain English, inspect their work, test results and correct errors.
Is this right for our agency, and who should take part?
The approach suits agencies that can nominate two to four curious consultants or operations colleagues, back them with practice time and use suitable approved data. We discuss the starting workflow and fit in the first conversation; there is no fixed staff-count threshold.
Our ATS already has AI. Why build anything else?
Keep using the AI in your ATS where it works well. We focus on agency-specific work your existing systems do not fully cover, such as a client’s report format or desk research. If an ATS feature solves the problem, we will say so.
What if our ATS or database cannot connect to the AI tools?
We start with workflows your agency can support using approved tools and permitted data. Not every system or integration will suit a first build. More complex connections may need technical assessment or specialist help.
Can we get a quick win in the first full-day session?
We aim for a quick win by starting with a small, familiar desk task. Your champions begin building and testing a real first version alongside us—not just listening to presentations. In the documented Dublin afternoon, the CEO installed the tools and ran his first builds. A finished, production-ready tool on day one is not guaranteed.
How much time will our people need to practise between the two days?
Your champions need protected time to experiment, test and improve their work over at least a couple of weeks. We agree Q&A via email, chat or a WhatsApp group between the two full-day sessions, and discuss what is practical for your agency.
How will we know our champions can build without you?
We agree an independence check before starting. A champion takes a new desk problem, builds a useful first version, tests it on examples, catches and corrects errors, and explains the method while we watch rather than drive.
How do you handle candidate data and GDPR?
We start with agency-approved tools, permitted data and agreed access controls. Champions inspect what an AI agent can read and send to external services, then test and verify outputs. AI may help recruiters organise information and design sharper searches, but never rank, reject or contact candidates automatically. Recruiters review the evidence and make every decision. The agency remains responsible for its data-protection and regulatory obligations.
What happens after I send the email?
The link opens a draft to Shay and Lia. Send it with any times that suit you, and we’ll reply to arrange a 20-minute conversation. No preparation is needed. We will discuss the workflow that costs your desk most, whether it is a suitable first build and what your champions would learn. Scope and commercial terms are discussed in that conversation.
Will we still depend on Domino Effect Lab to change our tools?
The goal is low dependency. Champions learn to build, test and improve suitable tools independently of us, without asking us or an outside developer for every small change. Complex integrations or production systems may still need specialist help. Wider company rollout is a separate later conversation, not part of this programme.
What would your people build first?
Spend 20 minutes with Shay and Lia on the one workflow that costs your desk most. No preparation needed. We’ll explore whether it’s a good first build and what your champions could learn.
Contact Shay and Lia: info@dominoeffectlab.com
Opens an email draft to Shay and Lia. If it doesn’t open, email info@dominoeffectlab.com; we’ll reply to arrange a time.