The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Most AI builds break at the wiring, not the model: no retries, no log of what the model actually did, and no clear line for what the agent may do without a human.
I build that layer. My own products run on it -- an ordering platform where an AI agent handles real orders, and a coding agent that writes no file until someone approves the change.
The question worth answering first is what your agent should never decide on its own.
For the coding agent, I'd bind approval to both the proposed diff and the file's current version. If another edit lands between review and write, does it stop and request a fresh review?
For me the line is anything irreversible or anything that costs money: refunds, deleting data, messaging customers in bulk. Below that line the agent can act, as long as every tool call is logged with its input and result, so you can replay what it actually did. The detail I'd...
This is exactly right — the model is rarely the problem, the wiring is. Approve-before-any-write on the coding agent is the detail that matters. A file write looks small until it's the wrong file. My line building agentic systems is similar: anything irreversible, or anything a...
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
AI Automation SaaS Landing Page
A modern AI automation SaaS website design focused on simplifying workflows, automating repetitive tasks, and helping businesses work smarter. Designed with a clean interface, intuitive user experience, and a scalable visual system to communicate the value of AI-powered automation.
Amazing
Check it out
Just published: 0-1 design for an Energy Management System (EMS) ⚡
Designed a desktop platform monitoring wind turbines, PV parks, and battery storage across multiple sites. Keeping the interface mostly white, so green and red status indicators do the talking.
Operators managing critical infrastructure need to assess system health instantly. No visual noise, no cognitive overload. Just unmistakable state communication.
Cool work 👌
🚀 New project added to my Contra portfolio - Dynamiq AI!
I’ve just added Dynamiq AI to my portfolio, a project in the AI space that gave me another opportunity to explore the intersection of technology, product design, and user experience.
When designing digital products, I believe the goal goes beyond creating an attractive interface. It’s about making complex ideas easier to understand, helping users navigate the experience intuitively, and presenting information in a way that feels clear and purposeful.
🤖 What makes the product interesting:
→ Building AI agents and automating complex workflows
→ Connecting AI to business tools, knowledge bases, and data
→ Managing AI agents with security controls and human approvals
→ Monitoring performance and evaluating results
These are the kinds of challenges that make product design exciting for me — turning complex functionality into experiences that feel simple and natural to use.
✨ Always happy to explore new ideas, experiment with different approaches, and share more of my work along the way.
🔗 Check out Dynamiq AI: https://www.getdynamiq.ai/
The interviews with engineers and compliance reviewers caught my attention. What did the reviewers need to see before approving an agent action, and how did that change the approval screen?
