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
Research OS — Evidence-Driven AI Research Workflow
The problem
AI makes research dramatically faster, but speed creates new problems: duplicated work, unclear authority, unsupported conclusions, model drift, expensive context reconstruction, and findings that become detached from their evidence.
What I built
Research OS is a structured AI-assisted research workflow designed to move from question to evidence to decision while preserving state and provenance.
The system combines:
exploratory research
model routing
structured evidence capture
machine-readable state
staged escalation
independent review where useful
durable research artifacts
explicit closeout and recovery
Local models are used where they are sufficient, with more capable systems deliberately introduced when the task justifies them.
Result
The workflow has been applied across software architecture, technical research, security research, market analysis, and other long-running investigations.
Its purpose is not to automate judgment away. It is to make research easier to resume, inspect, challenge, and reuse.
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
Weekend fun 😺. The new MacBook Pro has a Dynamic Island, and I kept wondering: what if that little pill was where your AI agent lived?
Try it live: https://lab.behzad.me/agent-mew/
So I spent the weekend prototyping Agent Mew. You drop a file on the notch, give it a task, and it quietly works up there while you keep doing your thing. It shows progress in the wings, asks a quick question when it needs one, and always waits for your OK before doing anything that leaves your Mac, like sending an email.
A few things I explored:
- Progress that stays out of the way, readable at a glance
- Questions as one tap, not a chat window
- Approval cards for anything with real consequences
- Paused and stuck states that tell you what to do next
- A small animated cat, because agents should have some personality
Approval cards for anything that leaves the Mac is the right line to draw. Does the card show the full email before you tap OK, or a short summary? I'd lean toward full text for anything going to another person.
Designed an AI research workspace where the agent plans the work and fills a sheet with cited sources as it goes.
Same screen, two modes. Which one would you work in all day? 👀
13 voted
72%
5 voted
28%
18 votes
Closed
Keeping the research plan and cited results side by side is a useful way to make the agent’s work easier to inspect. I’d lean toward light mode for reading the table in daylight. How do you distinguish a verified source from one that still needs review?
Your Business Doesn't Need More Tools. It Needs Better Systems.
Most businesses don't have a software problem. They have an integration problem.
Their CRM is disconnected from their database. Their spreadsheets require manual updates. Their reports take hours to prepare. Their teams waste valuable time moving information between platforms.
That's where I come in.
I'm a Full-Stack AI Engineer specializing in building connected, intelligent business systems.
What I can build for your business:
CRM & ERP Integrations — Connect business platforms, databases, and third-party services through APIs.
Workflow Automation — Eliminate repetitive processes with intelligent automation.
Custom Dashboards — Turn scattered business data into actionable insights.
Data Pipelines — Build reliable systems for collecting, processing, validating, and synchronizing data.
AI-Powered Solutions — Integrate LLMs, intelligent agents, and automation into existing workflows.
Custom Software — Build complete web applications and business platforms from architecture to deployment.
I've independently developed complex digital platforms, including government-level systems, international applications, and AI-powered products.
My approach is simple:
Understand the business problem. Design the right architecture. Build the complete solution. Make it reliable.
I'm interested in working with founders, agencies, and businesses looking to modernize their operations and build software that actually solves problems.
Have a complex workflow or an integration challenge? Let's build something that works.
#FullStackDevelopment #AIAutomation #CRMIntegration #WorkflowAutomation #SoftwareEngineering #AIEngineering #Contra
When a CRM-to-database sync fails halfway, how do you handle it: idempotent retries, or a reconciliation pass afterward? That's usually where integrations start to drift.