Rethinking AI Progress: Governance and Trust Over Metrics

Rethinking AI Progress: Governance and Trust Over MetricsRethinking AI Progress: Governance and Trust Over Metrics

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

Irvan Gerhana

• Jun 25

I think we've been measuring AI progress the wrong way.

We compare:

• GPT vs Claude • Closed vs Open Source • Model benchmarks • Token speeds

But almost nobody asks:

Who owns the business logic?

How is AI governed?

Can decisions be audited?

Can outputs be reproduced?

Can the organization actually trust the system?

Those questions become far more important once AI moves from demos into production.

That's exactly what I've been exploring while building enterprise AI systems.

I shared my thoughts here:

https://dev.to/uigerhana/the-ai-model-isnt-your-competitive-advantage-1ib7

Data Science

dev.to

The AI Model Isn’t Your Competitive Advantage.

Your Governance Model Is. Every time a new AI model is released, the same conversation…

David Akanbi

• Jun 27

97% of website visitors leave before a human can follow up.

I just published the full case study on an autonomous sale engine I designed to replace the entire top-of-funnel sales motion. No human required.

Here's what made it hard to design:

→ Two completely different audiences. One needed to feel seen in real time. The other needed to feel in control at 11pm. → The machine had to feel legible not like a black box. That meant designing the Sandbox: a split-screen interface where you can change the AI's knowledge and watch it respond live, before a single real customer sees it. → Every screen had to prove ROI, not just look good.

This isn't a cool UI case study. It's about how design decisions directly affect whether an AI agent earns trust or loses it on first use.

Full case study on Contra 👇 Check it out