Human in the loop vs ad bots - the line we draw for fintech

We get asked one question a lot on sales calls right now:

“Will your system just run our ads on its own?”​

The honest answer is no. And that is a deliberate choice, especially for fintech.

We have all seen the screenshots:

• tools promising “fully autonomous” campaign management
• threads about accounts getting banned after a bot started taking rapid fire actions
• people exporting data into random LLMs and pushing decisions back into Google or Meta with no guardrails​

In low risk categories, some of that might be acceptable. In lending, where you rely on these platforms as core acquisition channels, we think the risk is too high.

What makes fintech different

Two things make us more conservative here.

First, the stakes. A bad decision is not just wasted spend. It can shift your risk mix in ways that hurt collections later. Performance and underwriting are more connected than most dashboards show.

Second, platform rules. Meta and Google are already on alert for suspicious, high frequency automated actions. If a third party tool starts hammering their APIs in a way that looks like a bot making uncontrolled changes, your account can and will come under review.​

We did not want Thirdi to be the reason someone lost their main growth channel.

What “human in the loop” means in our world

For us, humans in the loop is not a buzzword. It is a design choice.

The agents in Thirdi:

• read your data across Meta, Google, and your CRM
• spot waste, fatigue, and odd patterns
• prepare clear action suggestions: what to pause, what to scale, what to test

They do not hit the “publish changes” button for you.

Your team still:

• reviews suggestions
• applies judgement on context, brand risk, and internal constraints
• pushes the actual changes in the ad platforms via official, supported paths

This slows things down by a few clicks, but it keeps you on the right side of policy.

Why we believe this is the right tradeoff

Could we wire Thirdi to take full control of your campaigns? Technically, yes.

Would we trust that model for a lender spending crores on Meta and Google, operating in a regulated market, and already nervous about bans? We would not.

By keeping humans in the final loop, we get a few clear benefits:

• no surprise policy violations from aggressive bots
• clear accountability on why a change was made
• more trust from performance, product, and compliance teams who need to live with the outcomes

We still give you the time savings and the sharper decisions. We just stop one step short of saying, “let the machine run your budget while you sleep.”

For some brands, that might sound less sexy. For fintech brands who know what a suspended account feels like, it is usually a relief.