Fintech performance video in UAE and KSA: four years of paid social demos for Pemo

The challenge

Spend management is a hard category to advertise. There is nothing to show — the product is a card and a dashboard, and neither photographs well. The buyer is a founder or a finance lead running a 10-to-200-person company, and they are not browsing for accounting software on Instagram. You are interrupting them, which means the first two seconds carry the whole ad.

The second problem is fatigue. Performance creative in a market the size of the UAE saturates fast. A hook that works in week one is being served to the same audience for the fourth time by week three, and CPL drifts up whether or not anything about the offer changed. Volume alone does not fix this — twenty variations of a tired idea fatigue at the same rate as one.

The third is that the UAE and Saudi Arabia are not one market. What reads as credible in Dubai does not automatically read that way in Riyadh, so the two were run as separate creative lines rather than one set stretched across both.

Two markets, one engine

The UAE and Saudi Arabia were run as separate creative lines. Casting, setting and the way a business problem was framed were decided per market before anything went into production.

What the two shared was the engine underneath — a hook library tracking which openings had run, where, for how long, and against which audience. When a hook started to fatigue in one market, we knew whether it was genuinely spent or simply had not been tried in the other yet.

The AI engine, from 2025

The first AI creatives on this account went live in May and June 2024, and they were rough — a stiff synthetic presenter, illustrated stills standing in for scenes, a look that dated within weeks. They ran, they taught us where the tooling broke, and they are the honest starting point of everything that followed.

In 2025 we stopped treating AI as an occasional trick and built a production engine around it. That became the single biggest differentiator on the account. A concept no longer had to survive a budget conversation about casting, crew, location and a shoot day, so a scenario that would once have been dismissed as unaffordable became routine. The cost of a finished video fell to a fraction of what the same idea would have cost as a live-action shoot, and the range of ideas we could put in front of a buyer widened at the same time.

It also removed the slowest link in the chain. There were no schedules to align, no reshoots to negotiate, no talent to rebook when a hook needed one line changed. A revision that used to mean another shoot day became another render.

Three to four finished videos in five days

Concepts that needed a real person, a real office, a real reaction were shot. Concepts that needed a talking corporate card, a warehouse of receipts, or a scenario nobody would fund a shoot for were generated. Most batches mixed both, and the seam was deliberately invisible.

The commercial point was turnaround. A brief on Monday was a delivered set of three to four finished videos inside five working days, which made the feedback loop a week rather than a quarter. Four years inside a CPL band was not the result of one great ad — it was the result of never being more than a week away from the next test.

Outcomes are described in the framing agreed with Pemo. Absolute demo counts, spend levels and CPL values are covered by our confidentiality obligations and are not published — for this client or any other.