How Should You use AI for Ad Design? Perfogro Ltd's Experience

Generative tools made it cheap to produce a banner ad in under a minute. They did not make it easy to produce a good one. That gap, between speed and quality, is where most marketing teams are losing time without realizing it. Perfogro Ltd has spent the past year watching this play out across campaigns, internal experiments, and client work — and the team believes the conversation about AI in ad design has skipped over the part that actually matters: knowing which steps to hand over and which to keep human.

NVIDIA's 2026 State of AI research found that 88% of enterprises report AI has helped increase annual revenue, with 53% citing improved employee productivity as the largest operational impact. Those numbers don't mean every workflow benefits equally, though. In creative production, the wins are real but narrower than the hype suggests — and the failures are quieter than the wins.

Myth 1: "AI Can Just Write the Brief"

The most frequent error that Perfogro observes is that the designer does not formulate an idea clearly enough and gives the generative tool vague prompts for creating the final visual. The outcome might even turn out to be well-presented, but it will hardly solve the particular marketing task. AI speeds up implementation; it does not substitute strategy. In practice, the brief still has to specify:

• The audience segment, in detail, beyond demographics.
• The single action the ad should drive.
• The visual or tonal cues that are off-limits for the brand.
• The performance benchmark the asset needs to beat.

Perfogro suggests treating AI like a junior designer with no context — capable, fast, but in need of clear direction.

Myth 2: "More Variations = Better Results"

Generative tools make it trivial to produce 40 versions of a single ad. Many teams take that as a win. Perfogro Ltd believes the opposite is usually true: too many variations dilute statistical signals and slow learning. The Perfogro team highlights that the goal of A/B testing is to isolate what works, and that gets harder, not easier, when the test set explodes.

A more disciplined approach:

• Cap initial test rounds at 3–5 meaningfully different variations.
• Vary one element at a time — headline, image, or call to action — not all three at once.
• Use AI to expand the winners, not to create more candidates upfront.

Where AI Actually Earns Its Place in Ad Design

Perfogro uses AI most heavily in three specific stages, and the reasoning is clear:

• Concept exploration.Generating rough visual directions early, before a designer commits time to refinement. Speed matters more than polish here.
• Localization at scale.Adapting a winning creative into ten markets with appropriate copy, imagery, and cultural cues — a task that previously required weeks of coordination.
• Iteration on proven creative.Producing controlled variations of an asset that has already performed well to extend its runway before fatigue sets in.

In each case, AI shortens a step that humans were already doing — it doesn't introduce a step that no one was doing before. That distinction matters because adding AI to a workflow that wasn't working tends to produce faster, worse output, not better.

Evaluating that output requires more than impression counts, which is where advanced attribution — Perfogro Ltd's framework becomes load-bearing: AI-generated creative produces volume, and volume without proper attribution looks like performance when it isn't.

Where AI Quietly Hurts Ad Design

Three patterns have a way of undermining creative quality:

• Generic visual language.AI image tools trained on the open web tend to converge on similar aesthetics. Brands that lean too heavily on raw outputs start to look like every other brand using the same tools.
• Copy that sounds confident but says nothing.Generative text fills space convincingly. It rarely makes a sharp claim or takes a position, both of which are what make an ad memorable.
• Loss of institutional taste.When designers stop drafting from scratch, the team's shared sense of "what good looks like" can erode. Rotating manual creative exercises into the workflow helps preserve craft.

A Practical Workflow Perfogro Suggests

For marketing teams looking to integrate AI into ad design without losing rigor, this sequence works:

1. Brief first, tools second.The strategist writes the brief without AI input.
2. Concept exploration with AI.Generate 8–12 rough directions, then narrow to 2–3 with the designer.
3. Human refinement.A designer takes the selected direction and develops it into a final creative.
4. Variation generation with AI.Once a hero asset is approved, AI extends it into format, language, and audience variations.
5. Performance review.Results feed back into the brief library — what worked, what didn't, and why.

This sequence preserves the two stages where human judgment matters most: strategy at the start, and review at the end.

What Companies Should Measure

Speed of production is the easiest metric to celebrate and the least useful one to optimize. Better questions to track:

• Click-through and conversion rate of AI-assisted creative versus fully manual creative, controlling for placement.
• Time from brief to first live test (not first draft).
• Creative fatigue curves — how long an asset performs before requiring refresh.
• Cost per acquisition across creative cohorts.

Without these comparisons, "AI sped things up" can mask the fact that performance held flat or declined.

The Underlying Principle

The use of AI in ad creation is best applied as a force multiplier to augment creativity in existing skill sets, rather than a replacement. Teams with well-defined briefs, KPIs, and creative vision are able to capture the most benefit from AI. Teams with the expectation that AI will cover for deficiencies in any of those skill sets create more work that is of poorer quality and with less learning.

This concept is echoed in McKinsey's AI guardrails research, which finds that companies leveraging structured review and assessment alongside their use of generative AI experience better results than companies focused solely on speed. Perfogro employs the same philosophy within its creative process – all AI-generated materials pass through a human review phase to ensure a strategic, brand-aligned performance hypothesis.

Final Thought

Used effectively, AI frees up ad designers' minds by reducing the time they spend doing tedious tasks. When used ineffectively, however, it results in an endless supply of slick but ultimately forgettable ads. Businesses need to know what outcome they wish to achieve prior to choosing the tools they will use to achieve it. The discipline associated with AI is more important than the technology itself, and it is the businesses that use AI as their creative partner rather than as a creative replacement that have a real edge, according to Perfogro Ltd.