Our October 2025 test found Google Pomelli useful for fast, on-brand creative starting points—but not a replacement for growth strategy. It helped us move from a public brand website to campaign ideas and editable assets quickly. It did not decide which audience to pursue, what the business could afford to pay for a customer, how a campaign should be structured, or whether the resulting traffic was commercially valuable.
That conclusion is deliberately time-bound. Pomelli was an early public-beta experiment when we tested it, and its capabilities may have changed since then.
Evidence note — updated August 6, 2026: This article documents one Shopping Ads Solutions hands-on test conducted on October 29, 2025. We used the public website of an anonymized swimwear brand and reviewed the brand profile, campaign suggestions, generated assets, and edits required. No confidential performance data was used. One qualitative test is not a benchmark or a claim about Pomelli's current feature set.
What Google said Pomelli was designed to do
In its October 28, 2025 launch announcement, Google described Pomelli as a Google Labs and Google DeepMind experiment for small and medium-sized businesses. At launch, the workflow had three stages:
- Analyze a website and create a “Business DNA” from its visible brand cues.
- Suggest campaign ideas grounded in that profile.
- Generate editable branded assets for channels such as social media, websites, and ads.
That was also the useful way to evaluate it. We did not ask whether it could “do marketing.” We asked which part of our creative workflow it could compress—and which decisions still sat outside the tool.
What we observed in the test
We entered the public website of a swimwear brand. Pomelli identified several core visual and verbal cues quickly: the main colors, typography, existing imagery, and recurring messaging themes.
That first step was the strongest part of the test. It converted a public site into a workable creative starting point in minutes. For an agency, that can reduce the time spent collecting basic cues before ideation begins.
The campaign suggestions were also useful against a blank page. They gave us angles to react to, reject, combine, or refine. The generated visuals were not final assets, but several were credible enough to begin a conversation with a strategist or designer.
We still edited the copy, sharpened the call to action, and rejected ideas that did not fit the intended campaign. Pomelli accelerated the first draft; it did not remove the need for judgment.
What the tested version did—and did not—replace
| Observed in the October 2025 test | Not observed in the tested workflow |
|---|---|
| Extracting visible brand cues from a public website | Defining the commercially valuable audience |
| Suggesting campaign concepts | Setting an allowable CAC, CPA, or contribution target |
| Generating editable creative starting points | Building and managing Google or Meta campaigns |
| Helping a team move past a blank page | Measuring traffic quality, incrementality, or profit |
| Speeding up early creative discussion | Deciding what to test next from business outcomes |
This distinction matters because creative production and growth strategy are not the same job. A polished asset can still be shown to the wrong person, attached to the wrong offer, or evaluated against the wrong conversion signal.
Did the test make Pomelli a threat to agencies?
It made basic creative production easier; it did not make accountable growth strategy unnecessary.
That is a threat to an agency whose only value is producing simple first drafts. It is useful leverage for an agency responsible for the decisions around those drafts: audience, offer, unit economics, campaign architecture, measurement, and iteration.
Our practical response was not to resist the tool. It was to place it in the right part of the process:
- Use it to accelerate brand extraction and early concepts.
- Review every idea against the campaign's audience and commercial purpose.
- Edit or reject the output before production.
- Measure the published work through business outcomes, not visual polish alone.
That process is the point. AI can shorten a production step. It does not automatically improve the decision that the production is meant to support.
The verdict from this test
Pomelli was a useful creative assistant in the version we tested. It helped us understand a public brand surface, generate starting points, and reduce early creative friction. It was not the strategy surrounding the creative.
You can watch the original hands-on walkthrough on YouTube. For a separate example of why the advertising dashboard cannot be evaluated without the business outcome, read what we found when more paid traffic produced weaker orders.
Shopping Ads Solutions uses tools like this inside a broader creative-development process. If you want us to examine how your creative, traffic, measurement, and commercial economics fit together, request a paid growth audit. If we subsequently work together, the audit fee is credited toward the agreed contract.
About the author: Michael Chachashvili is a founder of Shopping Ads Solutions and works on paid acquisition, measurement, and profitable growth systems for eCommerce and service businesses. Published October 29, 2025; updated August 6, 2026.