Which Angle Actually Sells Your Product?

01 /The Concept

When a company launches a new product, figuring out how to position it is genuinely hard. A single product can have:

- multiple features the team believes are important,

- multiple potential target audiences,

- multiple ways to frame the same value proposition.

Validated helps companies cut through that ambiguity – not through surveys or focus groups, but through real ad performance data.

The process: upload a product description to the platform, and the AI engine automatically generates and runs a range of ad variants. Different features, different offers, different creative angles, different audience targets. Then it collects click data. What people actually click is the signal that matters.

A worked example: a mobile wellness app offering workout routines, mindfulness exercises, nutrition advice, and healthy recipes. Validated's engine created 80 ad variants, spent $418 running them, generated 66,000 impressions and 993 clicks – and delivered a set of concrete recommendations.

The strongest value proposition turned out to be "reduce illness, injuries, and other health problems." That message resonated roughly equally across genders. But it landed very differently across age groups: almost no interest from people in their thirties and forties, and strong engagement from 18–24 year-olds, women over 55, and men over 65.

Validated has also launched a beta feature for automated competitive analysis. Describe your product, and the AI identifies who your competitors are, then pulls traffic data and visitor demographics. The detailed breakdown per competitor covers geography, traffic sources, search keywords, ad copy performance, and referring sites – essentially a lighter version of Similarweb, but scoped to your specific competitive set and priced accordingly.

Founded only in the summer of 2023, Validated has already acquired its first paying customers and just closed $1.1M in its first funding round.

02 /Why It Matters

Someone will always say that Validated is "just another wrapper around ChatGPT" (Though they may be using a different model entirely)

And at a component level, they're not wrong – AI-powered ad generation, automated campaign execution, and performance analytics all exist separately. Nothing here is technically novel.

But a product isn't its components. A product is the ability to assemble known building blocks into a new configuration that solves a specific user problem. That distinction matters enormously.

The point is worth spelling out. The "GPT wrapper" criticism claims these startups invent nothing – they take a commodity AI backend, add a layer of UX, and sell it. But the core startup skill isn't invention; it's building products people want. Invention is sometimes required, but it's not the goal – it's an occasional burden.

The smarter frame: applaud startups that build genuinely useful products from commodity components. They're doing the hard thing – finding what people will pay for – not the impressive-sounding thing of engineering something novel. Building "GPT wrappers" quickly and cheaply is actually excellent training for product thinking, because all the effort goes toward finding the right problem, the right user, and the right framing. And occasionally, what starts as a quick wrapper turns into a real, defensible product that compounds over time.

Beyond the philosophical point, AI-powered product insight is emerging as a distinct category.

Outset, from a recent Y Combinator class, was [covered here](/review/insajty-dvigatel-biznesa) in November. They built an AI interviewer that autonomously conducts user interviews and extracts structured insights – no human moderator needed. Outset raised $4.9M.

Knit, [covered](/review/budut-li-oni-jeto-pokupat) in the summer of 2022, started with a video interview platform focused on Gen Z users, then added AI-driven analysis in 2023, pushing in a direction similar to Outset. Knit has raised $5.6M.

03 /Opportunities

Start by stopping the reflexive dismissal of AI wrappers. Many excellent products are, at a technical level, a set of fairly simple operations on top of a relational database – but nobody calls them "MySQL wrappers"

The practical implication: consider actually building one. The exercise forces you to spend all your effort on the question that matters most – what to build, for whom, and in what form. That discipline is the best training for product thinking available right now. And the result might be a real, saleable product with room to grow.

More broadly, AI-powered platforms for understanding demand, capturing product feedback, and deriving actionable insight are coalescing into a real category. Startups are being built from scratch in this space, and existing players are evolving toward it.

Zappi, [covered](/review/v-10-raz-bystree-i-deshevle-v-100-raz-shire) at the end of 2022, is one of the larger players here – a platform for finding product insights that raised $192.7M, including $170M in a single round once it leaned more heavily into AI. The category has room for many more entrants.