Viktor vs Wispr Flow: Two Opposite Ads Playbooks, Head to Head

Viktor has 828 live Meta ads. Wispr Flow has 1,226. Two of the most copied AI ad accounts on the internet, both shipping more than ten new ads a day.

And they run on opposite operating systems.

Viktor built a factory. It shoots in-house, finds a winner and clones it under a dozen ad objects. Wispr built a marketplace. It almost never clones, because a new creator file costs it almost nothing until that file converts.

I already did the full teardown of each - Viktor's ads strategy and Wispr Flow's ads strategy. This is the two of them side by side. Same questions, two answers, and what to steal from each.

Same feed, same 9:16, same ten-plus a day. You can tell them apart from across the room.

1. The creative supply chain

Viktor clones. Wispr doesn't.

Viktor's most-cloned file runs under eleven ad objects. A walking shot, a Slack notification, zero cuts. When it works they don't shoot another one. They push the same file into more ad sets.

Wispr's 466 files sit behind 574 ads. 375 of them run once and never again. Nobody is cloning winners because nobody needs to. The next creator is already uploading.

Who actually makes the ads

Viktor pays creators up to $10,000 a post in cash, or 50% more in product credits. Most take the credits. A $15,000 credit payout costs Viktor inference, not $15,000, and the creator ends up a heavy user who makes the next video out of real work.

Wispr doesn't run a creator program at all. Its help centre says so. 52% of its live ads carry a Trybe creator ID in the URL - a marketplace where creators shoot for brands and get paid on what the video converts. The ID is the receipt.

Viktor pays for distribution out of margin. Wispr pays for it out of results.

Both are cheaper than cash up front. That is the whole trick.

2. The first six seconds

Both accounts open with text before they open with movement. That is where the similarity ends.

Viktor's caption is a subtitle the size of a headline. It sits on the speaker's face and it is there before they say a word.

Wispr's caption is a TikTok title card. Small, high, written by the creator. "POV: you quit typing." "Tech Luxury List." "Productivity hack for ADHD creators."

Neither one cuts. Median one cut in six seconds, both of them. Every playbook says cut fast. The two biggest accounts in the category hold the shot.

Viktor's most-cloned file, beat by beat:

Wispr's most-cloned live file, beat by beat:

Viktor puts the product on screen before the first word. Wispr puts a person in two rooms and one word on screen at a time. One is a demo with a face on it. The other is a creator video with a product in it.

The hooks, side by side

Viktor runs eleven hook formats at once. Tap any cell to watch the ad.

Wispr runs thirteen.

Viktor's list reads like a production schedule. Notification over life, prop in hand, money counter, tier list. Formats you can brief.

Wispr's reads like a creator's camera roll. Two-hander, POV card, developer screen, listicle, coined word. Formats a creator already makes.

The statics are both somebody else

Here is the one place they agree completely. In both accounts the video is messy and the static is art directed, and the two never meet.

Viktor's statics are a 1950s illustrated campaign, gold objects on a fork and a lock screen at 7:02am. Wispr's are serif type on cream and forest green, a desk at night, a message bubble with the ums struck out.

The video says someone like you uses this. The static says a company you can trust made it. Both accounts need both sentences.

3. What stays still

You can't test everything at once. Each of them picked something to hold still.

Viktor holds the headline. "Not a Chatbot. A Coworker." runs on 119 ads. "The AI employee that lives in Slack" on 109 more. 26% of the account says one of two things, and everything around it changes.

Wispr holds the numbers. 494 headlines across its live ads, the top one on just 34. But every creator says "4x faster than typing", "2x more accurate than Siri", "zero edits" in their own words.

If your positioning is new, do it Viktor's way. You need the market to learn one sentence. If your category is already understood - everyone knows what dictation is - do it Wispr's way. Let a thousand people say it like themselves.

4. Campaign structure and naming

Two campaign names. Read them as a schema, not a label.

Both use the same o prefix for the optimisation event. Same convention, two companies, two continents.

Viktor's name carries the bidding logic. Value bidding with Meta holding budget on cold traffic. Volume bidding with a human holding budget on the warm pool. Two philosophies, one account, readable at a glance.

Wispr's name carries history. S4 means three structures came before it. A second link resolves to EvReset_PP - an evergreen reset, named as one. When they rebuild, they say so.

Where the record lives

Viktor stamps every ad with an eight-field creative record: creative ID, test variant, production batch, angle ID, in-house or creator, the hook word for word, concept and funnel stage, launch date. It still reads the same a year after the campaign is deleted.

Wispr stamps every ad with a Trybe hash and a dub_id. Less about the creative, more about who gets paid. Creator pages carry their own promo code and UTMs - /steven lands with promo_code=STEVEN-BARTLETT and utm_source=doac. YouTube runs yt_shanik against yt_baseline, a creator cut against a control.

And both route ads through Dub short links. The landing page lives in the redirect, so it changes without touching the ad object. No edit, no learning phase reset. If you steal one thing from this whole post, steal that.

5. Landing, funnel and paywall

No password in either product. No trial clock in either product. Both walls fire on usage, after the product has done real work in front of you.

The difference is the card. Viktor takes one early - a $1 hold that it voids, dressed as verification, taken at peak investment. Wispr never asks until you want more words.

That follows from the bill. Viktor's cost of goods is metered model work, so an unverified user burning $100 of credits is a real loss and annual prepay is a bet against its own costs. Wispr's marginal user is cheap, so it can give away 15 minutes a week forever and sell the year up front.

6. Channels

Viktor sells to teams, so LinkedIn is people. Its co-founder, its growth lead, outside B2B creators. Zero ads from the company page. A company page reads as an advertiser. A person reads as a peer.

Wispr sells to individuals, so YouTube is the second engine. "Siri vs. Wispr Flow." "Get 'Voice-Pilled' With Flow." Two creatives past 10 million impressions.

And the Meta direction since September runs opposite. Viktor went from 875 live to 828. Wispr went from 1,120 to 1,226 and has launched 4,936 ads in its lifetime, nearly three times Viktor's 1,671.

Watch both live on the Viktor ads library and the Wispr Flow ads library.

7. What to steal from each

From Viktor

• A creative record on every ad. ID, angle, hook verbatim, in-house or creator, launch date. It outlives the campaign.
• Clone the winner before you shoot the next one. One file under twelve ad objects is cheaper than twelve files.
• Two headlines, held still. Test the wrapper, not the claim.
• Pay creators in product. Credits at 1.5x cash costs you margin, not money.
• A question before the email. A micro-commitment and a firmographic, even if they bounce.

From Wispr Flow

• Pay creators on conversion. A hash in the URL turns a UGC budget into a performance budget.
• Let the creator write the headline. Hold the numbers still instead.
• Ship in waves. Three days made up 45% of two months of launches.
• Name your rebuilds. S4, EvReset. The account carries its own history.
• Make the free tier a habit. A weekly allowance puts the wall back in front of you every week.

From both

• Caption on screen before the first word.
• Hold the shot. One cut in six seconds is enough.
• Route every ad through a redirect you control.
• No password. No trial clock. A wall that fires on usage.

Which one is right

Depends what a video costs you.

If your best ads come from inside the building - a founder, a product only you can demo, a claim the market hasn't heard yet - build Viktor's factory. Make fewer files and clone them hard.

If anyone with a phone can show your product working, build Wispr's marketplace. Make it cheap to say yes, pay on results and stop cloning.

Viktor is at $50M ARR. Wispr is worth $2B. Both answers work.

The account that runs neither is the one shipping four ads a quarter.

Break down your own market

Everything above is public ad activity, read at scale. The Sprites ads library does it for any advertiser, and creative analytics does it for your own account.

The full teardowns: Viktor AI's ads strategy and Wispr Flow's ads strategy. More in the series: Clay, Lovable, Suno and Speak.