Khandoker Mursalat creates faceless, AI-powered UGC-style ads for brands, using scripts, generated product visuals and voiceovers to build creative variants for testing.
They are most useful for script-led, faceless AI UGC ad variants built for paid social testing.
Best for
Direct-to-consumer brands with one hero product, one benefit to prove, and no need for a person on camera.
Not for
Brands needing on-camera people, regulated claims, or several benefits proved in one ad.
Key takeaways
The script decides whether an ad converts; visuals are the replaceable layer.
One claim, one proof, one ask.
One script can become five ads that differ only in the opening hook.
Check first: Confirm client results, ad performance figures, rates, and partnership terms first.
Main ideas
The script decides whether an ad converts; visuals are the replaceable layer · One claim, one proof, one ask · Disclose the AI and the absence of a person on camera up front
Khandoker Mursalat works on faceless, AI-assisted UGC-style ads.
Read more
Their posts mostly argue what makes that kind of creative work, rather than promote the tools.
The through-line: the script decides whether an ad converts and everything visual is replaceable; an ad should prove one claim instead of gesturing at three; and the unusual parts of the setup — AI production, no one on camera — are worth disclosing at the start rather than hiding.
Two smaller threads apply the same instinct to selling: listings and outreach messages should match how the other side behaves, whether that means how a buyer searches or how a platform reacts to a burst of messages.
The script decides whether an ad converts; visuals are the replaceable layer
Mursalat argues that writing is the part of an ad that cannot be substituted; production choices come later.
Read more
The pipeline runs in a fixed order — script first, visuals last — because better footage, voiceover or editing will not rescue an ad that cannot sell on paper.
Next come generated product visuals as B-roll (the product in real-life contexts, macro shots, before/after setups), an AI voiceover written to sound like a real user, with natural pacing, short sentences and slight imperfections, and an edit built for the feed, with captions on every word and a cut or zoom every 1.5 to 3 seconds.
They say the first three seconds do most of the work and describe the script as the moat; faceless production makes scale possible but does not make the ad good.
One script becomes five ads with the same body and different openings. The variation tests each hook on its own, helping a brand learn which opening earns attention without changing the rest of the creative.
The pipeline post presents this as a system, not a one-off method, suggesting the order of steps is a settled working preference rather than an experiment.
Mursalat's recurring argument about why ads fail is that they try to say too much.
Read more
Their One-Claim Rule: pick a single claim for the whole creative, spend its full length proving it, then close with a single ask.
Multiple claims read as multiple half-proved claims, they argue, leaving viewers who are offered three unproven benefits believing none of them.
Proof means demonstration, not assertion: show the product doing the thing, get close enough to make it unambiguous, and put the number on screen.
Keep the call to action to one so the viewer is not asked to choose.
A water bottle ad that claims cold retention for 24 hours, cupholder fit and leak-proofing in one spot illustrates the failure: three claims competing for belief. The fix is to choose one — say, the 24-hour cold claim — and spend the entire video proving it, with the 24-hour figure on screen and a single shop-now close.
This pairs with the practice of cutting one script into five hook variants. In their framing, the single claim is what the ad argues, while the variants test different openings to that same argument rather than different messages.
Disclose the AI and the absence of a person on camera up front
Mursalat says to disclose the use of AI on day one, before work starts, rather than leave the brand to discover it.
Read more
Brands are buying creative that converts, not a production method, so the label does not change the result, they argue.
Faceless production is an advantage, not a defect: it allows more variants than a filming schedule can match and keeps one consistent character across ads.
A brand that wants face-to-camera content should say so immediately, while a wasted week is still avoidable.
The disclosure filters enquiries: brands that agree are, in their view, buying results in the first place.
The post opens with the disclosure that the UGC is AI-powered and Mursalat never appears on camera.
It contrasts this with AI creators who bury the detail and hope the brand does not ask.
Losing a lead at the start is presented as the point, not a cost.
Write a marketplace listing the way the buyer actually behaves
A separate thread applies the same reasoning to selling: match the other side's behaviour instead of describing yourself.
Read more
After auditing two Fiverr gigs that had received 13 impressions between them in 30 days, Mursalat set out rules for the listing.
Titles should use action plus outcome, in the buyer's language; the platform rejects the 'I will' prefix, and buyers do not search in the first person.
Treat the five tag slots as search positions and fill them with distinct phrases a buyer would type, not variations of one word.
Since only about two description lines appear before the 'see more' cut, lead with the outcome and put biographical material below the fold.
The description filter blocks a keyword repeated three times.
The 13 impressions figure is the concrete starting point: Mursalat is explicit that it counts appearances in search, not clicks.
The replacement title offered as a model is 'I create scroll-stopping UGC ads for DTC brands', contrasting with the self-describing alternative the rules are meant to rule out.
The low number is framed as a data point to optimise and retest, not a verdict on the service.
Send outreach at a pace that does not look like spam
Mursalat's outbound rule follows a specific failure: roughly 30 outreach messages sent in one morning on Facebook, after which later ones silently failed to send, though the account was never restricted.
Read more
They conclude that sending velocity affects whether messages arrive.
Their Warm Batch Rule caps a session at five new message threads, requires real gaps between sends rather than back-to-back messages, and rules out sending an entire batch at once.
The framing is that platforms act on patterns resembling spam, not on outreach itself, so the control is over rhythm rather than the number of prospects.
The failure mode carries the argument: nothing about the account was flagged, no warning was issued, and the messages that failed simply did not arrive. In Mursalat's account, that makes volume a deliverability question rather than a compliance one.
This post answers the question "what does this person actually do?" in one place, laying out the production model from script to variants and explaining each step.
Read more
Khandoker Mursalat outlines a five-step process for brands: write the script first and create visuals last; use AI-generated product footage in real-life contexts, macro shots and before/after setups; add an AI voiceover written to sound like a real user, with natural pacing and slight imperfections; edit for the feed with captions on every word, a cut or zoom every 1.5-3 seconds and most of the work done in the first three seconds; then turn one script into five variants by swapping openings. The system rests on two ideas: the script, not the visuals, decides whether an ad converts, and testing the hook separately from the rest of the ad makes variants useful.
For brands planning a campaign
Expect script-led creative, AI product visuals and voiceover rather than a filmed creator, with hook-level A/B variants. If a campaign requires face-to-camera footage or a single finished spot, this approach is not a fit.
This post explains how they position AI in client conversations, helping brands judge whether that approach fits before the first call.
Read more
Khandoker Mursalat says they tell every brand upfront that the UGC is AI-powered and that they never appear on camera, while other AI creators bury the detail. They give three reasons: brands buy creative that converts, not a production method; faceless production allows more variants and one consistent character across ads; and brands that want only face-to-camera content can find that out on day one instead of after a wasted week.
For brands planning a campaign
Khandoker Mursalat says any brief should accept faceless AI production. Brands that require on-camera creator authenticity should not proceed; the post presents that as an intentional filter.
This post shows how they approach a script before choosing production tools, which they describe as decisive.
Read more
It also offers a quick way to judge whether their creative reasoning fits a brand's messaging.
The post argues that most creatives fail because they make three claims instead of one.
A cup example illustrates the point: "keeps water cold 24 hours, fits any cupholder, never leaks" reads as three half-proved claims, so the viewer believes none.
The proposed rule is one claim, one proof, one ask: pick a single claim for the whole creative, spend the full video proving it with a demo, a close-up and the number on screen, then close with one call to action.
For brands planning a campaign
Briefs that pile several product benefits into one asset conflict with this approach. Brands should narrow the message to one provable claim per creative and plan variant testing around hooks rather than feature lists.
This post gives concrete evidence of how they manage their own marketplace presence, including numbers and a platform mechanic, rather than offering advice in the abstract.
Read more
It also shows their operating style.
Khandoker Mursalat reports that their two oldest Fiverr gigs received 13 impressions combined over 30 days and describes auditing and rewriting both.
Three rules come out of the exercise: write the title the way a buyer searches (action plus outcome, since Fiverr rejects the "I will" prefix and buyers do not search in first person); treat the five tags as search slots filled with phrases buyers actually type, such as "ugc ads" or "tiktok ad creative"; and lead the first two visible description lines with the outcome rather than a biography.
A platform quirk is noted as well - the description filter blocks a keyword repeated three times.
For brands planning a campaign
The creator is actively pursuing inbound gig discovery. Brands sourcing through freelance marketplaces should expect buyer-language titles and outcome-led descriptions; the same discipline shapes how they present their work.
This post shows them testing a channel, hitting a platform limit and changing their method.
Read more
It also shows they handle outbound work as well as creative production.
Khandoker Mursalat describes sending roughly 30 outreach DMs in one morning on Facebook.
Later messages silently failed, although the account itself was not restricted, leading them to conclude that outbound velocity affects deliverability.
Their warm batch rule is to send at most five new threads per session, leave real gaps between sends rather than message back-to-back, and never blast a full batch at once.
The reasoning: platforms punish spam-like patterns rather than outreach itself.
For brands planning a campaign
This approach is relevant to briefs that include outreach or lead generation support. It favors paced sending over bulk blasting: brands wanting high-volume cold outreach should treat the ceiling of five new threads per session as a constraint.
Mursalat's main offer is UGC-style ad creative made without anyone appearing on camera.
Read more
The pipeline follows a fixed order: write the script first (problem, failed alternatives, proof, call to action), generate product visuals as B-roll, add an AI voiceover with natural pacing and short sentences, then edit for the feed with captions on every word and a cut or zoom every 1.5–3 seconds.
One script becomes five ads that differ only in the opening hook.
They call the script the "moat" and treat generated visuals as the replaceable layer: if the script cannot sell on paper, no visual rescues it.
Script → generated product visuals → AI voiceover → feed edit → five hook variants from one script body. The first three seconds are treated as doing most of the work.
For brands planning a campaign
A brand can get several testable hooks per script without a shoot day, plus one consistent character across ads. The approach suits product-led creative where the product, not a person, is the subject; it is not suited to real-person testimonials or founder-led ads.
One claim per creative
The recurring argument is that weak ads fail by trying to say too much.
Read more
Mursalat's "One-Claim Rule": one claim → one proof → one ask.
An ad that claims cold retention, cupholder fit and leak-proofing at once, they argue, makes three half-proved claims and leaves viewers believing none of them.
The prescribed fix is to pick a single claim, spend the whole creative proving it with a demo, a close-up and the number on screen, then make one call to action.
Choose one claim, prove it throughout the ad, and close with a single ask.
The rule governs the ad's core claim, not its testing plan: the same one-claim script is still cut into five different openings, varying the hook rather than the body.
For brands planning a campaign
Useful for fixing underperforming creative that is technically well produced: the key review question is whether the ad proves one thing or gestures at three. A brand briefing this creator should bring the single claim it most wants proved.
Disclosing AI production up front
Mursalat says every brand is told on day one that the UGC is AI-powered and they never appear on camera.
Read more
Their reasoning: brands buy converting creative, not a production method; faceless production allows more variants than a human filming schedule; and brands that specifically want face-to-camera "authentic" content should say so before either side spends a week. They say brands that agree after disclosure tend to be buying results.
For brands planning a campaign
This approach avoids discovering the format mid-brief and makes any mismatch with a face-to-camera requirement clear early. Brands that require a real creator on camera should treat it as a mismatch, not a negotiation.
Freelance marketplace positioning
A separate thread of the material covers how Mursalat sells on Fiverr: auditing two gigs that had 13 impressions in 30 days and rewriting them.
Read more
The rules given are to title a gig the way a buyer searches (action plus outcome, not "I will…"), to use the five tag slots as distinct buyer search phrases rather than variations of one word, and to lead the first two description lines with the outcome since buyers see roughly two lines before "see more." Mursalat also notes that Fiverr's description filter blocks a keyword repeated three times.
For brands planning a campaign
Shows how the service is packaged and described to marketplace buyers, useful context for brands sourcing through a freelance platform rather than commissioning directly. The posts do not show the gigs; ask for package details, pricing and delivery terms before commissioning.
Outbound volume treated as a deliverability constraint
After Facebook rate-limited them, Mursalat says the cause was sending roughly 30 outreach DMs in one morning.
Read more
Later messages silently failed while the account itself stayed unrestricted.
Their resulting "Warm Batch Rule": a maximum of five new message threads per session, real gaps between sends, and never pushing a whole batch at once.
They frame the issue as platforms punishing spam-like patterns rather than outreach itself.
For brands planning a campaign
Relevant when a campaign involves coordinated outbound or community seeding: the approach is to throttle sending rather than scale it, setting expectations for how quickly a DM-led push would ramp. This is a working-style observation, not evidence of campaign results.
The five posts name no sponsor, client, affiliate or investment partner.
Read more
Khandoker Mursalat’s clearest commercial relationships are with the platforms they work through: Fiverr, where they maintain and sell gigs for UGC ad creative, and Facebook, which they use to send outbound lead-generation messages.
They describe client work only in general terms — brands buying faceless AI-assisted ad creative — without naming a brand.
Ask for named client examples before assessing any specific client or sponsorship relationship.
Fiverr
Mention
Mursalat sells UGC ad-creative services through Fiverr gigs and describes auditing and rewriting two of them after they received 13 impressions in 30 days.
Read more
They treat Fiverr as their storefront rather than as a client or sponsor, and set out their own rules for gig titles, tags and first description lines aimed at buyers on that marketplace.
Mursalat describes being rate-limited by Facebook after sending roughly 30 outreach direct messages in one morning, with later messages failing to send silently.
Read more
They use the experience to set a slower outbound pace — a maximum of five new threads per session, with gaps between sends — and present Facebook messaging as a lead-generation channel rather than as a client or sponsor.
Khandoker Mursalat's posts focus on producing UGC-style ad creative with AI and no one appearing on camera.
Read more
Start with the pipeline post — script first, generated product visuals, an AI voiceover, a feed edit, then five hook variants cut from one script — and read the One-Claim Rule post alongside it for the reasoning behind the method.
A third post on disclosure explains why they tell every brand about the AI production and the missing on-camera presenter before work begins.
Two posts step outside ad production: one on rewriting two dormant Fiverr gigs, the other on Facebook outreach limits.
Both apply the same habit of matching how the other side actually behaves rather than describing yourself.
Follow the tags below for more on faceless AI production, ad scripting and hook testing; the marketplace-selling and outbound-outreach tags point to the narrower threads.
AI UGC adsFaceless content productionUGC-style ad creativeAd scriptingVideo hooksAd creative testingShort-form video adsAI-generated videoDTC brandsFreelance marketplace sellingOutbound outreach
Grace Eva
Writes about making UGC-style ad video with AI instead of a creator and camera — avatars, natural-sounding scripts, AI voiceovers, product demos and B-roll — and how much faster that lets a brand test hooks and product angles in paid social.
Read more
Eva's posts are short guide promotions that list what the format can produce and name a specific AI tool, leaving strategy to the brand. Mursalat lays out craft rules — script before visuals, one claim per creative — and treats the AI, faceless setup as something to disclose to a client on day one, not a selling point to lead with.
Works the same ground of hook-led creative testing for DTC paid social: diagnosing an ad by where viewers drop off, then re-cutting one clip into five ads that each open differently to test the hook.
Read more
Soriano starts with footage a brand has already paid for, re-edits it and frames the work as an ad-by-ad check against hook rate, hold rate, click-through and conversion. Mursalat starts with the script and generated visuals, without footage, a shoot day or an existing clip.
Makes the same case that ad editing is judged by whether it produces action, not how polished it looks: a strong hook, the problem introduced quickly, proof made obvious.
Read more
A larger volume of video only helps if the story lands.
Ganzina writes for brands briefing or hiring editors, framing the gap as the difference between video editing and direct-response editing skill. Mursalat's premise is that no camera or creator is needed; the deciding work is in the script and opening-hook variants.
Khandoker Mursalat’s posts cover scripting, product visuals, ad variants, and creative testing, making them a potential fit for DTC brands evaluating faceless, AI-powered UGC ads and performance creative.
Read more
If a brief depends on audience demographics or professional roles, request recent audience analytics; for employer, occupation, or seniority, ask for audience survey results or another documented source.
Khandoker Mursalat sells a done-for-you AI-powered faceless UGC ad service: script-first creative, generated product visuals and voiceover, feed-ready editing, and multiple hook variants.
Read more
They state the work is AI-powered and that they never appear on camera, so brands needing face-to-camera creator content should rule out the fit before scoping.
The posts ask interested brands to DM “ADS” on LinkedIn.
Before scoping, ask for sponsorship terms, a rate card and a direct partnership inquiry route; the posts publish no official sponsorship policy, rate card or standalone partnership page.
AI-powered faceless UGC ad creative
Buyer requirement
The brand must accept upfront that the creative is AI-powered and Khandoker Mursalat does not appear on camera. If face-to-camera authenticity is required, raise it on day one so both sides avoid a wasted week.
Script-first creative system
Buyer requirement
Provide the product, chosen claim, proof points and desired call to action.
The script comes first: problem, failed alternatives, proof, CTA.
Expect it to carry the ad, not the visuals.
Faceless AI UGC pipeline
Buyer requirement
Supply product context and enough real-life usage detail for generated B-roll and macro or before-and-after visuals. Approve the AI voiceover tone—natural pacing, short sentences, slight imperfections—instead of expecting a studio shoot.
Hook-variant testing from one script
Buyer requirement
Test openings, not whole ads.
One approved script body becomes five variants with different hooks.
Plan for feed editing with captions, cuts or zooms every 1.5–3 seconds, and a first three seconds that do most of the work.
One-claim creative discipline
Buyer requirement
Choose one claim for the whole creative and supply one proof point. Expect a single ask such as “shop now”; multiple claims in one ad are treated as a reason the creative fails.
Brief: a direct-to-consumer brand with one hero product and a single benefit it wants proved, buying creative for paid social at volume, open to AI-generated visuals and voiceover, and with no requirement for a person on camera.
Read more
This fits the process described in Mursalat's posts.
The production model matches the brief: script first, then generated product visuals in real-life contexts, macro shots and before/after setups; an AI voiceover written to sound like a real user, with natural pacing and short sentences; and a feed-ready edit with captions on every word and a cut or zoom every 1.5–3 seconds.
The testing unit is the hook: one approved script becomes five ads that differ only in the opening, giving a paid-social buyer a clear basis for spend.
The One-Claim Rule gives the brief a natural constraint: name the single claim the ad must prove, supply the proof and let the creative serve that claim instead of stacking benefits.
The AI disclosure happens on day one, so a brand comfortable with faceless production won't encounter the format halfway through the project.
They position face-to-camera authenticity requests as something to settle before either side spends a week.
The Fiverr rewrite episode shows their working style: audit a dead asset against how buyers actually search, rewrite from zero and treat a small number as a data point rather than a verdict.
When to choose a specialist
Campaign need: a brand that needs a real person on camera — a founder-led spot, a customer testimonial or an on-camera creator endorsement for a product where claims carry regulatory weight.
Read more
Mursalat states plainly that they never appear on camera and that their UGC is AI-powered.
They frame a face-to-camera authenticity requirement as a mismatch to surface on day one, so a campaign built around a named person on screen should go to a different creator.
Before committing budget, ask for client results or case studies, ad performance figures, rates and partnership terms, and audience composition data.
A campaign that wants several product benefits proved in one asset also conflicts with the One-Claim Rule.
For high-volume cold DM outreach, plan around paced sending — the stated ceiling is five new threads per session — because platforms punish spam-like patterns rather than outreach itself, not a bulk blast.
A worked campaign example
Brief: a DTC insulated drinkware brand has one product and one claim to prove — that it keeps water cold for 24 hours — and wants paid-social creative volume plus a hook test.
Angle: build the campaign around the single claim, not the product's full feature list. An ad that claims cold retention, cupholder fit and leak-proofing at once reads as three half-proved claims.
Step one, the script: open on the problem (warm water by mid-afternoon), show the failed alternative (a thin bottle that sweats), prove the claim with a close product demo and the 24-hour figure on screen, then make one ask — shop now.
Step two, visuals: generated B-roll of the bottle in real contexts (gym bag, desk, car cupholder) plus a macro pour shot, with no shoot day.
Step three, voiceover: short sentences, natural pacing and slight imperfections so it reads as a user rather than an announcer.
Step four, edit for the feed: captions on every word, a cut or zoom every 1.5–3 seconds, with the first three seconds doing the heavy lifting.
Step five, variant test: keep the body identical and produce five different openings — an ice-still-there reveal, a temperature-timer countdown, a 'this is why your bottle sweats' cold open, a failed-alternative open and a straight claim-and-proof open — then test hooks, not whole ads.
Before briefing, decide on the single claim, the proof asset (does the brand have demonstrable 24-hour data or footage to show on screen?) and whether AI-generated visuals and voiceover are acceptable to the approval chain.
Then set the test budget.
Ask for benchmarks or client results to guide expectations; without them, read results as data points and iterate rather than treating them as guaranteed outcomes.
If the brand's process requires a real person holding the bottle on camera, this brief belongs with a different creator.
A DTC brand already using Mursalat's ad-making method but needing help choosing and stacking AI tools — avatar and spokesperson generators, script and voiceover tools, product-to-video platforms, and creative-variation platforms — rather than another production framework.
Read more
Eva's posts target that step: how to make a first AI UGC ad without a creator or camera, and a guide comparing AI tool categories for UGC-style ad creative and what to check before choosing one. Choose this approach when the question is which tools to use and in what order, and the brand wants a walkthrough of the stated idea-to-script-to-video-to-testing-to-optimization workflow.
The material focuses on tool guidance and beginner enablement.
One post names and promotes Tagshop AI as a way to turn product ideas into UGC-style video, so the recommendation comes with a vendor attached.
Eva's writing covers what AI can produce — avatars, natural scripts, voiceovers, demo B-roll, captions and multiple variants — and why speed matters in paid social.
It does not offer the concrete ad-craft rules Mursalat gives: the one-claim rule, the 1.5–3 second cut-or-zoom cadence, first-three-seconds framing, or a scripted problem/failed-alternative/proof/CTA structure.
Brands that already know the tools but need production discipline will get less from this material; brands still deciding what software to buy will get more.
A DTC brand that owns footage and is deciding whether to re-cut it or shoot more.
Read more
Soriano addresses the same buyer decision as Mursalat — how to get more testable ads without another shoot — but starts with diagnosis: check hook rate, hold rate, CTR and conversion, find where viewers leave, then fix what broke.
His re-edit list turns one raw UGC clip into five ads by swapping the opener, leading with the problem, product or proof, and tightening the pace, all without reshooting.
Choose Soriano when ads have stopped working and the brand needs to know what to change, or when it has a library of clips, no budget for new production and wants a free hook rewrite as an entry point.
The starting point is human-shot footage.
Soriano assumes a clip already exists and works through the edit; Mursalat assumes there is no shoot day and instead builds generated product visuals, an AI voiceover and faceless variants.
Brands without footage cannot use Soriano's method; brands seeking AI-generated creative at scale are better served by Mursalat's.
Soriano's key contribution is a funnel-metric diagnostic that separates a hook problem from an offer, CTA or landing-page problem — a distinction Mursalat's posts do not make.
The tradeoff is production format: choose Soriano to re-cut owned footage, Mursalat to produce ads without a camera.
A DTC brand with production capacity — an editor and a growing volume of UGC and AI footage — whose polished output is not converting.
Read more
Ganzina argues that volume is not the constraint: as AI and UGC make video cheaper, the differentiator is direct-response editing, where every cut serves a hook, a fast problem statement, clear product value, trust-building, objection handling and a reason to act. Choose this approach when the brand has the assets and the editor, but its ads look good and underperform, and it needs someone who treats editing as a persuasion problem rather than a motion-design one.
The focus is craft and creative judgement, not a repeatable generation pipeline.
Ganzina explains what a performance edit should do — including cutting a beautiful opening shot or putting an ugly customer quote on screen because proof beats polish.
But the posts do not lay out Mursalat's scripted production sequence (script first, generated B-roll, AI voiceover, captions, five hook variants from one body) or a hook-testing protocol.
Brands starting from scratch will get less here than brands needing a firm decision framework for edits already underway.
Choose Ganzina to improve an existing edit and establish editorial standards; choose Mursalat to generate and test variants without a shoot.
Five LinkedIn posts by Khandoker Mursalat, along with the authored profile sections based on them, describe a creative process and advice, but provide no client results or campaign performance figures.
No post dates are published, so the timing or development of these ideas cannot be established.
