AI video models can now produce footage that would have needed a crew a few years ago. What they cannot do on their own is remember your brand. Every model starts from a blank canvas, so unless you give it your identity as real inputs, it invents one. The workflow below is how we build adverts on Reigniter, and it works as a checklist whichever tool you use.

Why most AI video ads drift off brand

There are four common reasons an AI advert stops looking like the company that paid for it.

• The logo is described, not supplied. A prompt that says "our green leaf logo" gets a green leaf. It will not be yours, and it will change from shot to shot.
• Colours live in prose. "Warm, premium, a bit of teal" means something different to every model and every seed.
• Each scene is prompted separately. Five scenes means five chances to drift, and nobody compares them side by side until the video is finished.
• Text is drawn by the model. Video and image models still misspell words and prices. A wrong price in an advert is worse than no price.

The fix for all four is the same: decide the brand once, store it as files and values, and make every scene read from that one source.

Step 1: Lock the brand kit before you write a prompt

A brand kit is the smallest set of decisions that makes something look like you: the logo, the brand name, a primary colour, a short palette and your fonts. If you have not written one down yet, our guide to what a brand kit is and how AI uses it covers it in detail.

For AI work, three rules matter most:

1. One logo. Pick the version that works small and on both light and dark backgrounds.
2. A fixed palette. Three to five colours with their hex values. Fewer choices left open means more consistent output.
3. Real font names. So on-screen type can be set in them rather than imitated.

On Reigniter you can import colours, fonts and the name from your website or a brand guidelines PDF, then review and save. If you are starting from nothing, the Brand Generator proposes logo directions from a name and one sentence about the business. Either way, the kit is saved once and every advert, still, motion graphic and training module reads it.

Step 2: Add the real product

If the advert shows a product, give the model a real photo of it. A clean shot on a plain background works best. Without one, each scene has to guess the packaging, the label and the proportions, and the guesses will not match.

This matters most for:

• product adverts and hero shots,
• UGC-style adverts where a creator holds the product,
• anything with a label, a screen or a printed name.

Step 3: Pick the format that fits the job

Different formats solve different problems. Choosing one before you write the script saves rewriting it later.

For more detail on the first format, read UGC-style AI adverts: what works. To see what each format looks like on Reigniter, the AI video page has examples.

Step 4: Plan the advert as scenes

Write the advert as four to six scenes before anything is generated. Each scene needs two things: what we see, and what we hear. A dependable shape for a short advert is:

1. Hook. One line or image that earns the next three seconds.
2. Problem or desire. Why the viewer should care.
3. Proof. The product doing the thing. This is where the real product photo pays off.
4. Call to action. One instruction, with the logo and the offer.

On the Reigniter canvas the plan appears as a list of scenes you can rewrite, reorder, add to or delete. Editing the plan costs nothing, so this is the cheapest place to get the advert right. A storyboard score shows how ready the plan is and lists what would raise it, such as adding your brand or your product.

Step 5: Storyboard before you render video

This is the step most teams skip, and it is the one that saves the most money.

Before any video is made, generate one still frame per scene. Put the frames side by side and check them against the brand kit:

• Is the logo the real one, in the right place?
• Do the colours match the palette?
• Is the product the right shape, with the right label?
• Does each frame look like it belongs to the same advert?

If a frame is wrong, change that one scene and draw it again. On Reigniter you can also type a single change on a frame, for example "make the background darker", and keep the rest of the frame as it is. Frames cost less than video, and fixing a scene at this stage means you never pay to animate a mistake.

Step 6: Keep text as type, not pixels

AI models draw letters as shapes. Sometimes the shapes are right. For anything that must read exactly, such as a price, a discount code, a product name or a legal line, set it as type instead.

Two ways to do that:

• Motion graphics scenes animate your headline, figures and interface panels from real type, in your fonts and colours, with no AI video model involved. See motion graphics.
• Static web adverts let the AI paint only the background while your headline is set on top in real type.

A good rule: if a word would cost you money when misspelled, it should be type.

Step 7: Render, check and make versions

Once the frames are right, build the video. Choose the aspect ratio for where it will run, the voice, and whether music plays under the voiceover. On Reigniter the render window shows the credit cost and your balance before and after, and nothing starts until you confirm. If a scene fails, its share of the cost is refunded automatically.

Then make versions rather than starting again:

• Hooks. Test two or three openings against the same body.
• Placements. A vertical cut for Reels, TikTok and Shorts; a wide cut for YouTube and web.
• Languages. A language version is rendered again from the same storyboard, so a presenter's mouth matches the new words.

Which video model you pick for each shot also matters. We compare the main ones in Kling vs Veo vs Seedance for product adverts.

A checklist before you publish

Run through this list before an advert goes live:

• The logo is your real file, not a redrawn version, in every shot where it appears.
• Every price, name and offer is set as type and spelled correctly.
• Colours stay inside the palette from the first scene to the last.
• The product looks the same in every scene.
• The first three seconds work with the sound off.
• The call to action names one thing to do.
• You have the rights to every face, voice and piece of music in it.

Where Reigniter fits

Plenty of tools now apply a brand kit, including Canva and VEED, and that is a good thing. What Reigniter adds is the whole loop in one place. It can build the brand identity itself, with a brand kit and brand guidelines, and then make every video advert, product still, motion graphic and AI-presenter training module from that same brand. Change the kit, and the next thing you make follows it.

If you want to try the workflow, the free plan covers stills, brand kits and brand guidelines, and video starts on Pro. The pricing page has the details, and every action shows its credit cost before it runs.

Every brand has a look. The trouble is that most brands keep it in people's heads, scattered PDFs and old slide decks. That works until you ask software to make something for you. A brand kit is how you hand your look to a tool without losing it on the way.

What a brand kit is

A brand kit is the smallest set of decisions that makes something recognisably yours, saved as files and values rather than as descriptions. It usually holds:

Some teams add a product photo library, a tone of voice line, or approved imagery. Those are useful, but the five items above are the core. If they are fixed, most of the look follows.

Brand kit vs brand guidelines vs brand identity

These three terms get used as if they mean the same thing. They do not.

• Brand identity is the whole system: how you look, sound and behave.
• Brand guidelines are the written rules for that system: clear space, minimum sizes, how to use colour, tone of voice, examples of right and wrong.
• A brand kit is the set of assets the rules refer to: the logo files, the colour values, the fonts.

A designer reads the guidelines. A tool reads the kit. You need both, and they should agree. If you have neither yet, our guide on how to write brand guidelines gives you a structure to start from.

How AI tools use a brand kit

Generative models do not know your brand. Each image or video starts from noise and a prompt. Anything you only describe, the model invents. A brand kit changes the inputs from descriptions to facts. In practice, a well-built tool uses each part differently.

The logo is placed, not drawn

The most important rule. A model asked to draw "our logo" will produce a plausible new logo every time. A system that holds your real file can lay it over the finished frame, in a set position and size, so the mark on screen is your file rather than a lookalike. On Reigniter the brand appears as its own node on the canvas, where you choose where the logo sits (top left, bottom right and so on) and how large it is.

Colours become constraints

Hex values give the model and the layout engine a fixed range. Backgrounds, accents, lower thirds and motion graphics can be set inside the palette rather than drifting towards whatever the model likes. This is where a short palette beats a long one: five colours leave less room for surprises than fifteen.

Fonts turn text into type

Image and video models draw letters as shapes, and shapes can be wrong. When the kit names your fonts, headlines, prices and captions can be set as real type on top of generated imagery. On Reigniter, motion graphics scenes and static web adverts both work this way, so the words are always spelled as you wrote them. See motion graphics and image ads for examples.

The kit is read every time

The difference between a template and a brand system is when the brand is applied. A template bakes your logo into one layout. A brand system reads the kit each time something is made, so if you update the kit, the next advert, still or training module follows the new version. Reigniter works the second way: the brand is applied as each scene is made, not painted into the plan and left there.

What a good brand kit looks like for AI

A kit that works for a designer can still trip up an AI tool. These habits make the output far more consistent.

1. One logo, not five. Pick the version that reads at small sizes. Keep variants for print.
2. Transparent background. A PNG or WebP with no box around it can sit on any frame.
3. Contrast in the palette. At least one dark and one light colour, so text always has something to sit on.
4. Real font names. "Clean sans-serif" is a description. "Inter" or "Archivo" is a font.
5. A real product photo. If you sell a physical product, a clean shot on a plain background stops every scene inventing its own packaging.
6. One kit per brand. Agencies and multi-product companies should keep a separate kit for each client or product line so nothing crosses over.

How to build a brand kit

There are three quick routes, depending on where you start.

You already have a website. Import colours, fonts and the brand name from your site, then review what was found. Imports are a starting point, not the final word, so check every value.

You have brand guidelines as a PDF. Import the colours and fonts from the document, upload the logo, and save.

You have nothing yet. Generate one. On Reigniter the Brand Generator asks for the brand name, what the business does and how it should feel, then proposes four logo directions in words. Proposing costs nothing. When you pick one, it draws the mark, writes the brand kit and opens a guidelines document you can save as a PDF.

Whichever route you take, finish by checking the kit against a real piece of work. Make one still, look at it next to your website, and adjust before you make anything bigger.

Common brand kit mistakes

• Storing the logo as a screenshot. Low resolution and a white box around it. Export a clean file instead.
• Too many "primary" colours. If everything is primary, nothing is. Choose one.
• Fonts you do not have a licence for. Check the licence covers video and advertising use.
• Letting imports go unchecked. A website may use a colour for one button that is not part of the brand at all.
• Different kits in different tools. If your design tool, video tool and presentation tool each hold a slightly different kit, you will get three slightly different brands.

One kit, every format

The last mistake above is the one that costs most over time. Many tools now apply a brand kit, including Canva, VEED, Google Pomelli and Typeface, and that is progress. But if your adverts come from one tool, your product stills from another and your training videos from a third, you maintain three kits and they drift apart.

Reigniter's approach is one brand system for the whole loop. It can build the identity (the kit and the guidelines), then make video adverts, product stills, motion graphics and AI-presenter training modules from that same kit. To see how the kit carries into an advert, read how to make on-brand video ads with AI. Brand kits, guidelines and stills are on the free plan; the pricing page shows what each plan adds.

Almost every organisation already has its training. It lives in PowerPoint decks, PDFs and Word documents that someone clicks through in a meeting room, or emails round and hopes people read. Turning those files into short videos that ask questions back is one of the most useful things AI can do for a training team. It is also easy to do badly.

Here is a workflow that produces a course people finish, not a deck with a voiceover.

Why a narrated slide deck is not a training video

The quickest conversion is to record a voice over every slide. It is also the least effective. Slides were designed to support a person in a room. On their own they are dense, text-heavy and passive. Nobody checks whether the learner understood anything.

A good training video changes three things:

• One idea per scene. A slide with six bullet points becomes several short scenes.
• A person explaining. A presenter talking to camera holds attention better than text.
• Questions that matter. Checkpoints prove understanding, and wrong answers loop back to the point that was missed.

Step 1: Write one objective

Before you import anything, write the objective as a sentence that starts "By the end, the learner can…". For example: By the end, the learner can log a safety incident in the reporting system.

One objective per module. If your deck covers four things, that is four short modules, not one long one. Short modules are easier to finish, easier to update, and easier to track.

On Reigniter the objective is the one required field when you start from a file. It shapes how the scenes are planned and where the questions go.

Step 2: Prepare the deck

Ten minutes of tidying saves an hour of editing later.

• Cut the agenda, title and "any questions?" slides. They add length without teaching.
• Split crowded slides. One point per slide gives one point per scene.
• Put the real words on the slide. The import reads the text on each slide, so speaker notes that never made it onto a slide will not be used.
• Collect your screenshots. If the training shows a system, grab clean screenshots of every screen you teach.

Step 3: Import the file

On Reigniter open Interactive Training, then From PDF or deck. You can start from:

Up to 12 scenes are planned from one file. Add the audience if you know it (for example, new warehouse staff, first week), pick a length of one to two, five or ten minutes, and check that your brand kit is selected. The document-to-video flow is described in more detail on the document to video page.

Step 4: Shape the scenes

The canvas opens with a planned module. Steps are numbered, a checkpoint question appears after every two or three steps, and wrong answers are wired back to the step that teaches them.

Now edit like a trainer, not a typist:

1. Rewrite for speech. Slide bullets read badly out loud. Turn them into short sentences a presenter would actually say.
2. Open with why. The first scene should say why this matters to the learner.
3. Teach one thing per scene. If a scene needs "and also", split it.
4. End on a recap, not a question. Tell learners what to do next time.

Step 5: Add the presenter and your screenshots

An AI presenter delivers the scenes with lip-sync to your script. Pick one presenter and keep them for the whole course, so modules feel like a series. If you want your own face, you can create a private digital twin from a photo of yourself, with a consent check. Presenters are covered on the AI presenter page.

Then attach your real screenshots to the scenes that show a system. Reigniter never invents a screen: if a scene teaches a step in software, it needs the real screenshot, and the build button tells you how many are still missing.

Step 6: Write checkpoints that teach

Checkpoints are where a video becomes training. A checkpoint scene pauses on its last frame and shows answer buttons, up to six per question. Each answer jumps to a scene, or opens a web page.

Good checkpoint questions:

• Test the decision, not the vocabulary. "A customer asks for a refund after 40 days. What do you do?" beats "What is our refund window?"
• Make wrong answers plausible. If nobody would choose them, the question teaches nothing.
• Send wrong answers back. Point each wrong answer at the scene that teaches the fact, so the learner sees it again and gets another go.

The score is the percentage of questions answered correctly, and if a learner answers again, their latest answer counts.

Step 7: Storyboard, then render

Run the storyboard to draw one still frame per scene and check them before any video is made. Frames are charged per frame and the cost is shown before anything is spent. Fix a scene's words and draw it again if it is not right. When every frame looks right, build the video.

Step 8: Deliver it

You have two ways to get the course to people:

• A share link. The interactive player runs in any browser and learners need no account. On the Interactive Training and Agency plans you can also ask for names, invite named people, set one attempt, a timer and a pass mark, and see results per person. We cover that in training without an LMS.
• SCORM 1.2 for your LMS. Download a package and upload it to the LMS you already use, with optional xAPI statements to your own LRS. The package streams the video from Reigniter, so a re-render updates it without re-uploading. See SCORM vs xAPI for small teams.

If your learners speak different languages, render language versions of the module. When every scene has a version in a language, learners can switch language mid-video from the player, across nine voice languages.

What this costs

Rendering video needs a paid plan, and every action shows its credit cost before it runs. The Interactive Training plan is $149 a month with 238 credits and includes unlimited SCORM 1.2 export, xAPI and the tracking features, with no sales call. See pricing for the current plans, or read about the whole training product on the employee training page.

Every few weeks a new AI video model tops a leaderboard, and every few weeks someone asks which one to use for their adverts. The honest answer is that the question is slightly wrong. An advert is a sequence of different shots, and different models are good at different shots. The useful question is: which model for which shot?

This guide covers the models we use most for product adverts on Reigniter, what each is good at, and how we choose. It is based on how we route shots in our own product, not on a benchmark.

The short version

The rest of this article explains the reasoning, because your product may behave differently from ours.

What matters in a product advert shot

Before comparing models, it helps to be clear about what "good" means for a product advert. It is not the same as what makes a viral demo clip.

• Identity. The product must look like the product: same shape, same label, same colour, in every frame. A beautiful shot of a slightly different bottle is a failed shot.
• Brand. If a logo or wordmark is in shot, it has to survive the camera move without being redrawn.
• Motion that fits the format. A luxury hero shot wants a slow, controlled move. A UGC insert wants a little handheld wobble. The same model is rarely best at both.
• Cost per usable shot. A cheaper model that needs three attempts can cost more than a pricier one that works first time.

Kling: animating a strong still

On Reigniter you can choose Kling 3 Pro, Kling 2.6 and Kling 2.5 Turbo.

What it is good at. Taking one product photo and giving it smooth, cinematic motion. Kling animates the image you give it, which is exactly what you want when you have a good product shot and need it to move. Kling 3 Pro is our strongest single-image option, and on the Premium quality setting shots with people route to it (on Standard they route to MiniMax H3). Kling 2.5 Turbo has narrower jobs: UGC product inserts, the presenter walk-in, and shots that move from a first frame to a last frame.

Where to be careful. Because Kling animates one image rather than composing a new shot from several references, it cannot combine a product photo and a separate logo into one frame the way reference-based models can. Get the frame right first, then animate it.

Use it for: product close-ups, lifestyle moments built from a strong still, UGC-style inserts, and cutaways that need to match a Kling-rendered presenter walk-in.

Veo 3.1: when the shot needs sound

Veo 3.1 generates audio natively along with the video. That changes what a single generation can deliver: the fizz of a can opening, footsteps, room tone, a line said on camera.

What it is good at. Moments where sound and picture belong together and would be awkward to add afterwards.

Where to be careful. In a narrated advert you usually want control over the mix: a voiceover, a music bed and sound effects you choose. Reigniter adds those with ElevenLabs v3 voice (70+ languages), music and sound effects, so native audio is a tool for specific shots rather than a default.

Use it for: atmosphere, product sounds, short on-camera moments.

Seedance 2.0: choosable, but test it on your product

Seedance 2.0 is available on Reigniter and some teams like its look. We keep it in the picker because choice matters and models improve fast.

What we can report honestly is one comparison of our own. Given the same logo and the same watch shot, MiniMax H3 held the brand and produced the full parts-explosion move we asked for, where Seedance only hinted at it, and Seedance cost noticeably more for that shot. That is one test on one product, not a verdict on the model. If Seedance's style suits your brand, draw a storyboard frame and a short clip with both and compare.

MiniMax H3: the one we default to for product shots

The model that most often answers "which model for a product advert" in our routing is not in this article's title. For hero product moves (orbits, reveals, product films and stylised worlds built around a real product) we default to MiniMax H3.

Why. A product film shows one object from several angles. The model that keeps it the same object from the first shot to the last wins, and in our testing that has been H3. It also composes the product and the brand mark in one shot, so the logo is still legible after the camera moves. MiniMax Hailuo 2.3 is also available.

Presenters are a different problem

Talking-head and presenter shots are not a choice between general video models. The mouth has to follow the voice track, so these shots use lip-sync models driven by the audio. On Reigniter, presenter scenes run on Wan 2.5 and 2.6 lip-sync, and walk-and-talk scenes combine a Kling walk-in with an OmniHuman delivery. A general model picked for a talking scene would not move the mouth to your script, so the picker does not pretend otherwise. For more on these formats, see UGC-style AI adverts.

Other models worth knowing

• Luma Ray 2 and Ray 2 Flash. Flash is quick and lower-cost, useful for drafting motion. Like every video model, it needs a paid plan; the free plan is stills only.
• Runway Gen-3 Turbo and Pixverse v4.5. Further options in the picker when you want a different motion character.
• Nano Banana 2 and Flux Pro Ultra. Image models for storyboard frames and stills; Flux Schnell for quick drafts.
• Topaz for upscaling finished video, and Whisper for captions.

The full list lives on the models page.

How to choose, in practice

1. Storyboard first. Draw a still frame for each scene and fix the frame before choosing motion. Our on-brand video ad workflow covers why.
2. Pick per scene. A hero orbit, a UGC insert and a presenter line can each use a different model.
3. Compare on your product. Render the one or two scenes that matter most with two models before committing the whole advert.
4. Watch the cost of a usable shot, not the price per second. Reigniter shows each scene's credit cost before it runs, and failed scenes are refunded.

Model choice is one part of making a good advert; the brand, the product photo and the plan matter more. To see the models in context, visit the AI video page or check what each plan includes on pricing.

Most training tools assume you have a learning management system. Synthesia, HeyGen and Colossyan make the video; an LMS is expected to hand it out and keep the records. For a large company with an L&D department, that is fine. For a 40-person business, a franchise, or an agency training a client's staff, buying and running an LMS just to find out who watched a video is a lot of overhead.

There is a simpler option: let the player keep the records.

What an LMS actually does for you

Strip away the course catalogues and the admin screens, and most small teams use an LMS for five things:

1. Assigning training to specific people.
2. Controlling access so only those people can open it.
3. Recording results: who completed it, their score, and whether they passed.
4. Evidence, usually a certificate or an exportable record.
5. Chasing, seeing who has not started yet.

If the training player can do those five things, you may not need the LMS at all.

How tracked share links work

A share link is a web address that opens the training in any browser, with no app and no account. On its own, a share link is anonymous: you know how many people opened it and which answers were clicked, but not who clicked them.

The change is making the link know who is watching. On Reigniter's Interactive Training and Agency plans, each module has three access settings:

The third setting is the closest to an LMS. Each invited person gets a personal link that acts as their sign-in. It lists the modules assigned to them, how they did, and their certificates.

Setting it up, step by step

1. Add your people

Add learners by hand, or import them from Excel, a CSV or a pasted list, with the columns worked out for you. If you train people for several organisations, file each person under a client.

2. Choose who can open each module

Open the module's access settings and pick one of the three options above. For staff training with a known list, choose Only people I add and select the learners.

3. Set the rules

Three controls turn a video into an assessment:

• One attempt. Once a learner finishes, the module closes for them and their result stands.
• A timer. A countdown from the limit you set. If time runs out, the attempt is marked as failed.
• A pass mark. The score, as a percentage of questions answered correctly, that counts as a pass.

4. Invite

Send invitations by email, or copy each person's personal link and send it however you like. Anyone without an email address can be given their link directly.

5. Watch the results come in

The training dashboard shows who is taking a module right now, and for each person their answers, score, time taken and whether they passed. People who were invited but have not started show that too, which makes chasing easy.

6. Export the records

Download results as CSV for your own records, a spreadsheet, or a client report. Each learner can print their certificate, or save it as a PDF, from their own page.

Training clients' staff: the agency setup

Agencies and consultants have an extra problem: several clients, each with their own people and their own modules, and none of them should see each other's.

Filing learners and modules under a client solves that. Each client can be given one page that lists every module anyone at that client may open. Invited-only modules stay on each person's own page. The Agency plan includes up to ten client workspaces, and because Reigniter also makes adverts and brand assets, the same account can deliver a client's marketing video and their staff training in their brand. The campaign series page shows the marketing side of that.

When you should still use an LMS

A tracked share link is not a replacement for every LMS. Keep or buy one if:

• you need learning paths across dozens of courses from many vendors,
• your organisation requires single sign-on into training,
• HR systems need to read training records automatically,
• a regulator specifies a particular system or format.

If you do have an LMS, export the module as a SCORM 1.2 package instead, optionally with xAPI to your own LRS. We compare the two standards in SCORM vs xAPI for small teams.

Making the training worth tracking

Records only matter if the training asks something. A few rules that make results meaningful:

• One objective per module, written as "by the end, the learner can…".
• A checkpoint every two or three steps, testing a decision rather than a definition.
• Wrong answers loop back to the scene that teaches the point, so a learner who gets it wrong sees the step again and tries once more.
• Keep it short. A few focused minutes beats a long module people abandon.

If your training starts life as slides, PowerPoint to training video walks through turning a deck into a module with checkpoints.

What it costs

Tracking (clients, named learners, invites, one attempt, the timer, live results, records and certificates) is on the Interactive Training plan at $149 a month with 238 credits, and on Agency at $199 a month with 318 credits. Interactive Training also includes unlimited SCORM 1.2 export, with no sales call. Credits pay for making the video; every action shows its credit cost before it runs, and failed jobs are refunded. See pricing, or the employee training page for the whole training product.

UGC-style adverts, the ones that look like a person filming themselves talking about a product, became a staple of social advertising because they feel native to the feed. AI has made them far cheaper to produce and test. It has also made it easy to produce a lot of bad ones: uncanny presenters, invented products and claims nobody can stand behind.

Here is what we have learned about making UGC-style AI adverts that perform, stay honest and still look like your brand.

What makes UGC-style ads work

The format works because it breaks the rules of traditional advertising. It looks cheap on purpose. It talks like a friend, not a brand. It gets to the point in seconds.

The adverts that work usually share five traits:

1. A hook in the first line. A question, a surprising claim you can back up, or a relatable problem. "I stopped buying three different cleaners" beats "Introducing CleanCo".
2. One benefit. Not a feature list. One thing the viewer will remember.
3. The product in hand. Being shown, used and held, not floating in a void.
4. A casual setting. A kitchen, a car, a desk. Real-looking light.
5. One call to action. "Link below" or "Try it free". One instruction.

What to avoid

These are the mistakes we see most often in AI UGC adverts.

Over-polish

The biggest giveaway. A perfectly lit studio shot, a sweeping camera move and cinematic grading look like an advert, and the viewer scrolls past. Match the motion to the format: a little handheld movement, a phone-camera framing, natural light.

An invented product

If you do not give the model a real photo of the product, it will draw one. It will look plausible and be wrong, and it may change between scenes. Always supply the real product photo.

Fake testimonials

An AI presenter is not a customer. Scripting one to say "I have used this for six months and it changed my life" presents an invented experience as real, which misleads viewers and can break advertising rules and platform policies. Write scripts as a presenter explaining or demonstrating the product instead: "Here is what this does", "Three things to know before you buy". It is just as persuasive and it is honest.

Faces and voices you do not have rights to

Never recreate a real person, a celebrity or a competitor's creator. If you want a real face, use your own or that of someone who has given explicit permission. On Reigniter, creating a digital twin from a photo requires you to confirm it is you, or that you have the person's permission.

Claims you cannot support

AI makes it easy to write bold claims. It does not make them true. Every claim in a script needs the same evidence it would need in any other advert.

A script structure that works

A reliable shape for a 20 to 30 second UGC-style advert:

Write it the way people talk: short sentences, contractions, one idea at a time. Read it aloud before you use it. If it sounds like a brochure, rewrite it.

Keeping UGC-style adverts on brand

This is the tension: the format should look casual, but the advert still has to carry your brand. The answer is to be exact about some things and relaxed about others.

Be exact about:

• The product. The real photo, the real label, the right colour, in every scene.
• Captions. Most people watch with the sound off. Set captions in your fonts and colours.
• The end card. Your logo, the offer and the call to action, set as type so it is spelled right. A short motion graphics scene does this well.
• The presenter. One presenter across a campaign makes a series feel like a series.

Be relaxed about:

• The setting, the lighting and the delivery. These should feel human.

This is why we build UGC-style adverts on a brand kit rather than from a blank prompt. The kit supplies the logo, colours and fonts; the real product photo supplies the product; the format supplies the casual look. Our guide to what a brand kit is explains how that works.

How to make one on Reigniter

1. Set up the brand and product. Brand kit and a clean product photo.
2. Start a UGC advert from Discover. Pick a UGC flow and describe the product, the audience and the one benefit.
3. Choose a presenter from the avatar library, or your own digital twin.
4. Edit the plan. Tighten the hook and the call to action. Editing the plan is free.
5. Storyboard. Check a still frame for each scene: the presenter, the product, the setting.
6. Build. Presenter scenes are lip-synced to the voice. Product inserts use a model chosen for believable handheld motion, so they cut cleanly between the talking beats. The render window shows the credit cost before you confirm.

The AI video page shows the formats, and we explain how we pick models for each shot in Kling vs Veo vs Seedance for product adverts.

Test hooks, not whole adverts

The cheapest way to improve UGC-style adverts is to keep the body and test the opening. Make three hooks for the same advert and let performance data choose.

On Reigniter the Batch option renders an advert across several presenter and hook combinations, and shows its price before it runs. Language versions are rendered again from the same storyboard, so a presenter's mouth matches the new words rather than being dubbed over.

A final checklist

Before a UGC-style AI advert goes live, check that:

• The first line works without sound, with captions.
• The product is your real product in every scene.
• Nothing in the script presents the presenter as a real customer.
• Every claim is one you could defend.
• You have the rights to every face, voice and track.
• The end card has your real logo and one call to action.

Video starts on the Pro plan at $30 a month with 48 credits, and every action shows its credit cost before it runs. See pricing for details.

Most brand guidelines fail in one of two ways. Either they are a beautiful 80-page PDF that nobody opens, or they do not exist and the brand lives in one designer's head. The fix is the same for both: a short document of specific rules, and a brand kit that stores the assets those rules describe.

This guide gives you a structure you can copy, explains what to write in each section, and shows how to turn the finished guidelines into assets.

What brand guidelines are for

Brand guidelines exist so that someone who has never met you can make something that looks and sounds like you. That someone might be a freelance designer, a new marketing hire, a printer, an agency, or, increasingly, a piece of software.

That last reader changes how you should write them. A person can interpret "use our colours generously". Software cannot. Rules that are specific enough for a tool to follow are also clearer for people, so write for the stricter reader.

A useful test: for every rule, ask "could someone check a finished asset against this and say yes or no?" If not, rewrite it.

A free brand guidelines structure

Copy these sections in order. Skip any that do not apply to what you make.

1. Brand basics

Write the brand name exactly as it should appear, including capitals and spacing. Add a one-sentence description of what you do and who it is for. This sentence is useful far beyond the guidelines: it is the seed for adverts, scripts and product copy.

2. Logo

This section carries most of the weight. Include:

• The primary logo, as a file, with a transparent background.
• Clear space, as a rule someone can measure, for example "the height of the letter R on all sides".
• Minimum size, in pixels for screen and millimetres for print.
• Backgrounds, stating which versions go on light and dark, and that the logo never sits on a busy photo without a panel behind it.
• Placement defaults for video, such as "bottom right, small, from the first frame to the last".

Keep to one primary logo if you possibly can. Every extra variant is another thing to get wrong.

3. Colour

List each colour with its hex value (plus RGB, and CMYK or Pantone if you print). Then say how each one is used:

• Primary: the colour people remember you by.
• Secondary: two to four supporting colours.
• Neutrals: a dark for text and a light for backgrounds.

Add a ratio if it helps, for example "mostly neutral, primary for emphasis, secondary for charts". Vague instructions such as "vibrant" are not rules.

4. Typography

Name the fonts. Give a heading font and a body font at most, the weights you use, and a simple scale (for example, headline, subhead, body, caption). Note the licence, because a font licensed for print is not always licensed for video or advertising.

5. Imagery

Describe the photography or illustration style in concrete terms: lighting, setting, colour grade, how people are shown, and what you never show. Two or three reference images do more than a paragraph.

6. Tone of voice

Three to five adjectives are a start, but examples are what make tone usable. For each trait, write one sentence that is on brand and one that is not. Include words you always use and words you never use.

7. Motion and video

If you make video, add a few rules: how the logo appears, how long an end card holds, how captions look, how fast transitions are. These rules are what make a series of adverts feel like a series.

8. Right and wrong

Show common mistakes: a stretched logo, a colour outside the palette, the wrong font, text on a busy background. People learn faster from a wrong example than from a rule.

How to write rules people and tools can follow

A few habits make the difference between guidelines that work and guidelines that decorate a shared drive.

1. Use numbers. "Clear space of 24px at 1080p" beats "give it room".
2. Prefer one answer. One logo, one primary colour, one heading font.
3. Say where, not just what. "Logo bottom right on every video frame" is a placement rule.
4. Write the exceptions down. If the logo goes top centre on vertical video, say so.
5. Date the document. When the brand changes, people need to know which version is current.

Turn the guidelines into assets

A guidelines document describes the brand. A brand kit stores it. The step most teams miss is connecting the two, so that assets are made from the kit rather than from someone reading the PDF and doing their best.

Here is the order that works:

1. Store the kit. Logo file, brand name, primary colour, palette and fonts, in the tool you make things with. Our explainer on what a brand kit is covers what belongs in it.
2. Make one still. A product shot or a poster. Check it against every rule in sections 2 to 4.
3. Fix the kit, not the asset. If the colour looks wrong, change the kit value, so the fix carries into everything that follows.
4. Then make the bigger things. Video adverts, motion graphics, training modules. Each one reads the same kit.

On Reigniter that loop lives in one place. You can import an existing guidelines PDF to fill the kit, or start from nothing with the Brand Generator, which proposes logo directions, draws the one you pick, writes the kit and opens a guidelines document showing the mark, the palette with its colour values, typography and pattern. That document can be saved as a PDF to hand to a designer or printer. From the same kit you can then make product stills and image ads, video adverts and training.

A quick checklist

Before you call the guidelines finished, check that:

• The logo section has a real file, clear space, a minimum size and background rules.
• Every colour has an exact value and a stated use.
• Fonts are named, with weights and a licence note.
• Tone of voice has examples, not only adjectives.
• There is at least one page of wrong examples.
• The same logo, colours and fonts are saved as a brand kit in the tools you use.

If you only do one thing from this guide, do the last one. Guidelines that live only as a PDF depend on everyone reading them. A kit that every asset is made from does not. When you are ready to put the kit to work, how to make on-brand video ads with AI walks through the next step, and brand kits and guidelines are on Reigniter's free plan (see pricing).

If you have been asked to "make it SCORM compliant", you are not alone in wondering what that means. SCORM and xAPI are the two standards most training tools mention, and the jargon hides a simple idea: how a course tells a system who took it and how they did.

This guide explains both in plain terms, compares them, and covers the option many small teams overlook: not needing an LMS at all.

What SCORM is

SCORM stands for Sharable Content Object Reference Model. In practice it is two things:

1. A package format. Your course is a zip file with a manifest that tells the LMS what is inside and how to launch it.
2. A conversation with the LMS. While the learner takes the course, it reports back: started, progress, score, completed, passed or failed.

The LMS stores those results against the learner's account. You see them in the LMS's reports.

SCORM 1.2 vs SCORM 2004

There are two main versions. SCORM 1.2 is the older and simpler one, and it is the version almost every LMS supports. SCORM 2004 added more detailed sequencing rules, but support varies between systems. For most workplace training, SCORM 1.2 reports everything you need.

Reigniter exports SCORM 1.2 only. If your LMS insists on SCORM 2004, check whether it also accepts 1.2 (most do).

What xAPI is

xAPI (sometimes called Tin Can) is a newer standard that does not care where learning happens. Instead of a package talking to an LMS, the course sends short statements to a Learning Record Store (LRS). Each statement reads like a sentence: Sam answered question 3 correctly.

That makes xAPI more flexible. It can record learning outside an LMS, detail every interaction, and feed a data warehouse or a dashboard you build yourself. The catch is that you need an LRS, and someone has to make sense of the data.

SCORM 1.2 vs xAPI at a glance

They are not mutually exclusive. A SCORM package can also send xAPI statements.

How Reigniter packages training

When you export a rendered module from Reigniter, you get a zip ending in -scorm12.zip. You set a pass mark from 1 to 100 (the default is 80), download the zip and upload it to your LMS as a SCORM 1.2 course, without unzipping it.

What your LMS receives depends on the module:

Three things are worth knowing before you choose this route:

• The video streams from Reigniter. The package is small and learners need internet access while they take the module. It is not a self-contained offline package.
• Updates flow through. Re-render the module and the package in your LMS plays the new version, with no re-upload.
• The share link matters. A package for a module with checkpoints plays through the module's share link. If you revoke that link, the package stops working.

If you also run an LRS, tick the xAPI option when you export, enter your statements endpoint and the authorisation header your LRS gave you. The package then sends a statement when the learner starts, answers a question and completes. Reigniter does not host an LRS; statements go only to the endpoint you enter.

SCORM export and xAPI are included on the Interactive Training plan with unlimited SCORM 1.2 exports and no sales call. See pricing.

When you do not need an LMS at all

Here is the question worth asking before you buy an LMS to host a handful of videos: what do you actually need to know?

For many small teams the honest answer is:

• Who was asked to take it.
• Who finished, and when.
• What they scored, and whether they passed.
• Proof, such as a certificate, for audits or client reporting.

An LMS gives you that, but it also brings a licence, user management, an admin interface and a learning curve for everyone. If your training is a set of short modules for a known group of people, a tracked share link can cover the same needs.

On Reigniter's Interactive Training and Agency plans, the share player can:

• ask each learner for their name and email, or admit only people you invited,
• limit each person to one attempt and set a timer and a pass mark,
• show who is taking a module live, with every answer, score and time taken,
• export records as CSV, and give each learner a page with a printable certificate.

We go through that setup in detail in training without an LMS.

Which should you choose?

A quick way to decide:

• You already have an LMS: export SCORM 1.2. It fits into the reports you already use.
• You have an LMS and an LRS, and want detailed data: export SCORM 1.2 with xAPI switched on.
• You have no LMS and a known list of learners: use the tracked share link and skip the LMS.
• You need anonymous, open training (for customers, say): a plain share link with views, completions and answer counts is enough.

Whichever route you pick, the content work is the same. If you are still turning slides into modules, start with PowerPoint to training video, and the employee training page shows the whole product.

Synthesia is one of the best-known names in AI video, and for good reason: it made AI-presenter video practical for training teams. If you are reading this, you are probably weighing whether it is the right fit, or whether you need something that also covers the marketing side.

We make Reigniter, so we are not neutral. We have tried to be fair anyway, including about where Synthesia is the better choice. Synthesia's features and prices change, so check their site for the current details before you decide.

The short answer

• Synthesia is built around AI-avatar video. It is strong on avatars and languages, and it is a mature, focused product.
• Reigniter is built around a brand system. It builds or imports your brand kit, then makes video adverts, product stills, motion graphics and AI-presenter training from it, with SCORM 1.2 and learner tracking included on the Interactive Training plan.

If training video is the only thing you make, Synthesia deserves a close look. If the same team also makes adverts and brand assets, or you want SCORM and tracking without a sales call, read on.

Where Synthesia is stronger

Honesty first.

• Languages. Synthesia supports 160+ languages. Reigniter's narration voices (ElevenLabs v3) cover 70+ languages, and learners can switch language mid-video across nine voice languages. For a global rollout in dozens of languages, Synthesia is ahead.
• Avatars. Synthesia's avatars are a core strength and a big part of why teams choose it. If the presenter is the whole product for you, that matters.
• Focus. Synthesia does one category of video and has refined it for a long time. A focused tool is often simpler to adopt for a single use case.
• Entry price. Synthesia's Starter plan is $29 a month, a dollar less than Reigniter Pro at $30.

Where Reigniter is different

Marketing video from the same brand

This is the main reason teams look at Reigniter. The same account that makes a training module can make a UGC-style advert, a product advert, a faceless advert, motion graphics for a price or a feature, product stills and posters. All of it comes from one brand kit, so the onboarding video and the launch advert look like the same company.

Reigniter can also build that brand in the first place: a brand kit and brand guidelines, generated or imported. Many tools apply a brand kit, including Canva and VEED. What is unusual is the whole loop, with training inside the same brand system. See the brand kit guide for how that works.

SCORM without an enterprise contract

Synthesia offers SCORM export on its Enterprise plan. Reigniter includes unlimited SCORM 1.2 export on the Interactive Training plan at $149 a month, with no sales call, plus optional xAPI statements to your own LRS.

Two things to know about Reigniter's packages: they are SCORM 1.2 only (not 2004), and the video streams from Reigniter, so learners need internet access and a re-render updates the course in your LMS without re-uploading. We explain the trade-offs in SCORM vs xAPI for small teams.

Training without an LMS

Most AI video tools, Synthesia included, expect an LMS to hand out the video and keep the records. Reigniter's share player can do that job itself on the Interactive Training and Agency plans: ask for names or admit only invited people, set one attempt, a timer and a pass mark, watch results live, export CSV and give each learner a printable certificate. For a small business or an agency training a client's staff, that can remove the need for an LMS entirely. More in training without an LMS.

Language switching mid-video

When every scene has a version in a language, learners can switch language in the player without reloading, and carry on from the same moment. Each language version is rendered again from the same storyboard, so the presenter's mouth matches the new words.

Side by side

Prices are monthly in US dollars as we understand them at the time of writing.

How pricing compares in practice

The plans are not like for like, so compare on what you need:

• Avatar training videos only, no LMS export: Synthesia Starter or Creator is likely the simpler choice.
• Training with SCORM for your LMS: compare Synthesia Enterprise with Reigniter Interactive Training at $149 a month.
• Training plus marketing video on one brand: Reigniter Interactive Training, or Agency at $199 a month with 318 credits if you also run campaigns for clients.

On Reigniter, a credit is roughly 25 cents of AI provider cost, every action shows its credit cost before it runs, and failed jobs are refunded. See pricing for the current plans.

Which should you choose?

Choose Synthesia if:

• you need many languages,
• the avatar is the heart of your content,
• you only make training and presenter video.

Choose Reigniter if:

• the same team makes training and marketing video,
• you want one brand kit behind every asset,
• you want SCORM 1.2 without an enterprise contract,
• you would rather track learners from a share link than run an LMS.

To see the training product, visit the employee training page. If you want to talk it through, contact us.

Creatify is one of the tools performance marketers reach for when they need a lot of AI video adverts quickly. If you are comparing it with other options, the right choice depends less on which tool is "better" and more on what problem you are solving: ad volume, or brand consistency across everything you make.

We make Reigniter, so we have a point of view. We have tried to be fair about where Creatify is the stronger choice. Features and prices change, so check Creatify's site for current details.

The short answer

• Creatify is built around AI adverts at scale: avatars, URL-to-video and a performance focus. It is rated 4.8 on G2.
• Reigniter is built around a brand system. It builds or imports your brand identity, then makes video adverts, product stills, motion graphics and AI-presenter training from it.

If you mainly need to produce and test a high volume of ads, Creatify is well suited to that. If you need the ads, the product stills and the training to look like one company, keep reading.

Where Creatify is stronger

• Ads at scale. Creatify is designed around producing many ad variations quickly, which is exactly what performance teams need.
• Performance focus. The product is shaped around direct-response advertising and testing.
• Avatars and URL-to-video. Both are core to how it works, and they make it fast to go from a product page to an advert.
• Reviews. A 4.8 rating on G2 reflects a product many customers are happy with. Reigniter is new and does not have public reviews yet.

Where Reigniter is different

It starts with the brand, not the ad

Most ad tools start from a brief and produce an advert. Reigniter starts one step earlier. It can build the brand identity itself, with a brand kit and brand guidelines, or import an existing one from your website or a guidelines PDF. Every asset afterwards is made from that kit: the logo is your file, the colours come from your palette and on-screen type is set in your fonts.

That matters most when you make a lot of different things. A brand kit applied inside one ad tool keeps that tool's ads consistent. A brand system behind every format keeps the adverts, the Amazon stills, the launch poster and the onboarding video consistent with each other. Our guide to what a brand kit is explains the difference.

Stills and brand assets, not only video

Reigniter makes product shots, posters, static web adverts with headlines set in real type, logos and brand guidelines documents. Brand kits, guidelines and stills are available on the free plan. See image ads.

Motion graphics for exact text

Prices, offers, feature call-outs and app screens need to be spelled exactly. Reigniter's motion graphics scenes animate real type in your fonts and colours, with no AI video model drawing the letters. See motion graphics.

Training inside the same brand

This is the part ad tools do not usually cover. The same account can turn a PDF, PowerPoint, Word file, URL, prompt or script into an interactive training module with an AI presenter, quizzes and branching, then export SCORM 1.2 or share it as a tracked link with named learners and certificates. For a brand that onboards staff, trains resellers or teaches customers, that keeps the training on brand too. More in training without an LMS.

Side by side

Keeping identity consistent at volume

Whatever tool you use, consistency at volume comes from the same habits:

1. Fix the brand before the brief. One logo, a fixed palette and named fonts, saved as a kit.
2. Always supply the real product. A clean product photo stops every variation inventing its own packaging.
3. Vary the hook, not the brand. Test openings, presenters and calls to action; keep logo, colours and end card fixed.
4. Set exact text as type. Prices and offers belong in motion graphics or typeset overlays.
5. Check a storyboard frame before rendering. Catching drift in a still is much cheaper than in a finished video.

Our step-by-step on-brand video ad workflow goes through each of these, and UGC-style AI adverts covers the format most performance teams test first.

Pricing, briefly

Reigniter's plans, in US dollars per month:

A credit is about 25 cents of AI provider cost, every action shows its credit cost before it runs, and failed jobs are refunded. Check Creatify's site for their current plans, and our pricing page for ours.

Which should you choose?

Choose Creatify if:

• you mainly need a high volume of performance adverts,
• avatars and URL-to-video are the core of your workflow,
• public reviews matter to your buying process.

Choose Reigniter if:

• you need one identity across adverts, stills, motion graphics and training,
• you do not have a finished brand kit yet and want it built,
• your team also trains staff, partners or customers.