Showreels used to settle the question. A brand watched a director's best work, liked it, and signed the contract.

In 2026, a showreel tells you very little about an AI video partner. A clip generated in ten minutes can look as polished as one a studio directed for a week. The difference shows up later, in consistency across forty clips, rights paperwork, and how revisions are handled.

This checklist helps marketing teams compare providers before they sign. It covers nine checks, a scorecard, and a pilot plan you can run before committing real budget.

Start with the job, not the vendor

The right provider depends on the work. A single launch film and two hundred product variants need different partners.

Write down three numbers before you talk to anyone.

How many assets do you need each month? Which formats and channels matter? What is your budget range?

Those answers narrow the field fast. They also stop sales calls from drifting toward whatever the vendor sells best.

Add a fourth item if you can: the date the first asset must go live. A fixed launch date changes which providers are realistic. A studio with a four-week slot cannot rescue a launch that is ten days away.

The four types of AI video provider

AI video providers fall into four groups. Each solves a different problem and carries different risks.

Specialist AI studio

Typical pricing: Quoted per project

Strength: Art direction and a consistent brand look

Watch-out: Custom quotes and limited capacity

Best for: Campaigns and hero films

Agency-network production arm

Typical pricing: Enterprise contracts, rarely published

Strength: Scale across many markets

Watch-out: Minimum commitments and slower access

Best for: Global brands with large calendars

Self-serve platform

Typical pricing: Monthly subscription or credits

Strength: Low cost and instant output

Watch-out: No creative direction, so you own quality control

Best for: In-house teams testing volume

Freelance AI operator

Typical pricing: Day or project rates, often EUR 500 to 5,000

Strength: Flexibility and low price

Watch-out: Limited depth, rights cover, and backup

Best for: Small one-off tasks

A specialist studio designs a look, directs production, and delivers finished files. Trippy Pictures sits in this group, so weigh our view with that in mind. Typical turnaround is two to four weeks, against four to eight for a traditional shoot.

Traditional production often costs EUR 10,000 to 50,000 or more per asset. Studios position themselves below that range for comparable output. Self-serve platforms cost a fraction, but your team does the directing.

An agency-network arm suits brands already inside a large agency group. Access usually runs through that relationship, and minimum commitments are common. A freelancer can be a smart choice for a small, low-risk task.

Many brands combine groups. A studio sets the look, and a platform produces cheap daily variants.

The checklist at a glance

Use these nine checks to compare every provider on your shortlist. The next section explains each one.

Proof of craft: Published client work and named people who direct it.

Brand consistency: A clear method for keeping products identical across clips.

Workflow transparency: Defined review points between brief and delivery.

Rights and licensing: A written license and clear ownership of final files.

Compliance plan: A view on disclosure labels and provenance data.

Data security: Written rules for how your assets are stored and used.

Itemized pricing: A quote broken down by deliverable, format, and revision.

Realistic turnaround: Milestone dates instead of a single vague deadline.

References: Past clients willing to talk about what went wrong.

The nine checks in detail

1. Proof of craft

Generation is easy now. Direction is not. Ask who actually makes the creative decisions on your project, and look for a named director or creative lead.

Look for published, approved client work rather than demos. Spec reels show what a tool can do. Client work shows what a team can deliver under real feedback.

Ask for:

Two or three published campaigns with the client's name attached.

The names and roles of the people working on your project.

Awards or festival recognition, if the team claims craft heritage.

2. A method for brand consistency

Brand consistency is where weak providers fail. A product that changes color between clips reads as an error, not a style.

Strong providers train custom models on your own reference imagery. That locks in packaging, materials, and palette across dozens of clips. Our guide to brand consistency in AI video explains the mechanics.

Ask for:

A description of how packaging, logos, and colors are controlled.

A sample of ten clips showing the same product from different angles.

The process for correcting a clip that drifts off-model.

3. A transparent workflow

You do not need to know every tool. You do need to know where humans review the work.

Ask for a walkthrough from brief to delivery. Good providers can name their review points and who signs off at each one.

The process should also start with a proper brief. Our guide on how to brief an AI studio lists what to prepare. Providers who ask sharp questions early usually deliver cleaner results.

Ask for:

A stage-by-stage timeline with approval points.

The number of revision rounds included.

Which steps use AI and which use human craft.

4. Clear rights and licensing

Rights are the quiet risk in AI video. Contracts written for traditional shoots often say nothing about generated footage.

The US Copyright Office's January 2025 report concluded that prompts alone do not make a user an author. Purely AI-generated material is not protected, but human-authored expression and creative selection can be.

That affects how much exclusivity you can claim over a finished ad. A provider who directs, edits, and composes the footage strengthens your position. A provider who simply hands over raw generations does not.

Laws differ by country, so ask counsel to review your contract. Our explainer on who owns AI commercial footage covers licensing in more depth.

Ask for:

A written commercial license covering your channels and territories.

A statement of who owns final files and project assets.

Indemnity terms if a third party challenges the footage.

5. A disclosure and compliance plan

The rules are changing quickly. The European Commission says the AI Act's transparency rules take effect in August 2026. Providers must make AI output detectable as machine-made, and deepfakes must be clearly labelled.

Ask how the provider supports labeling. Provenance standards such as Content Credentials attach origin and edit history to a file. A provider with no view on this is behind the market.

Ask for:

Their position on disclosure for your markets and channels.

Whether delivered files can carry provenance metadata.

Who handles it if a platform asks you to label content.

6. Data security and GDPR

Your product images, brand books, and unreleased launches are sensitive. Find out where they go once you hand them over.

Ask whether your files train shared models or stay in a separate project environment. European brands should also ask where data is processed and request a data processing agreement. Independent certification helps, and ISO/IEC 27001 is a common benchmark.

Ask for:

Written confirmation that your assets do not train shared models.

The processing location and a signed data processing agreement.

Access controls for your project files.

7. Itemized pricing and scope

A single number on a quote hides a lot. Ask for pricing broken down by deliverable, format, and revision round.

Pricing models differ widely, and each has a catch. The list below covers the four you will meet most often.

Per project

How it works: Fixed quote for defined deliverables

Watch for: Scope creep and revision limits

Retainer

How it works: Monthly fee for steady output

Watch for: Unused capacity and minimum terms

Tiered packages

How it works: Published bundles by length or volume

Watch for: Template-style output

Credits or subscription

How it works: Pay for generation capacity

Watch for: Credits burning fast on retries

Compare quotes against the traditional benchmark of EUR 10,000 to 50,000 or more per asset. A very low quote deserves questions about quality and rights. A very high one should come with proof of craft.

Ask for:

Cost per deliverable, including every format you need.

Included revision rounds and the price of extras.

Whether delivery covers 16:9, 9:16, and 1:1 versions.

8. Realistic turnaround

AI shortens timelines, but it does not remove them. Specialist studios typically quote two to four weeks, against four to eight for a traditional shoot.

Self-serve tools return clips in minutes, yet review and correction still take time. Ask for milestone dates rather than a single deadline.

Approvals are the usual bottleneck. Name one decision-maker on your side and agree on feedback windows early.

Ask for:

Milestone dates from kickoff to final delivery.

Turnaround time for each revision round.

Any rush fees for urgent changes.

9. References that hold up

Talk to two past clients if you can. Ask what went wrong and how the team responded. Problems are normal, and the recovery tells you more than the success.

A confident provider will offer references without being pushed. One that deflects deserves a harder look.

Ask for:

Two named clients willing to take a short call.

An example of a project that needed major revisions.

Evidence of repeat work with the same brand.

Questions to ask on the first call

The first call sets the tone for the whole relationship. Prepare a short list and listen for specifics.

Start by asking who will direct your project day to day. A named person is a better answer than "our team." Then ask how the provider keeps your product identical across clips, and listen for a concrete method rather than a promise.

Ask what happens when a clip looks wrong. The answer reveals the revision process and its cost.

Follow up on ownership by asking who holds the final files and project assets. Expect a written license in reply.

Ask how your data is handled. Look for storage rules, processing locations, and training exclusions. Finish by asking to speak with a past client, since confidence here usually signals real delivered work.

Take notes and compare the answers across providers. Vague replies are data too.

Mistakes buyers make most often

Even careful teams repeat the same errors. Knowing them in advance saves weeks.

The most common error is choosing on price alone. The cheapest quote often omits revisions, formats, or usage rights. You pay for those later.

Skipping the brief is the next one. Vague direction produces vague output, and the invoice grows with every extra round. Buyers also judge a single clip, although one striking sample says nothing about consistency across thirty.

Others ignore the legal team. Contracts that predate AI rarely cover generated footage, so involve counsel before kickoff.

Some buy a tool when they needed a team. A cheap subscription still needs a person to direct, review, and fix output. Leaving approvals open is the last trap, because slow feedback adds more delay than any generation step.

These mistakes share one habit. Buyers compare outputs when they should compare processes.

Score your shortlist

Rate each provider from one to five on the nine checks. Use this list to judge the quality of the answers you receive.

Proof of craft

Strong answer: Published, approved client work

Red flag: Only demos and spec reels

Brand consistency

Strong answer: Custom models trained on your assets

Red flag: "We will just prompt it carefully"

Workflow

Strong answer: Named review points and human sign-off

Red flag: Black-box delivery

Rights

Strong answer: Written commercial license and clear ownership

Red flag: Silence on usage terms

Compliance

Strong answer: Labeling plan and provenance support

Red flag: No view on disclosure rules

Security

Strong answer: Defined data handling and certification

Red flag: Brand assets pasted into open tools

Pricing

Strong answer: Itemized scope, formats, and revisions

Red flag: One number with no breakdown

Turnaround

Strong answer: Milestones with approval dates

Red flag: "Fast" with no schedule

References

Strong answer: Named clients willing to talk

Red flag: No references offered

Weight the criteria by your priorities. A luxury label should weight craft and consistency heavily. A volume retailer should weight pricing and turnaround.

Run a paid pilot before you commit

Pick one real product line and one real channel. Brief the provider as you would for a full program. Review the output against your scorecard.

A good pilot should do four things:

Produce three to five finished clips in your real formats.

Include one deliberate revision round to test responsiveness.

Show how the team handles consistency across several clips.

End with a written summary of costs, timeline, and lessons.

Decide against the scorecard, not the demo. A pilot that runs smoothly is the best predictor of a healthy long-term relationship.

Read the results with a cool head. Did the product stay accurate in every clip? Did the team answer questions quickly and plainly?

Compare the final invoice with the original quote. Any gap tells you how the provider handles scope. If the pilot passes, scale up in stages rather than signing a twelve-month commitment at once.

Frequently asked questions

How many providers should I shortlist?

Three is enough. Pick two close comparisons and one wildcard from a different category. More than that slows the decision without improving it.

Should I choose a studio or a self-serve tool?

Studios suit launches and hero content. Tools suit testing and volume. Many teams use both, with the studio setting the look.

What should an AI video contract include?

Include scope, deliverables, formats, revision rounds, timeline, commercial license, ownership, data handling, and indemnity. Ask counsel to review it before signing.

How much should I budget for a pilot?

Scope it as a small fraction of your intended program. Ask for a fixed quote tied to three to five finished clips. That keeps the test cheap and the comparison fair.

Is AI video suitable for regulated industries?

Often yes, with the right controls. Check data handling, approval workflows, and disclosure rules. Fast iteration can help teams whose approval cycles run long.

Do I need in-house AI expertise?

Not with a studio, which handles generation for you. A self-serve tool needs someone who can brief, review, and fix output. Plan for that time before you buy.

What are the biggest red flags on a sales call?

Watch for vague answers about rights, data, and consistency. Be wary of providers who only show their best clip. A refusal to share references or a process outline is another warning sign.

Can a provider work from my existing brand assets?

Yes, and good providers will ask for them. Logos, product photography, color specifications, and past campaigns all help. The richer your reference library, the more accurate the results.

Final take: buy the process, not the demo

A strong demo is easy to produce. A strong process is what protects your brand, your budget, and your legal position. Judge providers on the nine checks, not on their best clip.

Teams with a flagship campaign should weight craft and consistency most. Teams with a high-volume catalog should weight pricing, turnaround, and workflow. Many teams end up needing both, so plan for a studio and a platform.

Want a partner that directs every project? Contact Trippy Pictures. We will scope a pilot around your products, your formats, and your timeline.