AI video · Case study

Root Flow Flex: making an invention ad before a prototype exists

An unmanufactured plant-pot insert needed to become understandable. The maker delivered a story with a fictional apartment, character and see-through pot shot. The images explain a proposed design; they cannot prove how a physical product performs.

Published · Reviewed

Watch the maker's 75-second ad. Its apartment, people and product-use scenes are generated, not footage of a working prototype.Poster source: Crossroad Media's original site cover and maker-embedded YouTube film; BitShovel does not host the media · Poster source · Watch original film
Delivered work · maker-reported75-second ad; maker says the client reviewed and signed off

The film is publicly watchable. Payment, review and sign-off are the maker's account; independent client confirmation, this project's fee, receipts and subsequent performance are not disclosed.

Check this result

The ad begins with overwatering, then moves through the pot wall to make an unbuilt mechanism visible. It shows a possible client task for AI video and raises a harder question: when a film can make an unbuilt product look real, how do makers keep product claims, client review and business evidence straight?

01

Verified material

Watch the 75-second film: what does it show?

The public film presents a problem, a proposed product and an improved-life payoff.

Calixte LeBlanc embeds the Root Flow Flex ad on Crossroad Media's site and gives its length as 75 seconds. In a Vancouver apartment, a character discovers standing water around plant roots; an insert appears, followed by thriving plants and a visiting friend. Readers can inspect the film's story, pace and clarity themselves.

The maker says the apartment and character do not exist, the imagery is AI-generated and the product has not been manufactured. This is a visual explanation of a pre-prototype design, not filmed evidence of use. The published film establishes a completed work, not the insert's horticultural effect, market demand or sales.

Sources for this section: Calixte LeBlanc: No Prototype. No Problem. · Root Flow Flex film embedded on Crossroad Media's page

02

Public account

The client's communication problem and the product's real state

The maker says the client had a patent and drawings but no physical prototype to film.

In the maker's account, the client had a patent and schematics for Root Flow Flex, but no prototype. The proposed flexible insert sits at the bottom of a decorative pot, lifting roots above standing water; vertical tubes are intended to carry air and water. Claims such as use in pots without drainage holes remain product claims. We have not examined the patent, a prototype test, a growing trial or customer feedback, so the ad cannot validate function.

The assignment was concrete: help prospective partners understand the design and use setting without a physical item to film. The maker lists licensing, investor meetings, crowdfunding and retail buyers as possible uses for this kind of video. He does not report that this film secured a licence, investment, crowdfunding result or order.

Sources for this section: Calixte LeBlanc: No Prototype. No Problem.

03

Public account

How drawings became a legible scene

The maker built product renders from structural references and placed the central claim in a pot cutaway.

The maker shares a product render based on the client's drawings and an image cutting through the pot. He says the vertical tubes needed precise prompting to preserve the patented geometry. The underlying drawings, exact prompts and iteration history are not public, so readers cannot independently compare every shape with the patent. What can be inspected is the chosen illustration, not a design-validation result.

He gives the explanation to a slow push that appears to pass through the pot wall, revealing the insert below the soil and a before-and-after view of water. This imagined interior is useful because a physical camera cannot pass through an opaque pot. It is a proposed mechanism shown visually; without a prototype, the shot cannot prove drainage, aeration or plant growth in practice.

Documented process

Turning an unfilmable design into a reviewable deliverable

The links come from one maker page. They show selected final assets and the maker's process account, not a controlled comparison with the original schematics.

  1. Input · client patent and drawings

    The maker says the project began with client materials; the original drawings, patent number and full brief are not provided on this page.

    Read the maker's account of client materials
  2. Design · product render

    The site shows a render of the insert and vertical tubes based on the drawings. It is generated imagery, not a prototype photograph.

    Open the maker's product render
  3. Explanation · pot cutaway

    The cutaway explains the proposed mechanism and is public to inspect; it does not measure real-world performance.

    Open the maker's cutaway
  4. Delivery · client review

    The maker says the client reviewed script, boards and final cut, gave notes and signed off; review records and sign-off documents are not public.

    Read the maker's account of review and sign-off

Evidence limitNo original drawings and reference history, shot-level prompts, failed footage, generation counts, full project cost or independent client materials are published. The public account cannot reconstruct a production ledger. BitShovel links to and embeds original media only.

Maker-published render of the proposed Root Flow Flex insert and vertical tubes; not a prototype photograph.Check visual source
A render made from client structural material. The underlying drawings are not public for a geometry comparison. · Crossroad Media's maker page; product render hosted by the original site · Visual source · Original material
Maker-published cutaway of a pot showing how the unbuilt insert is intended to work below the soil.Check visual source
The cutaway makes the hidden mechanism visible; it is an illustration, not a growing trial. · Crossroad Media's maker page; cutaway hosted by the original site · Visual source · Original material

Sources for this section: Calixte LeBlanc: No Prototype. No Problem. · Crossroad Media: Root Flow Flex product render · Crossroad Media: pot cutaway image

04

Public account

From reference images to shots and client review

The maker names the tools and shot-selection approach, without a reproducible shot log.

The maker lists Claude and ChatGPT for script and shot structure; Midjourney for character and setting references; Nano Banana Pro, GPT Image 2 and Magnific for stills, compositing and product renders; Higgsfield Seedance 2.0 for shot-by-shot motion; ElevenLabs for voiceover; and Milanote for boards, references and revisions. Tools make material, while the problem–solution–payoff story, treatment of product claims and deliverable choices remain maker and client decisions.

He says each shot was generated, reviewed, diagnosed and refined, while script drafts were checked against the client's product claims. He also describes client reviews, notes, revisions and final sign-off. The page does not publish bad takes, shot prompts, generation settings, review comments or costs. We can learn where checks sat in the workflow without pretending to know which prompt change fixed a frame.

Sources for this section: Calixte LeBlanc: No Prototype. No Problem.

05

Public account

What the commercial record does and does not show

The maker calls it a paid client project; the published $2,500–$10,000 range is a service price range, not this job's fee.

Unlike a self-initiated brand spec ad, the maker explicitly describes Root Flow Flex as a paying-client commission, reviewed, revised and added to the client's marketing toolkit. On the same page he gives a $2,500–$10,000 range for his AI commercial work. That is a service range, not this project's contracted fee, money received or profit. No independent client confirmation, agreement, invoice, receipt, labour hours or total tool bill is available here.

The narrower finding is that the maker published a watchable deliverable and gave a first-party account of the client need and paid engagement. It cannot establish a fee amount, licensing or investment gained through the film, campaign performance or the product's efficacy. Public client records, distribution results or project accounts would each change a different part of that assessment.

Sources for this section: Calixte LeBlanc: No Prototype. No Problem.

06

Our analysis

A bounded first attempt for a similar project

Start with a verified product brief and one shot explaining the central mechanism.

Begin with drawings or product material you may use, and confirm with the inventor which claims are design intent, tested facts or hopes. Choose one mechanism that a photograph cannot explain, make a structural reference and one short cutaway test, then ask the client to correct shape or function errors. Only after that review expand to characters, settings, voice and a full ad. This sequence is an editorial suggestion drawn from the case, not a step-by-step recipe published by the maker.

To turn such work into a repeatable service, record each test's inputs, attempts, rework reason, time and tool spend. Track client review rounds, asset rights, quoted price, money received and how the film was used separately. Generative tools can make an unfilmable design visible; whether clients pay, delivery repeats and a profit remains depends on your own approvals, contracts and accounts.

Sources for this section: Calixte LeBlanc: No Prototype. No Problem.

Sources and limits

Where the material comes from