KIJESTIC

Customer-facing AI + brand operations · Dasom Essence

A skin scan that never leaves the visitor's device.

Dasom Essence sells premium Korean skincare online. Kijestic built the skin scan a shopper uses to get routine guidance, and runs the script-to-publish content operation behind the brand. It is not a medical device and does not provide a diagnosis.

Get My AI Assessment See the operating system

Business need

A premium brand, sold without a counter or a consultant.

Dasom Essence is a direct-to-consumer Korean skincare brand: six core products and a set of bundles, sold entirely online. A shopper choosing between a handful of serums and toners has no esthetician in the room to ask, and a brand at this price point needs a way to guide that choice that does not read as a sales script.

Customer-facing AI

Input, analysis and guidance, on the visitor's own device.

The skin scan is embedded on the live Dasom Essence store with a single script tag. It is a working demonstration Kijestic built and operates, not a mockup.

Customer input

A photo, taken or uploaded

A shopper opens the scan and takes a photo, or uploads one. A face-detection check has to see a real, current face in frame before the shutter is allowed to open.

On-device analysis

Scored where it was taken

The photo is scored on the visitor's own device using tone-invariant rating bars across a small set of skin concerns. Nothing is uploaded to Kijestic or to Dasom as part of that scoring.

Skincare guidance

Routine and product suggestions

The result maps each concern to a suggested routine and, in shop mode, a product from Dasom's own catalog. It is not a medical device and does not provide a diagnosis.

The scan is built as a reusable, multi-tenant widget: the same engine runs in a booking-oriented mode for a service business and a shop-oriented mode for a retail brand like Dasom, configured per client rather than rebuilt.

The operating system behind the brand

Every piece of content passes the same eight steps before it goes out.

This is the pipeline Kijestic operates for Dasom's blog articles and product videos: scripts, video and captions, with a claims check before anything is posted.

Idea to publishing-ready asset Operated by Kijestic
  1. 1
    Idea
    A content need is identified: a blog topic tied to a product, or a product story for video.
  2. 2
    Brief
    The idea is scoped to a specific product and a specific claim boundary: what can be said about its ingredients, and what cannot.
  3. 3
    Script
    An article or a video script is written against that boundary.
  4. 4
    Production
    The asset is produced: an article is drafted, or a video is cut from real footage.
  5. 5
    Claims check
    An automated check compares every claim in the draft against the product's actual ingredients and a list of retired claims, and blocks anything that does not match.
  6. 6
    Creative review
    The draft is scored on its own before it is considered for release.
  7. 7
    Approval
    Kelvin reviews and approves. Video drafts are sent to the brand's own account as an unpublished draft for him to finish and post himself; nothing publishes itself.
  8. 8
    Publishing-ready asset
    A published blog article, or a video ready to post, with a record of what was checked along the way.

What this demonstrates

The same operating model, in two different places.

A shopper gets skincare guidance without a photo ever leaving their device. A piece of content reaches an audience only after it passes a documented, automated check and Kelvin's own approval. Both run inside rules set in advance, not judgment calls made in the moment they happen.

What goes out to a customer in your business without anyone checking it first?

An AI Opportunity Assessment starts with a Workflow Teardown of the one workflow that costs you the most: what it looks like today, what it could look like, and what it would take to build it.

Get My AI Assessment See a sample assessment