From a brief
to a media plan.
Bringing AI into an existing growth platform, with a clear path for customers to brief, review, refine, and generate a plan.
- Role
- Senior Product Designer · Contract
- Focus
- Product flows · Conversational UX · UI
- Timeline
- Mar 2025–Jul 2026 · Remote
Selected design screens · Illustrative content and data
Give AI a place
in the customer’s work.
FOSHO’s existing growth platform did not have AI assistance. I designed the interaction flows that connect customer inputs, AI proposals, review, refinement, and a usable plan.
This case focuses on the media-planning journey. My contribution covered product flows, conversational UX, and interface design, including a related strategy-editing state.
A conversation with decision points.
The path gives customers a way to inspect and refine the proposal before requesting a final document.
- 01Brief
Set the goal, audience, product, and budget.
- 02Propose
Review channels and individual publishers.
- 03Refine
Add campaign and market context.
- 04Confirm
Review the outline before generation.
- 05Deliver
View or download the completed plan.
An editorial summary of selected design states. The screens below contain illustrative content and data.
Structure the input.
Leave room for context.
A starting point, rather than a blank prompt.
AI Media asks for the campaign goal, product, audience, and budget in sequence. Suggested choices give customers something concrete to respond to, while free-text answers carry their own context.
The question and response stay visible in the conversation, so the developing brief can be followed.
Guide the first answer
Suggested choices help the customer begin without having to invent a prompt.
Carry the context forward
Visible question-and-answer pairs show how the brief is developing.
Keep input flexible
Free-text responses make space for product and campaign details.
Make the proposal
something people can inspect.
Move from the channel mix to individual publishers, then into the reasoning behind a recommendation.
Channel-level structure
Expandable groups organize the proposed allocation into a scannable hierarchy.
Publisher-level options
Cards expose individual placements without flattening everything into a chat response.
Progressive detail
Collapsed groups preserve the overall structure while one area is being reviewed.
Explain the recommendation at the point of review.
The publisher detail modal separates the recommendation rationale from the next action. A customer can read the explanation and choose whether to visit the media details.
A displayed fit score is illustrative; it does not establish recommendation accuracy. That would need validation in the working product.
A review step
before generation.
Keep the proposal open to revision, then make the transition into generation explicit.
Finish with a usable artifact.
The document is
the endpoint.
After the outline is confirmed, progress feedback leads to a completed state with separate View and Download actions. The design creates a plan document; it does not launch or spend money on a campaign.
What still needs to be validated.
Understanding
Can customers explain why a channel or publisher was proposed?
Revision
Do revised constraints carry consistently into the outline and final plan?
Trust
Are the recommendation rationale and underlying data useful and accurate?
This case documents the design work. Release status and business results have not been confirmed. Scores, budgets, and performance figures in the original screens are example content.