# A product that assistants can describe.

> A studio project. A visibility panel for a software product that ranks in search but goes unnamed in AI answers.

- Sector: Software
- Year: 2026
- Project: Plainsheet
- Client: Self-initiated
- Services: Marketing & growth, Development

Plainsheet is a studio project in marketing and growth: a panel that shows a software team which buyer questions name their product in AI answers, and which do not.

## In the project

- Name and wordmark
- Palette of four tones
- Question list with a mark per assistant
- Chart of who is named
- Component set

## The problem

Software buyers used to compare tools through a search and ten open tabs. Many now ask an assistant for a shortlist. A product can hold a good search position and still be absent from that answer, and nobody on the team finds out, because no report shows it.

Without a view of the questions, visibility work turns into guesswork: a new page here, a rewritten heading there, and no way to tell whether either mattered.

## What we made

Plainsheet is the planning tool we invented for the project, with its name, its wordmark set in mono and a palette of paper and ink blue. The panel is the work itself: buyer questions on the left, a mark per assistant, and a chart of who is named. The technical half of the method runs on this site: every page is rendered on the server and has a Markdown copy.

## Who it is for

- The marketing lead: Reports on visibility. Needs a number that can be followed from month to month.
- The content writer: Decides what to write next. Needs to see which questions leave the product out.
- The founder: Asks whether the work is paying off. Needs one chart.

## What we did

### Marketing & growth

Used: SEO, GEO and AIO, Reports.

1. List the questions: What buyers ask an assistant about the category.
2. Record who is named: A mark per assistant, for each question.
3. Chart it: The product against the others, on one screen.

### Development

Used: Front-end, Astro.

1. Build the screens in code: The panel and its chart are built in HTML and CSS inside Astro components. They are not exported images.
2. Make them scale: Every size is relative to its container, so one drawing works as a card, a cover and a figure.
3. Check on a phone: The same drawing is checked at 390 pixels wide.

## Design decisions

1. Questions are the rows. A question is something a team can act on: write the page that answers it. So the panel is organised by question, not by assistant. Instead of: A dashboard of scores per assistant.
2. A dot per assistant, filled or empty. Named or not named is a fact. Three dots per row show at a glance where the gaps are. Instead of: Sentiment scores, which invite argument.
3. The product’s bar is the only blue one. In the chart of who is named, competitors stay grey. The eye goes to one bar. Instead of: A colour per competitor.
4. No reading without a baseline. The panel is designed to be run before any change is made, so every later reading has something to stand against. Instead of: Reporting a single reading as a result.

## How it works

1. Questions first: The list of questions buyers ask.
2. Who names you: A mark per assistant for each question.
3. Compare: A chart of who is named most in the category.

## The principle

Start from the questions buyers ask.

## What it is built to change

- Share of voice: The share of panel questions in which the product is named, per assistant.
- Questions without a mention: The rows where no dot is filled. This is the list the content plan starts from.
