GPT OpenAi

A copywriting assistant to ensure language quality in early design stages

Within the design team, we did not have copywriters involved from the start of new initiatives. Text reviews had to be requested via a ticket to an external team, subject to availability, budget and backlog. As a result, the copy used in testing prototypes was often underdeveloped or inconsistent with the company’s voice and tone guidelines.

Open laptop showing a chat window titled "Asistente Voz y Tono", with a dark theme and sidebar menu on the left, displaying various chat options.

Designing without early access to copywriters created several risks:

  • We validated copy that later needed major revisions

  • Language consistency was lost during exploration and design phases

  • The team spent time reworking already validated copy

  • Dependence on an external team reduced agility

The Problem

Improve the quality of copy before testing, reduce external dependence and ensure consistency with the brand’s verbal identity from the design stage.

The Objective

A woman with long brown hair and glasses sits at a wooden desk, talking on a cellphone while looking at a silver laptop. She is wearing a red and white striped shirt under a denim shirt. The background features a brick wall, a white shelf with books and a camera, and a small blue cabinet with a plant on top.

Solution

I designed and configured a custom GPT-based assistant that acts as an integrated “copywriting team” within my design workflow.

To do this:

  • I organised and structured all existing documentation on voice and tone

  • I identified and resolved inconsistencies, gaps and ambiguities that could lead to incorrect responses from the model

  • I trained the assistant to understand and apply the core principles with good judgement

How it works

  • Design upload: I share a screenshot of the design

  • Text extraction: The assistant identifies and organises the copy by section

  • Guided evaluation: It analyses each text according to voice and tone guidelines

Results:

  • A list of strengths and issues with explanations

  • Specific improvement suggestions

  • The ability to ask questions about corporate language

This initiative allowed me to improve design quality through language, with a solution that is efficient, scalable and tailored to the team’s real needs. Thinking systemically and anticipating bottlenecks was key to improving both user experience and internal processes.

Key points:

  1. Agility: I could iterate faster without relying on other teams in early stages

  2. Quality: Prototype copy was much more polished by the time it reached testing

  3. Consistency: Stronger alignment with the brand’s verbal identity from the start

Impact

A screenshot of a chat interface with a message detailing instructions for uploading images in Spanish.

Sabadell Online Account: Designing the digital onboarding for new customers