Customise ChatGPT's tone, warmth, and structure: what European professionals need to know

What the controls actually do

The new settings live inside ChatGPT's personalisation panel and offer a level of granularity that previous versions entirely lacked. Users can move warmth from cool and neutral to notably friendly. Enthusiasm can be dialled from measured and concise to energetic and expressive. A separate toggle governs structure: responses can arrive as flowing prose or as organised headers and bullet points, depending on the task at hand.

For professionals who have spent months crafting repetitive prompt prefixes just to enforce a consistent house style, the time savings will accumulate quickly. Set your preferred structure once and subsequent queries follow it automatically. Set your preferred tone once and you stop specifying it in every request. Across thousands of interactions, that is a meaningful productivity gain.

The controls also interact with ChatGPT's memory system in useful ways. The model learns preferences over time and applies them contextually. A user who defaults to formal register for client-facing drafts but occasionally needs casual creative copy can have both needs served as ChatGPT learns which context calls for which style.

Why this matters more in Europe than elsewhere

Default AI tones trained predominantly on American English conventions do not suit every European working context. British English has its own formality conventions that differ markedly from US defaults. French and German business communication norms are more formal still. Multilingual teams in Brussels, Amsterdam, or Zurich are regularly producing outputs in two or three languages simultaneously, and a single default tone applied across all of them will feel off to at least some recipients.

Professor Luc Steels of the Vrije Universiteit Brussel, whose research covers language and AI cognition, has long argued that language models must account for sociolinguistic context to be genuinely useful rather than merely functional. The new controls are one practical step toward that goal, even if they do not fully resolve the deeper challenge of cultural register across European languages.

Equally relevant is the EU AI Act, which entered its phased enforcement schedule in 2025. The Act does not mandate specific tone settings, but it does impose transparency and accuracy obligations on AI-assisted outputs in high-risk categories. Dr. Kilian Gross, head of the European Commission's AI policy unit, has noted publicly that the Commission expects deployers, not just developers, to take responsibility for ensuring AI outputs meet applicable standards. That obligation does not disappear because a user has switched to a warmer tone setting.

Professional use cases across EU and UK sectors

The range of settings unlocked by these controls maps neatly onto the variety of tasks that European knowledge workers actually perform. Consider the following use cases and their optimal configuration:

• Customer service responses: warm, helpful, concise; moderate enthusiasm; prose format.
• Technical documentation: neutral warmth; measured enthusiasm; headers and bullet points enabled.
• Legal drafting: cool and precise; low enthusiasm; structured paragraphs without lists that might imply optionality.
• Internal communications: professional warmth; moderate formality; format matched to house style guide.
• Creative brainstorming: high warmth; energetic enthusiasm; flowing prose to avoid prematurely constraining ideas.
• Research synthesis: neutral warmth; measured tone; structured output with headers for navigability.

Each of these calls for a different configuration, and the new controls make it possible to encode that configuration once rather than prompt-engineering it on every query.

How ChatGPT compares with Claude and Gemini

OpenAI is not alone in offering personalisation, and European buyers should evaluate the competitive landscape honestly. Anthropic's Claude also provides customisation features, though with somewhat narrower bounds. Claude's constitutional AI approach produces responses with a characteristic tone that users can adjust, but the range of variation is more constrained than ChatGPT's current controls. Some enterprise users prefer that consistency; others find it limiting.

Google Gemini takes a different approach entirely, leveraging existing Google Workspace profile data to infer preferences rather than requiring explicit configuration. For organisations already deeply embedded in Google's ecosystem, this contextual personalisation can produce appropriately calibrated responses without manual adjustment. The trade-off is that it relies on Google holding and applying significant user data, a consideration that sits uneasily with GDPR obligations for many European data protection officers.

Enterprise deployment and compliance considerations

For enterprise deployments in the EU and UK, the personalisation controls introduce both opportunity and obligation. ChatGPT Business and Enterprise tiers allow administrators to establish organisational defaults while permitting individual users to adjust within defined limits. That architecture makes sense for large teams producing AI-assisted content at volume: it preserves brand voice and professional standards without forcing every employee into an identical configuration.

The compliance picture is more nuanced. Financial services firms operating under FCA or ESMA guidelines, healthcare organisations governed by clinical communication standards, and legal practices subject to bar council conduct rules all face a common challenge: tone settings affect style, not substance, but regulators do not always draw that line cleanly. An AI response that is warm and informal may still contain accurate regulated information, but the informal tone could itself create a mis-selling or duty-of-care problem depending on the context.

The practical advice for compliance teams is straightforward: treat personalisation settings as part of your AI governance documentation. Record which settings are applied to which workflows, test outputs at both ends of the warmth spectrum, and verify that regulatory obligations are met regardless of tone configuration. OpenAI's enterprise documentation provides guidance on administrative controls, but the substantive compliance work falls to the deploying organisation.

Accessibility and inclusion benefits

One dimension that deserves more attention than it typically receives is accessibility. The personalisation controls have genuine implications for users who do not fit default assumptions:

• Users who prefer highly structured, bulleted responses because of cognitive differences or learning preferences can enforce that format globally.
• Non-native English speakers can adjust toward simpler sentence construction and more literal phrasing, reducing comprehension load.
• Older users or those with limited technology experience may find warmer, more patient interaction styles reduce friction.
• Users who find AI's default cheerfulness grating can dial it down to something more professional and less performative.

None of this fully solves inclusion for every user group, but it shifts the default in a useful direction. AI assistants that adapt to users rather than requiring users to adapt to them are simply better products.

What European users should do next

For individual ChatGPT users, the immediate step is straightforward: open the personalisation panel and experiment. Users who have normalised the default tone may not realise how much better an adjusted configuration suits their actual working style until they try it. The OpenAI help centre documents each control and its effects in detail.

For enterprise teams, the more valuable exercise is to define optimal settings for each major workflow category before rolling out broadly. Pilot the combinations, compare outputs against your house style guide, and document the results. Establishing team defaults for specific use cases supports consistency without eliminating individual flexibility.

The broader point is that AI assistants are maturing into genuinely customisable professional tools. Personalisation is becoming a competitive differentiator, and the organisations that treat these controls as a serious configuration decision rather than a novelty will get measurably better outputs. OpenAI's December 2025 update is one step in that maturation; expect continued development as ChatGPT, Claude, and Gemini compete on how well they adapt to individual users, team contexts, and regulatory environments.