Two Prompt Structures That Make AI Work Harder for European SMBs

It is easy to get lost in the noise around artificial intelligence. Over the past two years it has shifted from a tech-insider curiosity into a standard operational tool across the continent — and the data confirms it. Eurostat's 2025 enterprise survey found that AI use among EU businesses climbed to 20%, up sharply from 13.5% a year earlier [1].

But the headline average hides a wide spread by company size. Only 17% of small enterprises use AI, against 30% of medium-sized firms and 55% of large corporations [1]. The real gap, in other words, isn't between SMEs and big business as a single block — medium-sized firms are adopting faster than most people assume. It's the smallest companies that are furthest behind.

One caveat is worth stating plainly, because most coverage skips it: these figures only count businesses with ten or more employees. The micro-businesses that make up the overwhelming majority of Europe's SMEs aren't captured at all — so if you run a five-person studio, you're not even in the statistics. In the UK, the Department for Science, Innovation and Technology (DSIT) puts deliberate, strategic AI deployment at around 16% of businesses with five or more staff [2]. That figure is on the conservative side; surveys that count any use of AI — including features quietly embedded in everyday software — run far higher.

If you are running a small business in London, Paris, Munich, or Dublin, you don't care much about any of these percentages. You care about practical efficiency, staying competitive, and managing tight margins.

You have probably opened a tool like ChatGPT, Claude, or a European alternative such as Mistral, typed a casual question, and received an answer that felt… fine. A bit generic. A little too robotic.

That is rarely a flaw in the model. It is usually a problem with how it was directed. There is no shortage of "prompt frameworks" online, and most are just sensible advice dressed up in an acronym. The two below are worth your time for a simple reason: they require no technical skill, and they map onto something you already do well every day — briefing a colleague. We'll call them RTC and CREATE.

The Core Idea: Match the Tool to the Job

Think of these as two tools in a shed. You don't grab a sledgehammer to drive a picture nail, and you don't reach for a pocket knife to fell an oak.

• RTC is your daily pocket knife — quick, sharp, and right for the bulk of everyday tasks.
• CREATE is your heavy machinery — built for complex, high-stakes assets like full marketing campaigns, onboarding templates, or multi-page policy drafts.

1. The Daily Workhorse: RTC

For quick, high-quality turnarounds on everyday items — drafting a client update, summarising a long PDF, brainstorming localised marketing copy — use RTC: Role, Task, Context.

When you hand a job to an employee or a freelancer, you don't shout a single verb and walk away. You give them a perspective, a specific goal, and some boundaries. RTC forces you to do the same with AI.

┌────────────────────────────────────────────────────────┐
│                      THE RTC PROMPT                     │
├───────────────┬────────────────────────────────────────┤
│ R – Role      │ "Who" the AI needs to be               │
├───────────────┼────────────────────────────────────────┤
│ T – Task      │ "What" exactly needs to be done        │
├───────────────┼────────────────────────────────────────┤
│ C – Context   │ "Why, how, and for whom" it's built    │
└───────────────┴────────────────────────────────────────┘

RTC in Action

• The lazy prompt: "Write a social media post for my boutique eco-hotel in Yorkshire."
• The RTC prompt:

[Role]: Act as a boutique hospitality copywriter specialising in sustainable luxury tourism. [Task]: Write a short, engaging LinkedIn post highlighting our new off-grid, solar-powered cabins. [Context]: Our audience is stressed city professionals in London looking for weekend eco-escapes. The tone should be serene, exclusive, and distinctly British. Focus on the feeling of disconnecting from digital noise.

The difference in output between those two is not marginal — it's the difference between something you bin and something you can post with light edits.

2. The Comprehensive Machine: CREATE

When you need a production-ready asset on the first pass, three steps aren't enough. Expand the blueprint into CREATE: Context, Role, Explicit task, Assumptions/constraints, Target audience, and Examples.

The acronym is a mnemonic, not gospel — but the six elements between them cover almost everything that separates a usable first draft from generic filler you have to rewrite.

┌────────────────────────────────────────────────────────┐
│                    THE CREATE PROMPT                    │
├─────────────────┬──────────────────────────────────────┤
│ C – Context     │ The background or broader project    │
├─────────────────┼──────────────────────────────────────┤
│ R – Role        │ The specific professional expertise  │
├─────────────────┼──────────────────────────────────────┤
│ E – Explicit    │ The exact deliverable needed         │
├─────────────────┼──────────────────────────────────────┤
│ A – Assumptions │ Constraints, boundaries, or rules    │
├─────────────────┼──────────────────────────────────────┤
│ T – Target      │ The specific end-user or reader      │
├─────────────────┼──────────────────────────────────────┤
│ E – Examples    │ A template or style reference        │
└─────────────────┴──────────────────────────────────────┘

CREATE in Action

Imagine you run a growing manufacturing supply company in Germany and need an internal procedure for handling customers who object once they realise they've been talking to an AI agent.

This is a timely example, because the rules are about to change. Under Article 50 of the EU AI Act, transparency obligations apply from 2 August 2026: people must be told when they are interacting with an AI system, clearly and at the start of the interaction [4]. A useful nuance for SMEs: that disclosure duty falls primarily on the provider of the AI system, not automatically on every business deploying it. But if you've deployed a chatbot on your own site, you still want a clear internal process for the moment a customer pushes back — which is exactly what the prompt below produces.

• The CREATE prompt:

[Context]: We are adding an automated customer-support triage chatbot to our website to handle initial inquiries. [Role]: Act as an experienced European corporate operations manager working alongside a compliance officer familiar with the EU AI Act. [Explicit task]: Draft a formal, internal, 500-word Standard Operating Procedure (SOP) describing how our team should handle complaints when customers realise they are talking to an AI agent. [Assumptions/constraints]: Assume our chatbot already discloses that it is an AI at the first point of contact, in line with EU AI Act Article 50 transparency requirements applying from August 2026. Avoid aggressive corporate jargon. Use clear, bulleted escalation steps showing exactly when a human manager must step in. [Target audience]: Our internal customer-success team based in Berlin and Dublin. [Example of tone]: "When a customer asks for a human, the handover should be seamless. We don't hide the technology — we use it to get them to the right person faster."

Notice how the constraint references the regulation correctly without overstating your own obligations. That precision is the difference between an SOP your compliance officer signs off and one they send back.

Why This Matters for the European Bottom Line

Many European business owners hesitate to use AI because of implementation costs, a shortage of specialist talent, or shifting regulation. These are real concerns — research prepared for the European Commission notes the same barriers around data accuracy, legal uncertainty, and ethics, while pointing to the efficiency gains available when firms apply AI to support functions like HR, finance, and customer relationship management [3].

The encouraging part is that the single highest-leverage change costs nothing. You don't need a software budget to treat the prompt line like a structured business brief instead of a casual search bar. That shift is purely a habit.

"The real divide in modern business is no longer between companies with large AI budgets and those without. It's between people who know how to issue clear, contextual instructions to these tools, and those who treat advanced models like a basic search query."

Anchor your daily tasks in RTC and your large-scale projects in CREATE, and you cut down the trial-and-error cycle. You stop losing hours to editing generic text, and you start getting real, measurable output from the first draft.

References & Sources

[1] Eurostat — "Use of artificial intelligence in enterprises" (2025 reference year, published December 2025). Source of the EU figures used above: 20.0% overall adoption among enterprises with 10+ employees, broken down as 17% (small), 30.4% (medium), and 55.0% (large). Note that the survey excludes businesses with fewer than 10 employees. News release: https://ec.europa.eu/eurostat/web/products-eurostat-news/w/ddn-20251211-2 Full data article: https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Use_of_artificial_intelligence_in_enterprises

[2] UK Department for Science, Innovation and Technology (DSIT) — AI Adoption Research. Survey of c. 3,500 UK businesses finding that roughly 1 in 6 (≈16%) of firms with five or more employees have deliberately deployed at least one AI technology with a defined business purpose. Broader definitions of "any AI use" report substantially higher figures. https://www.gov.uk/government/publications/ai-adoption-in-uk-businesses

[3] European Commission — "Unlocking the Potential of AI in Business Support Processes" (StepUp StartUps, 2025). A report prepared for the Commission on how AI is transforming business support processes (HR, finance, CRM) in EU companies, including the adoption barriers SMEs cite. Findings fed into the Commission's Apply AI Strategy. https://digital-strategy.ec.europa.eu/en/library/unlocking-potential-ai-business-support-processes

[4] EU Artificial Intelligence Act (Regulation (EU) 2024/1689), Article 50 — Transparency obligations. Requires that people be informed when interacting with an AI system, with the relevant transparency obligations applying from 2 August 2026. The duty to disclose chatbot interactions falls primarily on the system's provider; deployers carry related obligations. Official article text: https://artificialintelligenceact.eu/article/50/ European Commission transparency guidance: https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai