The basics of the Spring 2025 promotion – Here’s how to get started
What does a prompt mean?
A prompt is an input or prompt given to an AI to produce a response. It can be a simple question, a precise command, or a more complex instruction that defines the desired style, tone and content. A well-crafted prompt will help you get more accurate, higher quality and more relevant responses from your AI.
The structure of a good prompt: R.A.C.E.
- Role (Role)
Define the role you want the AI to play.
For example: you are a recruitment professional and recruit self-directed graphic design and marketing professionals for a company in Helsinki.
- Description of the order or task (Action)
Specify what is being done now.
For example: I want to write a LinkedIn post to gain visibility and attention for a graphic designer recruitment.
- Context (Context)
Give the AI more context so that the end result is not too generic, but fits your purpose.
For example: a recruiting firm emphasises deep and long-term expertise, a unique operating model with no hierarchies and where everyone gets to work directly with clients, but also as part of a team. The company is independent and long-established, founded in 2008. The company has been ranked among the top in the Best Workplaces of the Year competition for several years.
- Rules and limits (Execute)
Give clear instructions on what you want to achieve.
For example: write a Linkedin post for your company that is 300-400 words long and highlights the company as a good employer and interesting jobs. Don’t be too effusive, avoid terms like “Best, Leading, Exceptional, Phenomenal”. The text should be factual but relaxed.
- Model response (Examples)
You should include as much background material as possible to give the final result a more personal and branded feel.
For example: here’s an example of a previous company post that you can use as a model for your text style: are you a text wizard whose stories bring brands to life? Do you love language, communication and communication – and above all, how your message impacts people? Aste Helsinki is looking for a copywriter!
Read more: AI for creative work: how we use AI tools
Sequential Prompting – Try this too!
If the entity is complex, splitting it into several successive commands may give a better result. In that case, the AI can be better guided, with each prompt refining the original instruction.
For example, prompt like this:
- Write a LinkedIn post to recruit graphic designers.
Artificial intelligence responds to something - The company is <company website address>. Self-direction and working together are at the heart of the business.
Artificial intelligence is responsible for something - The company was founded in 2008 and has been a top-ranked Great Place to Work for several years.
Artificial intelligence answers to something
And so on, until you find the result works.
Chain of Thought Prompting – logical reasoning for demanding tasks
Chain of Thought Prompting(CoT) is a style of prompting that encourages AI to think in stages. It forces the model to slow down its thinking and process a complex problem step by step, reducing hasty answers and creating more accurate, reasoned responses.
CoT promotion is particularly useful for complex computational tasks, understanding coding and algorithms, or creative problem solving.
In its simplest form, this is done by simply adding a command after the prompt: think step by step and explain your reasoning.
You can also give the AI an example of how to solve the problem step by step:
Q: If I have 3 apples and I give one away, how many are left?
Chain of thought: I have 3 apples to start with. I start with 2 apples and give one away, so 3 – 1 = 2. Answer: 2 apples.
Now solve this question the same way: if you have 5 pens and lend 2 to a friend, how many do you have left?
Text: Maija Vaara, Aste Helsinki
Experts
