Skip to main content

The end of 2022 saw a generative AI storm hitting us in the form of Chat GPT developed by Open AI. Competitor tools such as Bard.ai too were soon released to the market. These tools are essentially large language models (LLMs) which have been trained on massive dataset of text and code.

They’re also referred to as conversational AI tools since they have been designed to have conversations with humans.

Many professionals around the world use these generative AI tools on a daily basis to do their jobs better. This includes marketing professionals too.

Here is a list of use cases for generative AI for Marketers.

  1. Can help with market research
  2. Can help with personalization and segmentation
  3. Can help with content creation
  4. Can help with ad campaign optimization
  5. Can help with marketing automation6. Can help provide 24/7 customer service on marketing channels

However, the big question is – to what extent are marketing professionals taking full advantage of these tools?

That’s where prompt engineering comes into the picture. Harvard Business Review described prompt engineering as follows:

Prompt engineering is the practice of crafting the optimal textual input by selecting the appropriate words, phrases, sentence structures, and punctuation to obtain the desired output from a large language model (LLM).

There are several techniques to improve the way you prompt engineer for better outputs from generative ai.

Foundation Inc highlights the below 4 steps

1. Defining the goal or objective

The more you clearly you can describe your end goal on your prompt – the better the output would be. This could include things like specifying the

  1. Format
  2. Target Audience
  3. Desired tone of the content

2. Setting the context

The Generative AI Chatbots would be able to provide you with more relevant and accurate responses if you could provide it with background information and other context relevant to the topic.

3. Providing examples and guidance

If you can provide the chatbot with some examples that would give it a template to follow when generating an output which would be better suited to you.

4. Iterating and refining

As marketers we know the importance of optimization. During prompt engineering too – optimization is important to continuously get better results. You could test different variations of a prompt, study and evaluate the results, and refine it so that you could get better outputs which are more relevant to you.

YouTuber Jeff Su has a slightly different formula for prompt engineering.

He outlines a 6-step checklist to help you get better at prompt engineering.

1. Task

You need to start this process by clearly defining the task. It is advisable to start the task sentence with an action verb, eg: Generate, Give, Write, Analyze, etc. Also, clearly articulate what your end goal is.  

2. Context

The primary purpose behind providing context to AI chatbots is to limit the endless possibilities of results that we could get from them otherwise.

3 main inputs here would be

  • What is the user’s background?
  • What does success look like?
  • What environment are they in?

3. Exemplar

This refers to examples you could be providing the AI chatbots to produce more relevant results for you. These could drastically improve the quality of the output.

An example of exemplars could be to ask the chatbots to use the Star answer framework when giving you the output.

4. Persona

Here you are suggesting who you want the AI chatbot to be. For example you might want the AI chatbot to be a ‘Senior Product Development Manager’ or ‘Batman’!

5. Format

AI Chatbots are capable of providing you with outputs in formats best suited for your requirement. For example, you could want something in

  • Email format
  • Bullet points format
  • Code blocks format

You could get results in tables, ask for concise responses or pretty much any other type of format.  

6. Tone

You can ask AI chatbots to give the outputs using casual or formal tone of voice. Or you could ask it to provide you with a witty output or show optimism. If you can’t seem to come up with an ideal adjective, then you can describe your feeling to the AI chatbot in detail and it could figure out the tone you’re going for.

Here’s what an example prompt looks like which has all 6 elements mentioned above.

Hope you will be able to make better use of AI Chatbots going forward. If you’re looking for any assistance with marketing – do book a free consultation with us here.

Leave a Reply

Open chat
💬 Chat with us
Hello 👋
How can I help you?