18 days ago - Technology and Innovation

The Art of Making a Good Prompt

By Martín Pefaur

The Art of Making a Good Prompt

Empowering LLM for Customer Service, Consultative Sales and Technical Support

In the dynamic world of technology, the effectiveness with which a natural language agent can assist in customer service, consultative sales and technical support depends largely on the quality of the prompt used. With large language models (LLMs) such as GPT-3.5 and GPT-4, prompt design has become a crucial skill for maximizing efficiency and reducing costs.

This article explores strategies for creating effective prompts and how good practice in this area can make the difference between efficient use of a low-cost model such as GPT-3.5 and the need to resort to more expensive models such as GPT-4.

The Strategy Behind a Good Prompt

Defining Agent Personality and Knowledge

For an AI agent to be effective in customer service, consultative sales and technical support, it is essential to clearly define his or her personality and expertise. An agent must be friendly, patient and clear in their explanations, adapting to the user's specific objectives. Creating a prompt that encapsulates these characteristics ensures that the agent can interact in a human and effective manner.

For example, a well-defined prompt might look like this:

Use Case: AgentHelper

Let's illustrate this concept with a concrete example: AgentHelper, an agent designed to assist in the creation of effective prompts. The description and configuration of this agent can serve as a template for other agents:


AgentHelper is a helpful, cooperative agent designed to assist humans in creating more effective prompts for conversational agents, based on industry best practices and specific customer needs.


Friendly, patient, clear in his explanations. Seeks to understand the human's goals well and adapts to them.


Is always up to date with the latest trends and developments in the creation of prompts. Understands how to create effective prompts that clearly define the personality, knowledge and skills of the desired agent.


- Can guide the human step-by-step in creating the prompt, giving examples and constructive feedback.

- Strictly maintains the parameters of the human-defined prompt, but proposes changes if he/she believes they could improve the effectiveness of the prompt.


1. Do not alter the prompt given by the human without their consent.

2. Respond from the perspective of the agent described in the prompt.

3. Ask for clarification if the human's instructions are unclear.

4. Provide example responses if the human requests them.

5. Make sure you understand the human's objective and stick to it.

Impact of an Efficient Prompt Strategy

Cost and Efficiency

The cost difference between using a model such as GPT-3.5 and GPT-4o can be significant. GPT-4o, being a more advanced model, generally provides more accurate and contextually rich responses, but at a higher cost. However, with a well-designed prompt, GPT-3.5 can be nearly as effective as GPT-4o in many customer service, sales and technical support applications.

By clearly defining the prompt parameters and training the agent to follow specific rules, the use of cheaper models can be optimized without sacrificing too much in terms of quality. This is particularly useful for repetitive tasks and frequent queries, where efficiency and consistency are key.

Integration with Communication Channels

To maximize the usefulness of AI agents, it is crucial to integrate them with common communication channels such as WhatsApp and Instagram. This integration allows companies to maintain a constant and accessible presence on platforms where their customers are already active.

For this task I never tire of recommending tools like Chatizalo.com who offer robust solutions for integrating AI agents with multiple messaging platforms, facilitating the implementation and maintenance of these systems.

Creating effective LLM prompts is an art that can transform the way companies handle customer service, consultative sales and technical support. By clearly defining the agent's personality, knowledge, skills and rules, efficiency can be maximized and costs can be significantly reduced. In addition, integrating these agents with communication platforms such as WhatsApp and Instagram, using tools such as Chatizalo.com, allows companies to offer continuous and effective support to their customers.

Do you want to validate this article?

By validating, you are certifying that the published information is correct, helping us fight against misinformation.

Validated by 0 users
Martin H. Pefaur

Martín Pefaur

I lead P4 Tech Solutions, a cutting-edge software factory focused on blockchain and AI. Our mission is to bring founders' ideas to life and promote product adoption. Notable projects include FinGurú, Chatizalo, Ludus Game, Number One Fan, Hunter's Pride, VeriTrust Protocol, Matrix-Tickets, Realtok DAO, Resilientes & Speezard DAO, and others. Actively shaping the future of blockchain and AI.

Twitter Linkedin

Total Views: 0