How to Use ChatGPT to Support Your Business

In a recent LinkedIn discussion led by David Rosendahl, the spotlight was on the transformative potential of AI technologies like ChatGPT in business. The conversation covered various applications—from optimizing recruitment processes to aiding in software development, and even enhancing writing tasks. This blog addresses the foundational questions: What are the pros and cons of ChatGPT, and how can it be used to streamline customer service as you automate your workflow?

This article is the first installment of a series focusing on how to take advantage of ChatGPT in your business. Whether it's enhancing customer service, expediting email responses, generating social media content, launching new products and services, crafting blog posts, conducting market research, or preparing managerial reports and presentations, ChatGPT has a role to play. 

Developed by OpenAI, ChatGPT is engineered to respond to user queries, often referred to as "prompts." These prompts access a database last updated in September 2021, so the content is about two years old. Both free and subscription-based versions are available. Our recommendation is to start with the free service and upgrade to the subscription service ($20/month) if you want more reliable access.

However, before you dive into the world of artificial intelligence, it's crucial to understand the ethical considerations surrounding its use, particularly the ambiguous area of originality and plagiarism. 

Is Using AI Considered Plagiarism?

It depends. 

Employing ChatGPT to produce academic papers, business reports, or other professional writings and claiming them as your own is deemed plagiarism. Such actions not only place you at risk of academic or professional sanctions, but also erode the foundational trust and integrity of these disciplines. Legal consequences and irreversible damage to your reputation are additional risks. On the flip side, ChatGPT can serve as a valuable tool for drafting, revising, and improving upon your original work without crossing into plagiarism territory.

Prompts serve as the primary mode of interaction with the platform and mastering them is as much an art as it is a science. Various resources, including classes and books, offer insights into effective prompt formulation. Since prompts and responses from ChatGPT are saved and available, if you write your own first draft there is little chance it would be considered plagiarism. The key lies in a balanced collaboration with ChatGPT. An emerging best practice is to not ask ChatGPT to write something from scratch, but to work collaboratively. Yes, I use ChatGPT, but I always write a first draft myself before I turn to the technology for further editing and refinement.

Optimizing Your Interaction with ChatGPT

Optimizing ChatGPT's performance requires a bit of trial and error. Think of it as an interactive experience. We already discussed starting with a rough draft. If that doesn’t work, try more general prompts and questions, and then tweak your initial prompts based on these responses for more accurate and helpful information. Combine original thoughts and writing and avoid copying answers to mitigate the risk of complaints. While learning through trial and error is effective, there are several YouTube tutorials to quicken the learning curve.

The Question of Accuracy

If plagiarism is the number one concern, then accuracy is number two. ChatGPT, like other AI technologies, isn't infallible. OpenAI acknowledges that the model may churn out "inaccurate or nonsensical answers." The issue of whether its accuracy is improving or deteriorating is subject to debate. OpenAI has reported an improvement in accuracy from version 3.5 to version 4. However, some studies challenge this claim, suggesting a decline in performance quality. 

One such study, featured in Popular Science, documents this conflict. Researchers from Stanford and UC Berkeley observed noticeable changes in ChatGPT's performance over time—changes that did not enhance its efficacy. Surprisingly, the reasons behind this decline remain somewhat elusive to researchers.

While these study results seem irrefutable, it’s important to recognize that accuracy is dependent on the clarity of the prompt. Vague questions will likely yield vague or irrelevant answers. For instance, a general question like "How can I optimize my operations?" will offer less targeted guidance than a more specific query, such as "How can I streamline workflow in an in-plant printing setup?" To elicit even more precise advice, consider instructing ChatGPT to adopt an expert role, as in "Act as an in-plant printer workflow consultant and explain how to streamline my workflow." As your prompting skills improve, so will the response accuracy.

Leveraging ChatGPT to Support Your Web to Print Automation

The automation trend for in-plant printers has seen considerable growth over the past two decades, especially in streamlining front-end processes like estimates, order entry, and customer interactions. While this improves efficiency, it can also risk isolating customers, making them feel they can’t get the answers they need. With an increasing number of companies embracing automated models, customer service challenges are inevitable. 

ChatGPT presents a solution to these challenges by serving as an effective tool for automating customer interactions, especially when traditional methods like email or social media prove burdensome. To capitalize on this, consider using ChatGPT to create an enriched customer service ecosystem that includes detailed responses to frequently asked questions on a variety of topics—ranging from how to get quotes to shipping options. 

Templates for Common Customer Queries

To begin this process, start by identifying common questions that your customer service team regularly fields. These could be about order status, estimated delivery times, or price adjustments. Categorize these questions and develop standardized response templates for each. For example, a template addressing an order status query could include details such as the customer's name, order ID, current status, and expected delivery date. 

For sorting incoming emails, implement a keyword-matching strategy to classify them into relevant categories. If you encounter inquiries that are too complex, create a manual escalation process. This ensures that sensitive or complicated issues are handled by human agents, preserving the all-important element of customer trust. 

Summary

In this blog we introduce ChatGPT as a strategy to automate business processes such as customer service, email automation, social media, and more. We start with the ethical implications, specifically around plagiarism, when using the tool for tasks like generating academic or professional content and recommend a blended approach in which the content produced starts as a first draft in the user’s own words, before being refined by ChatGPT.

Mastering the art of "prompting" is critical for effective utilization of the AI model and the blog stresses the importance of trial and error in obtaining accurate and useful responses. We discuss that ChatGPT is not infallible, but its accuracy can be improved by crafting precise, well-formulated prompts.

Finally, the piece offers practical advice for leveraging ChatGPT in Web to Print automation and customer service. It suggests creating detailed FAQs and email templates to enhance front-end automation processes. The blog also recommends implementing a manual escalation process for handling complex or sensitive queries, ensuring customer trust is maintained.