ChatGPT is the most viral application–literally of all time. It took just 5 days for ChatGPT to gain 1 million users and 2 months to gain its first 100 million users.
So, first, what is ChatGPT?
ChatGPT is a pre-trained language generation model developed by OpenAI. It uses deep learning techniques to generate human-like responses to questions, based on the input it receives. It can be useful to a mid-size or large organization by automating customer service tasks such as answering frequently asked questions, directing users to relevant information, and providing assistance with basic tasks. This can lead to increased efficiency and improved customer experience, as well as reduced costs and workload for human support staff.
To illustrate the impressive and human-like conversational ability of ChatGPT, the above paragraph was actually generated by ChatGPT.
ChatGPT also came up with the title for this article. Here is the full list of ideas it came up with:
- "Revolutionizing Management Consulting with ChatGPT Technology"
- "ChatGPT in Management Consulting: An Innovative Approach"
- "The Future of Management Consulting with ChatGPT"
- "ChatGPT: A Game Changer for Management Consultants"
- "Maximizing Efficiency in Management Consulting with ChatGPT"
- "How ChatGPT is Transforming the Management Consulting Industry"
- "The Benefits of Integrating ChatGPT into Management Consulting Practices"
- "ChatGPT-Powered Management Consulting: Streamlining the Process"
- "Enhancing Management Consulting with the Power of ChatGPT"
- "Management Consulting in the ChatGPT Era: A Look into the Future".
In summary, ChatGPT is a free long-form question-answering application that answers complex questions conversationally. To better visualize this, let’s look at what the ChatGPT interface looks like:
ChatGPT’s breadth and depth of knowledge is almost incomprehensibly vast. In fact, ChatGPT has already been used to pass the following exams, which require deep technical expertise:
- The U.S. Medical Licensing Exam
- A final MBA exam at University of Pennsylvania’s Wharton School of Business
- Law exams in four courses at the University of Minnesota
Former Treasury Secretary Larry Summers told Bloomberg that ChatGPT “could be the most important general-purpose technology, since the wheel or fire.”
ChatGPT Management Consulting Example
The use cases for ChatGPT are also extremely varied and versatile.
As the article suggests, let’s test out ChatGPT for purposes of management consulting. The below convo is verbatim from a ChatGPT discussion I had.
Prompt (Me): Write as if you were a management consultant with 10 years of experience working at a global consulting firm. The situation is: your client is the CEO of a global fast fashion company. He is looking for ways to reduce costs across the supply chain while minimizing the company’s carbon footprint. He would like to brainstorm some ideas. What would you recommend to him?
As a management consultant with a decade of experience at a global consulting firm, I would recommend the following strategies to reduce costs and minimize carbon footprint across your supply chain:
- Optimize Logistics: Optimize logistics and reduce transportation costs by implementing a more efficient supply chain network and reducing waste through better inventory management.
- Sustainable Sourcing: Source materials from sustainable suppliers who use eco-friendly production methods and have certifications for sustainability, such as the Global Organic Textile Standard (GOTS).
- Energy Efficiency: Implement energy-efficient technologies in your production processes, such as using renewable energy sources or utilizing energy-efficient equipment.
- Upcycle & Recycle: Utilize upcycling and recycling processes in your production methods to reduce waste and minimize the use of new raw materials.
- Collaborate with Suppliers: Collaborate with suppliers to jointly reduce carbon footprint and drive sustainability initiatives, such as reducing water usage or implementing waste reduction programs.
- Lean Manufacturing: Adopt Lean Manufacturing techniques to streamline production processes and reduce waste, resulting in cost savings and improved sustainability.
- Digital Transformation: Invest in digital technologies such as automation, predictive analytics and blockchain to increase supply chain visibility, improve transparency, and reduce costs.
These recommendations can be tailored to fit your specific needs and are not exhaustive, but they provide a good starting point for reducing costs and improving sustainability in your supply chain.
Prompt: These are great recommendations. I am interested in learning about the last point, Digital Transformation. Can you elaborate on some specific ideas around what processes we can automate?
Certainly! When it comes to digital transformation in the supply chain, there are several areas that can benefit from automation:
- Inventory Management: Automated inventory management systems can provide real-time visibility into stock levels, reducing overstocking and stock-outs, and helping to optimize inventory levels.
- Order Fulfillment: Automated order fulfillment processes can increase accuracy and speed, reducing the time and cost associated with manual order processing.
- Transportation Management: Automated transportation management systems can optimize routing and reduce costs by choosing the most efficient carriers and modes of transportation
- Supply Chain Visibility: Automated supply chain visibility solutions can provide real-time data on all stages of the supply chain, helping to identify inefficiencies and bottlenecks, and enabling faster decision-making.
- Quality Control: Automated quality control systems can reduce the risk of errors and improve product quality, red
- ucing the costs associated with returns and rework.Sustainability Tracking: Automated sustainability tracking solutions can provide real-time data on carbon footprint and other sustainability metrics, helping companies to monitor their progress and identify areas for improvement.
By automating these areas, companies can improve efficiency, reduce costs, and minimize the carbon footprint of their supply chain operations. Additionally, automation can help companies to more quickly respond to changes in demand and market conditions, allowing them to be more agile and competitive in a fast-paced, global market.
Prompt: This is great. Which of these areas is most relevant to the fast fashion industry?
In the fast fashion industry, inventory management and order fulfillment are the most relevant areas for automation.
Fast fashion companies often have a high turnover of stock and short product life cycles, making it important to have efficient and accurate inventory management systems in place. Automated inventory management can provide real-time visibility into stock levels and help to optimize inventory levels, reducing the risk of overstocking and stock-outs.
Order fulfillment is also a key area for automation in fast fashion, as it can help to increase the speed and accuracy of order processing. This is particularly important in an industry where consumers expect quick and efficient delivery. Automated order fulfillment processes can help fast fashion companies to meet these expectations, reduce the time and cost associated with manual order processing, and improve the overall customer experience.
These two areas, along with supply chain visibility, are key for fast fashion companies to stay competitive, reduce costs, and minimize their carbon footprint.
As you may have already gathered, crafting the “prompt” (i.e. the input phrasing) is extremely important. OpenAI even acknowledges this, stating ChatGPT is sensitive to tweaks to the prompt or attempting the same prompt multiple times.
Thus, there is a now a new term (and potential field) around the writing of prompts–known as Prompt Engineering. Prompt Engineering is the art and science of writing the prompt to elicit the desirable or useful results from the Artificial Intelligence (AI) tool you are using.
Like in most processes and systems, the quality of the inputs determines the quality of the outputs—i.e., garbage in, garbage out. With ChatGPT, the input is the prompt. Thus, it is imperative that the user write a well-crafted and articulated prompt to generate the highest quality, most desired output. Designing effective prompts increases the likelihood that the model will generate the most favorable and contextual response.
We can draw the analogy of asking a genie for a wish. It is important to phrase your wish precisely, so that the genie grants you the wish you actually hoped for (if not more).
Here are several guiding principles for Prompt Engineering:
- Be specific, descriptive and as detailed as possible about the desired context, outcome, length, format, style, etc. Be specific about the context, outcome, length, format, style, etc.
- Show and tell by giving examples in the prompt. Articulate the desired output format through examples. ChatGPT will respond better when shown specific format requirements.
- Minimize or remove imprecise and “fluffy” descriptive language.
- Rather than stating what not to do, specify what to do instead.
- Include acceptable responses in prompts for consistency.
- Align prompt instructions with the end goal. We need to put ourselves in the shoes of ChatGPT as if it were a person and ask: what will ChatGPT think the end goal of my task is?
- Use personas to get more specific voices. Note: this is the tactic we leveraged in our initial prompt in the Consulting example above.
- Experiment. Try and test out different prompts and prompt writing tactics to find what works best for your specific needs.
Limitations & Criticisms
Although a breakthrough technology, ChatGPT is not without its limitations and criticisms. The biggest one of ChatGPT is that the information it gives may not be accurate. ChatGPT sometimes writes plausible-sounding, but incorrect or nonsensical answers. To the unwitting user, he or she may take the response as true, when it is completely fictitious.
OpenAI has acknowledged this issue and said it is difficult to fix for a few primary reasons:
- During training, there’s no source of truth.
- Training the model to be more cautious causes it to decline questions that it can answer correctly.
- Supervised training misleads the model, because the ideal answer depends on what the model knows, rather than what the human demonstrator knows.
ChatGPT has already been banned by a number of websites, such as Stack Overflow, the Q&A platform for programmers, due to the inaccuracy of responses.
Are you interested in learning more about ChatGPT? Take a look at the ChatGPT presentations available on Flevy.
We also have a FlevyPro framework on ChatGPT, which this article is largely based on. The FlevyPro ChatGPT presentation additionally discusses GPT/GPT-3, GPT-3 use cases, additional ChatGPT limitations, ChatGPT controversies, and goes over more ChatGPT examples.
It looks like the IA prompter would be a very useful skill to learn these days.
Great question, Boris! I decided to pose this question to ChatGPT and here's what it responded with --
An online community of risk management professionals can use ChatGPT in a variety of ways to enhance their discussions and exchange of ideas. Here are some of the ways that ChatGPT can be useful:
Q&A: Members can ask ChatGPT questions related to risk management, and ChatGPT can provide relevant and up-to-date information and insights.
Knowledge Base: ChatGPT can act as a repository of information and best practices for risk management, allowing members to easily search for and retrieve information on specific topics.
Discussion Prompts: ChatGPT can generate thought-provoking questions and prompts to encourage meaningful discussions and exchange of ideas among community members.
Data and Insights: ChatGPT can analyze data and provide insights and recommendations on various risk management strategies and approaches.
Personalized Assistance: ChatGPT can provide personalized advice and recommendations to individual community members based on their unique needs and circumstances.
These are just a few examples of how ChatGPT can be leveraged to support the work of an online community of risk management professionals.
thanks for this informative article. How would you suggest to use ChatGPT for Global Risk community?