Enlist ChatGPT as your new startup partner

Photo by Jonathan Kemper on Unsplash

Enlist ChatGPT as your new startup partner

Randall Hancock, February 27, 2023

Starting or growing a new business? If so, consider enlisting ChatGPT as your startup’s next team member. Like other emerging generative AI models, ChatGPT can give you superpowers to unlock creativity and dramatically enhance productivity, helping accelerate your company’s growth and success.

On the unlikely chance that you’ve missed the news, ChatGPT is an artificial intelligence language model developed by OpenAI that has been trained on vast amounts of data, allowing it to generate human-like responses to a wide range of questions and conversations. Essentially, it is an AI-powered chatbot that can understand natural language and engage in helpful conversations.

We’ve been using ChatGPT extensively since it launched in November of 2022, during which time it has earned a full-time role on our team. We’ve discovered that ChatGPT is like having a young, extremely bright, and ambitious new team member – eager to please, very capable, hardworking, and not always right. And just like any new talent, if ChatGPT is provided with the right direction, nurturing, and feedback, it rewards us with some incredible results.

This guide provides step-by-step examples on how to put ChatGPT to work as your new AI partner. Using a fictional startup company as our example, we explore how we can ask ChatGPT to help us develop our elevator pitch, sketch out the business model, create an investor pitch deck, identify early-stage investors, write cold direct messages, negotiate investment terms, find customers, design our leadership team structure, create job descriptions, develop marketing copy, determine operating metrics, and prioritize our execution plan.

Getting started is easy and doesn’t take too much time. Let’s see what ChatGPT can do!

Sign up for ChatGPT

If you haven’t already, sign up for a free research preview account at https://chat.openai.com. You can upgrade to a ChatGPT Plus subscription later if needed.

Learn how to construct effective prompts

In the context of using ChatGPT, a prompt is a query used to start a conversation or request information from the AI. It can be a question, statement, or any other message. After a prompt is entered, ChatGPT analyzes the text and generates a response based on the knowledge and patterns it learned during its training. The response may be a single sentence, a paragraph or even multiple paragraphs depending on the complexity of the prompt and the AI’s ability to generate a suitable answer.

Some of the best practices for constructing prompts to get the best responses include:

  • Be specific: The more specific the prompt, the better chance of getting a relevant and useful response. Provide as much detail as possible about the topic you are interested in.
  • Be clear: Make sure each prompt is clear and use to understand. Avoid using jargon or overly technical terms.
  • Ask a question: Questions or requests tend to elicit better responses than statements. Make sure your prompt includes a clear request that the AI can answer.
  • Keep it concise: Avoid writing long and complex prompts. Keep your prompts short and to the point.
  • Provide context: If a request requires come background or context, provide it in the prompt. This will help the AI understand what you are asking and provide a more accurate response.
  • Avoid bias: Be careful not to include any biases in your prompt. This could affect the AI’s response and provide inaccurate information.
  • Be polite: ChatGPT is programmed to respond politely and professionally, so it’s always a good idea to be polite.

Understand knowledge is dated. The current version of ChatGPT was trained using a vast corpus of text from the internet, books, and other sources up until its knowledge cutoff date of September 2021. Therefore, while its training data is extensive and includes a wide range of topics and domains, the model may not have the most up-to-date information on every topic. Newer versions of ChatGPT and other generative AI tools will likely have much more current information, including real-time integration with internet search.

Do not share confidential information. Your prompts and responses could be integrated inadvertently into future versions of the software; therefore, do not input any information that you wouldn’t want to be publicly accessible.

Beware of AI hallucinations that provide incorrect answers. As you engage, keep in mind that the current version of ChatGPT doesn’t know if its answers are correct or not. Its algorithms determine responses based on probabilities of what the best answers might be, drawing from what the AI knows. It is not uncommon for ChatGPT to hallucinate, or make up answers that appear confident, but which are in fact entirely incorrect. Future versions of ChatGPT will likely address many of these challenges. In the meantime, use ChatGPT as an editor, research assistant, and brainstorming partner; however, make sure you verify data independently before using answers in final decisions or deliverables.

Experiment and learn. Once you see what the AI can and cannot do, experiment with prompt variations to determine which provide the best responses. For example, suggest a style in which you would like the response written, such as a Harvard Business Review article or LinkedIn post. Or provide restrictions, such as limiting the numbers of words or sentences in a response. Since ChatGPT remembers some of the current chat session, you can also ask it to revise a previous paragraph to be longer or shorter, use a different tone, or end on a different note. We have found that co-editing with ChatGPT often gets us the most valuable responses. It’s only through trial and error that you will learn what’s work best for you.

1. Develop an elevator pitch

Now that you’re familiar with ChatGPT, let’s write a short description of our business and what we’re trying to achieve, which we will use as context in our prompts. We’ve created the following fictional startup and situation to use in our examples:

We are a software-as-a-service startup that uses AI and Web3 technologies to improve the productivity of startup founders and leadership teams. Our innovative solution helps you do your job smarter, faster, and much more efficiently. We are seeking seed financing to build the MVP and demonstrate product-market fit.

Feel free to create your own description now if you want to use a real business situation; otherwise follow along with the fictional startup that we’ve created. We can ask ChatGPT to “Rewrite the following text as a compelling elevator pitch that convinces VCs to invest in our startup.” Notice that we have included the business description as context after the query, which provides ChatGPT with the added information it needs to provide a useful response.

After a few moments, ChatGPT begins writing a draft elevator pitch based upon our request and the information that we have provided it. In fact, we see that it has created a short elevator pitch, reorganizing the content order and including new language like “Imagine being able to revolutionize…”and “With your investment, we can bring our MVP to market, validate product-market fit and disrupt the industry.” If you’re experimenting with your own business description, the response you get will differ based upon the prompt and context you provided. Once the draft is completed, try asking ChatGPT to make it shorter or more detailed, adjust the tone, or emphasize different points. While it probably won’t produce our final elevator pitch by itself, it does give us language that we can further refine and finalize.

2. Sketch out your business model

Now that we have a working elevator pitch, let’s spend some time on our business model. I know we wanted to sketch this out on the back of a napkin, but before we do, how about enlisting some help from ChatGPT? We like the Business Model Canvas as a framework, so let’s prompt the AI to “Use the business model canvas framework to describe the above business,” referring to our previous prompt. If ChatGPT provides an error or can’t reference the previous prompt, simply copy and paste the original description into the new request.

Apparently ChatGPT already knows about the Business Model Canvas, as it responds with a decent high-level description of our business using the framework’s building blocks: customer segments, value proposition, channels, customer relationships, revenue streams, resources, operations, partners, and cost structure. Just like the elevator pitch, the results will depend on the details you provided in the business description, as well as the general knowledge ChatGPT already has about the industry. Try asking the AI to provide more detail or suggest ideas about what should be included in a specific building block. We can also experiment with other well-known business frameworks, such as SWOT or Porter Five-Force analyses.

3. Create your investor pitch deck

Armed with the completed Business Model Canvas, it’s time to raise some seed financing for our new venture. We’re going to need a pitch deck for investors, so let’s put ChatGPT to work again. Prompt it to “Develop the outline for an investor pitch deck, including headlines, key points, and proposed graphics for each page…” for our business.

Evidently ChatGPT has also reviewed a lot of pitch decks, because it quickly generates the outline and suggested content for an eight-page deck that includes an introduction, problem statement, solution, market opportunity, business model, team, financials, and conclusion. While the AI isn’t yet able to create pitch deck slides, it has given us a starting point on what to include. Try prompting to make the pitch deck longer, ask for more details on specific pages, or suggest alternative headlines to give more choices from which to choose.

4. Identify potential investors

Who are we going to pitch our new business to? Let’s ask ChatGPT for some ideas, with the prompt “Develop a list of potential early-stage investors for the following startup business, including one-sentence describing why each investor is included on the list.”

ChatGPT responds with six categories of early-stage investors that might be appropriate for our startup, including seed-stage venture capital firms, angel investor networks, incubators and accelerators, AI-focused venture capital firms, Web3-focused venture capital firms, and corporate venture capital arms, providing a few examples along with the rationale for each. If this was too generic and not as helpful as we hoped, try asking ChatGPT to generate a list of 10 investors in one of the promising categories, along with a summary description of each investor. Top 10 investors not enough? Ask it to generate 10 more. Remember that the AI is eager to please and prone to hallucinate, so we’re going to want to check these ourselves. Nonetheless, we now have a starting list of investors to contact.

5. Reach out to investors via cold direct messages

But how are we going to contact these investors? Of course, we already know the best way is through warm introductions provided by close professional contacts; however, we may run out of those before securing funding. Let’s therefore ask ChatGPT to “Write a short, professional, and compelling cold LinkedIn direct message to a prospective early-stage investor” for our startup.

Voila! Within a minute ChatGPT has suggested language we might use to reach out to investors cold using a LinkedIn direct message. Don’t like the tone? Ask the AI to make it more professional, friendlier, or upbeat. Need more detail? Prompt it to integrate additional context we provide. Want to send via Twitter instead? Instruct ChatGPT to shorten the previous message to an appropriate Twitter length and style.

6. Negotiate investment terms

We’ve had some successful meetings and are now looking to negotiate investment terms. Since we really think our startup is valuable, we want to be as prepared as possible. Let’s ask ChatGPT to “Develop an argument targeted to venture investors justifying a $20 million pre-seed valuation” for our business.

There is no guarantee that this logic is going to work. Nonetheless, ChatGPT does its best to produce compelling rationale that we can use in our investor meetings. Good luck!

7. Identify early-adopter customers

Let’s assume that we’ve now raised some money and have returned all our focus to developing our new business. We need to demonstrate product-market fit ASAP, and what better way to do that than attract some early-adopter customers for our new solution? Since we have a trip planned to Boston, we ask ChatGPT to “Identify a list of early-stage companies in Cambridge, Massachusetts who might become early-adopter customers for our business, providing a short description of each.”

Not surprisingly by now, ChatGPT comes back with a list of 10 companies that we might call on for our initial product. However, this list seems generic, so perhaps we’ll ask it to suggest 10 more potential customers. Or alternatively, maybe provide more guidance on the types and sizes of companies that we seek to target.

8. Design leadership team structure

The founders can’t do all the work to get this new venture up and running, so we turn our attention to expanding our leadership team. Ask ChatGPT to “Suggest the optimal leadership team structure for the next twelve months” for our business.

ChatGPT responds with a suggested leadership team structure that includes the CEO as well as chief technology, marketing, operations, and sales officers, along with a high-level description of each role. In this case the AI has also noted that the benefits of this organizational structure. But what if it missed something? If there are additional leadership positions we would like to include, then provide that as guidance in a follow-up prompt. We can also ask for more details, such as the potential direct reports for each position. For example, asking “What roles typically report into the CMO?” identifies and describes marketing, brand, product, public relations, and customer success managers.

9. Create job descriptions

Next, we need start recruiting to fill our leadership team gaps. Writing job descriptions is necessary but never fun, so we ask ChatGPT to “Create the job description for a Chief Marketing Officer” of our business.

The AI responds with a comprehensive CMO job description that provides an overview of the role, detailed job responsibilities, suggested education and experience requirements, an equal opportunity employer statement, and call to action to apply. Not bad. We could probably post this position right away but would be better served by co-editing it first. For example, we could prompt to add or delete responsibilities, change the experience requirements, or even refine the writing style tone, making it less or more formal. Then we could copy and paste it into Word to make final edits before sending it out.

10. Develop marketing copy

While we’re recruiting our new CMO, we still need to make progress marketing our MVP solution. Let’s ask ChatGPT to “Develop the marketing copy for a one-page brochure that introduces the company to potential early-adopter customers.”

ChatGPT generates a decent first draft of a marketing brochure, based upon the context that we previously provided it and its general knowledge of our industry. We could get better results if we provide it with more details on our solution’s target customer segments, value proposition, and functionalities. We could also request a different call to action, or even ask for potential graphics to accompany the marketing text. Use your imagination and see what you can co-create.

11. Determine operating metrics

Our MVP is about to launch, and thus we want to think about how we are going to track our performance. We prompt ChatGPT to “Suggest the most helpful operating metrics of interest to investors and the board of directors” for our business.

ChatGPT returns with seven operating metrics that are frequently used by SaaS growth companies: monthly recurring revenue (MRR), gross margins, customer acquisition cost (CAC), customer lifetime value (LTV), monthly active users (MAUs), user feedback and net promoter score (NPS), and product-market fit metrics. These metrics are all helpful and appropriate; however, we can push the AI to deliver more. Try asking how to calculate LTV for our business or ask what typical gross margins are for SaaS growth firms. Remember that ChatGPT may make things up (hallucinate) if it doesn’t know the answer, so we will use these as directional inputs to enhance and speed up our thinking process.

12. Prioritize execution plan

Let’s get help on one more thing, which is to determine what our near-term operating priorities should be. Prompt ChatGPT to “Propose the execution plan priorities over the next twelve months for the leadership team” of our business.

ChatGPT comes back with seven suggested execution priorities: fundraising, building the MVP, customer discovery, marketing and branding, partnering and network building, team building, and scalability planning. Hey, we’ve already started on many of those today, but this reminds us that there are additional priorities on which we need to focus. Try asking the AI to provide greater detail and examples on one of the priorities or to suggest additional priorities. You’re the CEO, after all, so you make the final call; however, ChatGPT is useful for suggesting things that we might not have thought about on our own.

As we’ve seen, generative AI models like ChatGPT are already proving themselves to be capable partners that we can put to work today. While there are limitations in their capabilities, expect the next generation of tools provided by OpenAI, Microsoft, Google, Meta, and others to become even more effective at generating valuable responses.

We’ve explored just a few possibilities in this guide. In addition to the examples we’ve covered, we’ve used ChatGPT to conduct strategic research, estimate total addressable market (TAM) opportunities, perform competitive analyses, draft corporate video scripts, create website code, edit business correspondence, create social media posts, summarize complex documents, write privacy policies, and many more. With these new tools, use classes are only limited by one’s imagination, with possibilities unfolding across every business function. As you interact with ChatGPT, you will discover the prompts that work best for you and your organization. How are you going to put ChatGPT to work as your new AI partner?

Randy is an entrepreneur, executive, thought leader and growth company advisor focused on the intersection of business, strategy and disruptive technologies, such as Web3 and AI.