How to Use AI Tools like ChatGPT for Financial Projections

April 14, 2023

Adam Hoeksema

With the pace of innovation in AI and tools like ChatGPT I wouldn’t be surprised if within the next few years a small business owner can simply ask ChatGPT to create a set of financial projections for their proposed business based on things like location, business model type, and industry financial data.  Who knows, maybe you will just be able to submit your coffee shop business plan to an AI and it will create your projections for you in an instant.  

Although this isn’t quite possible yet today, AI can still be incredibly useful in the financial projection process.  I wanted to share how I use ChatGPT to help create realistic financial projections.  I am going to walk through the following steps and show you how I use ChatGPT throughout the process. 

  1. Start with a Financial Projection Template
  2. Estimate Startup Expenses
  3. Forecast Revenue based on your Business Plan
  4. Project Expenses as a Percentage of Revenue
  5. Estimate Staffing Costs
  6. Review Whether Projections are Realistic

With that as a plan, let’s dive in. 

Start with a Financial Projection Template

Because AI tools can’t build a financial model from scratch just yet, the best place to start is going to be a financial projection template.  For this example, I am going to assume we are working on projections for a restaurant.  So I started by asking ChatGPT to send me a link to a financial projection template for a restaurant and although it didn’t send me a direct link, I was happy to see that it listed ProjectionHub’s restaurant financial projection templates as its second choice below:

With a template in hand, we can use ChatGPT to help with research to fill out the financial model with appropriate assumptions. 

Estimate Startup Expenses

To estimate startup expenses for a new business I like to start by asking ChatGPT to list common startup expenses for my type of business.  This will help me ensure that I am not forgetting about any common startup expenses.  So here is what I did:

Prompt: What are the common expense categories for a startup coffee shop? 

AI Response: Some of the common expense categories for a startup coffee shop are:

  • Lease and utilities
  • Construction and remodeling
  • Furniture and fixtures
  • Equipment
  • Initial inventory
  • Licenses and permits
  • Insurance
  • Marketing and advertising
  • Staff wages and benefits
  • Professional services
  • Operating expenses
  • Contingency fund

ChatGPT also did an excellent job of summarizing what could be included in each expense category which I have included below:

Lease and utilities: Rent for the location, utility bills (electricity, water, gas, internet, phone), and any additional property-related expenses (maintenance, property taxes, etc.) 

Construction and remodeling: Costs for building out or remodeling the space to fit the coffee shop's design and operational needs, including flooring, lighting, plumbing, electrical work, and painting.

Furniture and fixtures: Expenses for purchasing or leasing tables, chairs, display cases, countertops, shelving, and other necessary furniture.

Equipment: Costs for acquiring coffee brewing equipment (espresso machines, grinders, brewers, etc.), refrigeration units, kitchen appliances, and other required tools for the coffee shop.

Initial inventory: Costs for purchasing the initial stock of coffee beans, tea, milk, syrups, pastries, snacks, cups, lids, straws, and other consumables.

Licenses and permits: Fees associated with obtaining the necessary business licenses, permits, and certifications, such as health permits, food service permits, and liquor licenses (if applicable).

Insurance: Expenses for different types of insurance coverage, such as general liability, property, workers' compensation, and business interruption insurance.

Marketing and advertising: Costs for promoting the coffee shop, including signage, logo design, website development, social media, print materials, and local advertising.

Staff wages and benefits: Expenses related to hiring, training, and compensating employees, including salaries, benefits, and payroll taxes.

Professional services: Fees for legal, accounting, and consulting services to help with business setup, tax planning, and ongoing management.

Operating expenses: Ongoing costs for supplies, maintenance, repairs, and other miscellaneous expenses necessary for day-to-day operations.

Contingency fund: A reserve of funds to cover unexpected expenses or cash flow shortages.

Next, I wanted to see if ChatGPT could give me a range of the startup cost I might expect for each category.  In order to give it a bit more guidance I provided a city that the coffee shop would be located in, and I gave it an expected number of square feet and seats that the coffee shop would have.  

Prompt: Can you estimate a range of the startup cost for each category for a coffee shop starting in Chicago with 30 seats and 1,500 square feet?

AI Response:  ChatGPT estimated that the total startup cost for a coffee shop in Chicago would be between $90,000 and $270,000.  The breakdown of the startup costs from ChatGPT are as follows:

Lease: $2,000 - $5,000 per month

Utilities: $500 - $1,500 per month

Construction and remodeling: $20,000 - $80,000

Furniture and fixtures: $7,000 - $20,000

Equipment: $25,000 - $75,000

Initial inventory: $3,000 - $7,000

Licenses and permits: $1,000 - $3,000

Insurance: $1,500 - $3,500 per year

Marketing and advertising: $2,000 - $10,000

Staff wages and benefits (assuming 4 full-time and 4 part-time employees):

  • Full-time baristas (4): $25,000 - $35,000 per year each
  • Part-time baristas (4): $12,000 - $20,000 per year each

Professional services: $1,000 - $5,000 per year

Operating expenses (monthly): $2,000 - $6,000

Contingency fund: 5% - 10% of the total startup costs

I thought this was a pretty solid response.  I think the primary question and variation in your specific plan will be related to staffing.  Are you going to be working in the business yourself as the owner?  If so, this may reduce the amount of staffing required.  Also, in my past as an SBA lender, we typically did not lend for a full year of staffing costs when we made SBA Microloans for startups.  We would assume that the business should start to cover staffing costs within a few months, so at most we would want to lend 3 months worth of staffing costs. 

Unfortunately, AI isn’t quite ready to just ask it to take these expenses and fill out a spreadsheet template for you, but I am sure this will be possible someday soon.  For now, you will want to take these expenses and enter in the assets and operating expenses as seen below. 

Enter in assets like inventory, equipment, furniture and leasehold improvements:


Enter in startup operating expenses in the first column or first month of the operating expenses section:

Forecast Revenue based on your Business Plan

Once you have your startup expenses forecasted, the next step is to project your revenue based on your plans.  Again, I like to give ChatGPT as much information as possible to help me forecast revenue.  For this section, let’s assume we are planning to launch a restaurant.  You can see the types of details and assumptions I provide in my prompt to help it forecast. 

Prompt:  How much revenue do you think my steakhouse in Austin, Texas could generate assuming 150 seats, an average order value of $60 per person, and assuming the restaurant is open for dinner 7 days a week? 

AI Response: ChatGPT estimated that my steakhouse could generate roughly 2.6 million in annual revenue.  It assumed that 80% of our 150 seats would be filled each night, but it also assumed that each seat would only be filled once per night.  

I know that on busy nights, I might be able to turn tables more than once per night, in fact I might even be able to use the same table for dinner at 5 pm, 7 pm and 9 pm.  At the same time, I know that weeknights might not be nearly as busy as the weekend.  So I wanted to see if ChatGPT could do some research for me on the average number of table turns a fine dining restaurant would do on weekends and weeknights. 

Prompt:  How many times does the average fine dining restaurant turn tables per night? 

AI Response:  The average fine dining restaurant would expect to turn the same table 1 to 1.5 times per night and maybe up to 2 times per night during peak days and busy seasons.  Given that fine dining restaurants typically have a more leisurely atmosphere, the restaurant is not as likely to push guests for faster table turn times.  

Finally, I asked whether an 80% occupancy rate was average for a fine dining restaurant and ChatGPT said 80% would be in the high range and that 60% to 70% would be more in line with an average.  

With this guidance you ought to be able to fill out the key assumptions on the Input Revenue tab.  

Forecast Expenses as a Percentage of Revenue

Next we need to estimate operating expenses for our startup.  Similar to the process with startup expenses, I like to ask ChatGPT for a list of common operating expenses for my particular business type.  For this example, let’s use a different type of business model like ecommerce to see how well it can do.  

Prompt:  What are the typical operating expenses for a direct to consumer product business sold solely through ecommerce channels?  

AI Response: Typical expenses for an ecommerce, direct to consumer product include:

Cost of goods sold (COGS): This includes the cost of raw materials, manufacturing, packaging, and shipping of the products to your warehouse or fulfillment center.

Web hosting and maintenance: This covers the cost of hosting your e-commerce website and ensuring it remains functional and up-to-date.

E-commerce platform fees: This includes fees charged by e-commerce platforms like Shopify, WooCommerce, or BigCommerce, which provide the infrastructure for your online store.

Payment processing fees: These fees are charged by payment processors like Stripe, PayPal, or Square for facilitating online transactions.

Digital marketing: This includes expenses for search engine optimization (SEO), pay-per-click (PPC) advertising, social media marketing, email marketing, and content creation.

Third-party logistics (3PL) or fulfillment fees: These expenses involve the storage, packing, and shipping of your products to customers by a third-party logistics provider.

Inventory management: This includes the cost of warehousing, inventory tracking software, and associated staff salaries. 

Customer support: This covers the cost of providing customer support services, such as staffing a help desk, responding to inquiries, and managing returns and exchanges. 

Salaries and wages: This includes the cost of hiring and paying employees or contractors involved in the operation of your business.

Office expenses: This includes rent, utilities, office supplies, and other related expenses if you maintain a physical office location.

Insurance and licenses: This covers the cost of obtaining and maintaining necessary business insurance and licenses.

Taxes: This includes sales taxes, income taxes, and other tax obligations. 

Software and tools: This covers the cost of software and tools used to manage and optimize your business, such as project management software, analytics tools, and design software. 

Professional services: This includes fees for legal, accounting, and consulting services. 

Research and development (R&D): This covers the cost of developing new products or improving existing ones.

For operating expenses I like to ask ChatGPT to provide these expenses as a percentage of revenue so that I know my expenses should scale with my revenue growth appropriately.  

Prompt: Can you estimate these operating expenses as a percentage of revenue? 

AI Response:  

The average operating expenses as a percentage of revenue for an ecommerce, direct to consumer product company are as follows:

  • Cost of goods sold (COGS): 40-60% of revenue
  • Web hosting and maintenance: 1-3% of revenue
  • E-commerce platform fees: 2-5% of revenue
  • Payment processing fees: 2-4% of revenue
  • Digital marketing: 5-15% of revenue
  • Third-party logistics (3PL) or fulfillment fees: 5-15% of revenue
  • Inventory management: 2-5% of revenue
  • Customer support: 2-5% of revenue
  • Salaries and wages: 10-30% of revenue
  • Office expenses: 1-5% of revenue
  • Insurance and licenses: 1-3% of revenue
  • Taxes: Varies greatly depending on the tax structure in your location and the structure of your business
  • Software and tools: 1-5% of revenue
  • Professional services: 1-5% of revenue
  • Research and development (R&D): 2-10% of revenue

Our ecommerce financial projection template will allow you to enter in operating expenses as a fixed monthly amount or as a percentage of revenue.  If you know that a particular expense will be a specific dollar amount such as rent, then I would just enter in the specific expense, but if you aren’t certain, selecting a percentage within the range provided by ChatGPT can be a great way to estimate your operating expenses. 

Estimate Staffing Costs

Depending on your business you might want to simply forecast your staff salaries and wages as a percentage of revenue as outlined above, but if you would like to forecast specific positions with specific salaries, our templates allow you to do that as well. I like to ask ChatGPT to outline the different positions I might need for my particular business to make sure I am not forgetting anything and then estimate those salaries. 

Typical Staff Positions for a Fine Dining Restaurant

Prompt:  What staff positions does the typical fine dining restaurant have on staff? 

AI Response: The typical fine dining restaurant may include the following staff positions:

General Manager/Restaurant Manager: Oversees the entire restaurant's operations, including staff management, financial performance, and guest satisfaction.

Executive Chef: Responsible for creating the menu, managing the kitchen staff, and ensuring that the food quality and presentation meet the restaurant's high standards.

Sous Chef: Assists the executive chef in managing the kitchen, including overseeing food preparation, staff training, and inventory control.

Line Cooks/Chefs de Partie: Responsible for preparing specific dishes or parts of dishes under the guidance of the executive chef or sous chef.

Pastry Chef: Specializes in creating desserts, pastries, and other sweet treats.

Sommelier: A wine expert responsible for curating the wine list, making wine recommendations, and overseeing the wine service.

Maitre d'/Host: Manages the front-of-house staff, greets and seats guests, and handles reservations and guest requests.

Servers/Waitstaff: Responsible for taking orders, serving food and drinks, and attending to guests' needs throughout the meal.

Busser: Clears and resets tables, and assists servers with carrying dishes and refilling water and beverages.

Bartender: Mixes and serves alcoholic and non-alcoholic beverages, and often has extensive knowledge of cocktail recipes, wine, and beer.

Barback: Assists the bartender with tasks such as stocking the bar, cleaning glassware, and handling other behind-the-scenes duties.

Dishwasher: Cleans and sanitizes dishes, glassware, utensils, and kitchen equipment.

Food Runner: Delivers prepared dishes from the kitchen to the guests' tables, ensuring that the correct dishes are served to the appropriate guests.

Expediter: Coordinates the timing of food orders between the kitchen and front-of-house staff to ensure that dishes are served at the correct temperature and in a timely manner.

You can see how ChatGPT does a great job of providing a detailed list of staff positions that you can then decide whether or not to include in your specific restaurant business plan.  

Next, I asked ChatGPT to estimate a salary range for each of these positions.  To get more precise I typically just Google “average [position name] annual salary in [city name]” and typically you will get great, local data from sites like Ziprecruiter.  Nevertheless, the following ranges from ChatGPT can be helpful to make sure you are in the right ballpark. 

Average Salary Ranges for a Fine Dining Restaurant

  • General Manager/Restaurant Manager: $50,000 - $85,000 per year
  • Executive Chef: $60,000 - $100,000 per year
  • Sous Chef: $40,000 - $60,000 per year
  • Line Cooks/Chefs de Partie: $25,000 - $40,000 per year
  • Pastry Chef: $30,000 - $60,000 per year
  • Sommelier: $40,000 - $80,000 per year (including tips and bonuses)
  • Maitre d'/Host: $30,000 - $50,000 per year (including tips)
  • Servers/Waitstaff: $20,000 - $50,000 per year (including tips)
  • Busser: $18,000 - $30,000 per year (including tips)
  • Bartender: $20,000 - $60,000 per year (including tips)
  • Barback: $18,000 - $30,000 per year (including tips)
  • Dishwasher: $18,000 - $25,000 per year
  • Food Runner: $20,000 - $30,000 per year (including tips)
  • Expediter: $25,000 - $40,000 per year

You can enter in your salaried positions in the salaries tab in our templates as seen below:

Review Whether Projections are Realistic

Once you have entered in your startup costs, revenue forecast and projected expenses, it is time to review your finished projections and make sure they seem reasonable.  ChatGPT is really helpful in providing feedback on a set of projections.  Let me show you how I ask for feedback.  

First, all of our ProjectionHub templates include a Profit and Loss at a glance table as seen below: 

You can simply copy this data and paste it into ChatGPT along with a bit of context and it will provide some good feedback.  See below:

Prompt:  Assume the following data are financial projections for my startup coffee shop in Chicago.  Do these projections look realistic for a coffee shop in Chicago based on industry norms? 

AI Response: 

Based on the financial projections provided and my knowledge up to September 2021, the projections seem to deviate from industry norms in a few areas. However, please note that my analysis is not exhaustive and may not fully capture recent changes in the industry or specific factors that could affect your startup. Here are some observations:

Revenue Growth: The revenue growth rate is quite high in the initial years, particularly from 2023 to 2024 (78% growth) and then tapers off to more moderate growth rates. While it is common for startups to experience high growth rates initially, the coffee shop industry is generally more mature and competitive, making such rapid growth less common. It may be possible with an aggressive marketing plan or innovative product offerings, but it's worth considering if this level of growth is realistic for your specific situation.

Cost of Goods Sold (COGS): The COGS as a percentage of revenue decreases significantly from 70% in 2023 to 41% in 2027. While it is typical for COGS to decrease as a business scales and achieves economies of scale, this is a substantial reduction that may be difficult to achieve in the coffee shop industry, where margins are relatively stable.

Operating Expenses: The operating expenses as a percentage of revenue drop from 61% in 2023 to 30% in 2027. This is a significant decrease and may not be entirely realistic, especially considering the fixed costs associated with operating a coffee shop, such as rent, utilities, and labor. Operating expenses in the coffee shop industry typically range between 55% and 85% of revenue. 

Net Income: The net income margin increases from -31% in 2023 to 23% in 2027. This is quite an improvement and may be difficult to achieve in the coffee shop industry, where net income margins typically range between 2% and 15%, depending on the efficiency of the operation and the market. 

EBITDA: The EBITDA margin increases from -20% in 2023 to 29% in 2027, which is quite high for the coffee shop industry. EBITDA margins in the industry typically range between 10% and 20%. 

In summary, while some aspects of the projections seem plausible, others deviate significantly from industry norms. It's important to revisit your assumptions and consider any unique factors or competitive advantages that could influence the financial performance of your startup. Additionally, consulting industry experts and conducting thorough market research can help refine these projections and provide a more accurate representation of your coffee shop's potential financial performance. 

I thought this was a good initial summary, although many businesses do experience a net loss in the first year of business, so the change in performance from year 1 to future years doesn’t bother me too much.  I actually expect a business to lose money in the first year, but nevertheless it was good that ChatGPT identified this and other deviations from the norm so that you can address those in your business plan or pitch to potential investors or lenders. 

I hope this has been helpful for you to see how you can utilize artificial intelligence in the process of creating financial projections for your startup.  It can really serve as a research assistant and can provide you with reasonable feedback on your projections.  It can help you find where you deviate from norms and help you ensure that your numbers are reasonable for your specific business plan. 

About the Author

Adam is the Co-founder of ProjectionHub which helps entrepreneurs create financial projections for potential investors, lenders and internal business planning. Since 2012, over 50,000 entrepreneurs from around the world have used ProjectionHub to help create financial projections.

Other Stories to Check out

Common Troubleshooting Questions About ProjectionHub Templates

Check out the quick and easy ways to address some of the most common troubleshooting questions we hear about the excel templates at ProjectionHub!

5 Key Tips to Make Your Startup Business Plan Shine for an SBA Loan

Learn 5 key tips to make your startup business plan stand out and secure an SBA loan, from demonstrating market potential to creating realistic financial projections.

How to Know if Your Financial Projections are Realistic

It is important for financial projections for a small business or startup to be realistic or else an investor or lender may not take them seriously. More importantly, the founder may make a financial mistake without a reliable plan.

Have some questions?
Let us know and we'll be in touch.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.