How to Start Using Generative AI (Like GPT-4) in Your Business

Carlos Oliveira

AI in business blog featured image

Recently, we talked in-depth about what generative AI is and how your business can benefit from some of its more popular tools, such as ChatGPT, MidJourney, and Jasper.

Today we’ll talk about practical ways in which you can apply these platforms to your business, even in different industries.

No matter if you operate in healthcare, proptech, financial services, media, or another vertical: surely there’s an AI-enabled benefit around the corner. And these can come in different forms: maybe in streamlining the way the team works and delivers, or maybe in finding exciting new opportunities in the products and services you offer to your customers.

If you want to know more AI, check out these links:

The Impact of Generative AI on Businesses

We live in interesting times. Every day, more and more examples of artificial intelligence uses, previously only seen in sci-fi works or theoretical papers, become reality. From bots that are more empathetic with patients than doctors to chart-reading models that can summarize complex reports in seconds, AI is already reshaping the way businesses operate and the products they offer.

The first big impact is on productivity. Increased quality and speed through generative AI is already a reality in many industries. By leveraging generative AI, business owners can achieve higher productivity, reduce costs, and enhance overall efficiency.

But the big explosion is yet to come, according to many studies and analysts. Across different verticals, we’ll leverage generative AI to further optimize production processes, improve quality control, enhance demand forecasting, help retailers better understand consumer preferences, create tailored marketing campaigns, and much more.

In the report "The Economic Potential of Generative AI: The Next Productivity Frontier", from January 2023, McKinsey forecasts the potential value added to specific industries. This value will be added through increased productivity in domains such as Marketing and sales, customer operations, supply chain, and more.

Industry

Potential productivity lift

High tech

$240 billion to $460 billion

Banking

$200 billion to $340 billion

Pharmaceutical and
medical products

$60 billion to $110 billion

Education

$120 billion to $230 billion

Telecommunications

$60 billion to $100 billion

Healthcare

$150 billion to $260 billion

Insurance

$50 billion to $70 billion

Media and entertainment

$60 billion to $110 billion

Advanced manufacturing

$170 billion to $290 billion

Consumer packaged goods

$160 billion to $270 billion

Advanced electronics and
semiconductors

$100 billion to $170 billion

Retail

$240 billion to $390 billion

Real estate

$110 billion to $180 billion

Energy

$150 billion to $240 billion

Administrative and
professional services

$150 billion to $250 billion

Chemical

$80 billion to $140 billion

Basic materials

$120 billion to $200 billion

Construction

$90 billion to $150 billion

Agriculture

$40 billion to $70 billion

Public and social sector

$70 billion to $110 billion

The impact will be felt deeply across different levels – not only in overall industry trends but in the daily routine, tasks, and skill requirements of each worker.

In the US alone, up to 40% of working hours can be impacted by Large Language Models such as ChatGPT, according to a presentation by Accenture covering multiple industries.

Why is this the case?

Cta image

The Impact of Generative AI on Specific Industries

Let’s take a deeper look at the growing role of generative AI in four key industries.

Retail

When we think about genAI in business, retail is one of the first that comes to mind: intelligent, helpful chatbots that can talk just like humans, providing 24/7 availability. But there’s so much more that can be done.

A study from the World Economic Forum points out the possible benefits of artificial intelligence in retail. Here are the ones more connected to generative AI:

  • Automation – AI technology can be used to automate various tasks in the retail industry, such as inventory management, data analysis, and customer support.
  • Cost reduction – By automating tasks and optimizing processes, AI can help retailers reduce costs associated with labor, inventory management, and supply chain operations.
  • Supply chain optimization – Analyzing data from various sources, such as customer demand, production capacity, and transportation routes, to optimize supply chain operations and improve efficiency.
  • Customer satisfaction enhancement – personalizing customer experiences by providing tailored recommendations, and improving customer support by analyzing data and understanding their preferences.

Real-life example: CarMax reduced the time necessary to condense customer reviews from 11 years to a few months by using AI.

Automation potential: In retail, 34% of the tasks commonly performed have a high potential for being automated by AI, according to the Accenture report.

Banking and Finance

You may think genAI and financial services don’t have much in common. But, as a matter of fact, they do. A lot. Did you know we are at the frontier of a fraud detection revolution? It’s thanks to new language models that cut back analysis time from hours (or more) to seconds.

These are some of the biggest benefits of generative AI in banking and finance, according to McKinsey:

  • Risk assessment and fraud detection – Financial institutions can leverage AI to parse and summarize numerous complex reports, pinpointing trends and outliers that indicate unusual activity.
  • Customer service improvement – Customer-facing chatbots can assess user requests and direct them to the most suitable service expert based on various characteristics, improving processes and addressing customer requests more effectively.
  • Hyper specialization – Building virtual experts for environmental, social, and governance (ESG) topics by synthesizing and extracting information from long documents (including pictures and tables).
  • Augmenting employee performance - By training AI bots on proprietary knowledge, banks can provide deep technical support to frontline workers, improving customer experiences.

Real-life example: Morgan Stanley is building an AI assistant using GPT-4 to help wealth managers quickly access and synthesize information from a massive internal knowledge base.

Automation potential: Accenture posits Banking has the highest potential for automation across all industries, with 54% of tasks being highly automation-ready.

Medicine, pharmaceutical, and medical products

By definition, chatbots won’t ever replace the necessary human touch. But they can already do a great job in conversing with multiple patients at the same time and making them feel heard, freeing up the schedule of healthcare professionals.

These are some of the most likely genAI uses so far in healthcare and related verticals, as per McKinsey.

  • Enhancing indicator finding – Identifying and prioritizing new indications for medications or treatments by speeding up the analysis of indicators (real-world data, including patient clinical events, medical histories, and evidence-backed figures)
  • Improve automation of preliminary screening – Automating the initial screening of chemicals in the lead identification stage of drug development.
  • Customer service improvement – Communicating with patients in a clear but empathetic way, which also has the potential to relieve doctors and other medical personnel of non-specialized duties.

Real-life example: The health app JanuaryAI uses GenAI to estimate how each food will affect the glucose levels of the user.

Automation potential: Only 22% of the practitioners and technical tasks in Healthcare are highly prone to automation. When it comes to support staff in Healthcare, the figure is 27%.

Proptech

Being an industry with such a huge impact on the daily life of most people, proptech’s AI-enabled advancements will be seen with excitement by professionals and consumers alike.

Can we imagine a world where obtaining a mortgage is a simple process? Maybe, according to the genAI disruptions possibilities, as highlighted by TechCrunch:

  • Broker bots – Assisting or even replacing human agents in searching complex databases of real estate opportunities, creating personalized listings, and then connecting prospect buyers with the right real estate offer.
  • Mortgage approval simplification – Performing deep research and creating useful personas and projections for each client, speeding up the whole process.
  • Insurance policy automation – bots can improve the customer journey from acquisition and underwriting to policy administration and claims management, enhancing the efficiency and effectiveness of the insurance process.

Real-life example: AI-generated 360 virtual tours of real estate created with the help of an algorithm that stitches images and provides information such as floor planning and measurements. Implemented by the British startup Giraffe360.

Automation potential: Sales and related, the closest industry to proptech in Accenture’s report, is listed with 49% of tasks having high automation potential.

4 Ways Your Employees Can Use Generative AI In Their Daily Work

Considering how easy it is to find dedicated AI tools online, there are no excuses for not trying out new methods of optimizing different activities.

Below, we list some practical ways in which your team can start leveraging artificial intelligence to become more productive starting now.

1. Creating image captions or alt attributes

GPT-4 introduced the capability of parsing images. That’s great news to marketers, SEO professionals, and accessibility specialists.

One common task in these three roles is writing down descriptions of images so that they are easier to understand by robots and humans with limited eyesight.

  • Before: opening each picture, thinking about a description, writing it down, and fine-tuning it to ensure it’s relevant to the desired audience.
  • With AI: uploading one or more pictures to genAI service, receiving a description in seconds, editing it to follow necessary requirements.

2. Summarizing documents and reports

Analyzing dense reports is getting easier with services that even read charts and can spot outliers,

“I use services like PDF.AI when I need to know specific things about complex and long PDF documents. It’s also great that I can ask questions in a human language.”

isaac-quaino Isaac Quainoo Designer at Monterail
  • Before: Opening report, skimming document, reading it whole or browsing main sections, preparing a summary.
  • With AI: Uploading report to AI service, asking for a summary, and preparing relevant prompts that may reveal interesting data buried within it.

3. Taking smarter notes

The act of taking notes can be extremely time-consuming, but helps to ensure meetings are more productive and people, less forgetful.

There are now dedicated services that focus on transcribing and summing up online meetings in real-time (including dividing the information into actionable items).

But it can also be without specialized software.

“For example, the other day I was on a walk and took some voice notes on my iPhone. They were very chaotic, so I copy-pasted them to ChatGPT and received a nice, tidy summary in bullet points.”

grzegorz-hajdukiewicz Grzegorz Hajdukiewicz Chief Delivery Officer at Monterail
  • Before: Writing down notes during the meeting or while listening to a recording, reviewing notes, organizing them into contextual and actionable items.
  • With AI: Receive instant transcription during the meeting or from uploading a recording, ask genAI to organize text into the desired format.

4. Speeding up research and creating knowledge bases

Analysts and researchers spend valuable time creating, expanding, and maintaining tables and knowledge bases. Such work often involves connecting multiple sources of information, which can be cumbersome.

One growing set of tools demonstrates how the activity is about to change: Google Sheets extensions that populate columns with the requested information, saving time and effort.

By simply providing the necessary commands and available data, users can streamline the data population process within their spreadsheets.

It’s similar to when you replicate the same formula across multiple cells but now with tailored content. Examples: for a list of Fortune 500 companies, you can prompt the tool to write down their current CEO and CFO.

“For example, I add a list of companies and then task the extension with creating a brief description of each of them based on information found online.”

jakub-borecki Jakub Borecki Outbound Campaign Coordinator at Monterail
  • Before: Populating a table with the list of clients, researching information, such as industry and company size, of each client individually.
  • With AI: Populating a table with the list of clients and prompting an extension to fill in the required data.

What are the Limitations of GPT-4 and Similar?

All things considered, GPT-4 and similar tools have one characteristic which is both their strength and shortcoming: they are general and appeal to a wide array of uses.

For a more in-depth look at GPT-4's potentials and shortcomings, check 4 Features GPT-4 Is Missing – and What’s Next for OpenAI’s Generative AI.

Looking at the most successful cases in the industry, it’s easy to see that off-the-shelf solutions such as ChatGPT are not always ideal. Sure, they have a lower cost and fast time-to-market.

But they are limited and this wide appeal means they can’t be adapted to very specific use cases.

But that shouldn’t stop enterprising minds. Let’s briefly talk about how to move forward from here.

Addressing the Limitations

Option 1 - Developing yourself

The first option is to create a solution yourself. This is likely the most costly option in terms of time and resources since you will likely need a team of experts ready and focused on the task. But also it gives you complete control over the look, feel, and set of features of your program.

Since more and more APIs are being made available, it’s not necessary to start from scratch and compete head-to-head with big companies. Instead, you can leverage their resources and develop plugins, extensions, or other software with specific purposes.

How to access the GPT-4 API

As of writing, it’s necessary to join a waitlist to be granted access to the GPT-4 API. Note that, as with most APIs, it’s charged depending on how much you use it.

Read more in the documentation of GPT - OpenAI API.

Option 2 - Finding an AI-enablement partner

The safer and faster option is to find a partner that can help enable AI in your products and services.

At Monterail, we have interesting business use cases, both with partners and internally. To learn how we can boost success together with artificial intelligence, get in touch with us.

Cta image

Over to you

Generative AI is here to stay and business owners should see it as an opportunity to drive their companies forward.

That involves being prepared for the ways how your industry will be affected and ensuring your workforce is familiar with AI – especially considering how, in key industries, some roles will change completely in the near future.

The competition is fierce, but the good news is you don’t have to do it alone. You can find an AI-enablement partner to help explore how AI can change your business for the better.

Carlos Oliveira avatar
Carlos Oliveira
Tags