Tips for Unlocking the Potential of Generative AI

Many companies are starting to actively pursue Artificial Intelligence (AI) programs.??Operations, sales, marketing, customer service, new product, and service development all provide potential areas where a company that executes a winning AI strategy can gain a competitive edge.

The surge of generative AI tools and platforms has increased the accessibility of AI to businesses looking to rapidly advance their footprint.??

For those new to AI, generative AI refers to AI models and techniques that can generate new and original content and output, based on patterns and characteristics learned from a dataset.??This can be text, speech, music, art, images, video, and more.??Companies like OpenAI, Alphabet (Google), and Microsoft have released generative AI tools through internet-based websites and mobile applications.??These include ChatGPT [1], Bard [2], and Bing AI [3].??

Businesses can fast-track the development of more advanced generative AI-based solutions using technical solutions including the OpenAI API [4], Google’s Generative AI Studio [5], or Microsoft’s Azure AI Service [6].???These solutions can allow companies the ability to create generative AI solutions quickly and without the time and expense of building out a large IT footprint.?

The OpenAI API [1], for example, provides a way for a business to tap into the power of advanced AI models without having to train or maintain these models themselves. It offers a powerful way to access state-of-the-art language processing capabilities and integrate them into applications, products, or services.??An Application Programming Interface (API)??is a set of rules and protocols that enables different software applications to communicate and interact with each other.??It acts as a bridge between different software components, simplifying the process of integration.??

To help businesses unlock the potential of a successful generative AI program there are a few recommended tips and considerations.??

Many of these also apply to other forms of AI, including robotics, computer vision, machine learning, expert systems, natural language processing, and deep learning.

1.?????Know all your data

Data is vital for AI projects.??It is the foundation for training, algorithm development, validation, and continuous improvement of AI models.??Data helps ensure AI solutions have the right information to learn, provide insight, and replicate real-world scenarios.??This in turn leads to more accurate and precise predictions and outcomes, with relevant recall, and improved overall performance.??Without access to data AI projects are like swimming pools without water.

To make certain companies have the right data for today and tomorrow, they have to get imaginative about how they think about and manage their information assets.??Data does not always come in a neatly packaged format from a transactional or traditional record-keeping system, which may include an accounting, human resources, manufacturing, or sales system.??Companies have to carefully consider all the information flowing in and around an organization, anything with a digital pulse may be a viable asset.??This non-transactional information can include sensor, video, audio, geospatial, log file, and social.

All data needs to be understood, categorized, stored, protected, and available for use in a legal and ethically compliant manner and treated like an investment.

2.?????Build business use cases around what is on hand today

Determining where to start is often the most difficult step on any journey, although it should be the easiest.??Most businesses can quickly come up with a list of on-hand data, manual processes, high-touch and low-value service-based interactions, and a backlog of new features or enhancements for their products, and services.??

Reviewing and scoring that list in the context of potential generative AI projects provides a good starting point for additional analysis and business case creation.??

If all else fails, ask ChatGPT, Bard, or BingAI for ideas on business cases for a company in a specific industry and market.??It might just spark some ideas.??To see this exact example using ChatGPT jump to the “Appendix - Ideas in Action” section at the end of this article.

3.?????A culture open to experimentation and research needs to be fostered?

We are in the early days of the generative AI marketplace lifecycle.??No one with certainty can predict the future.??A decade from now or sooner there will be winners and losers.??The winners will be companies that create something significantly unique and bring it to market faster than their competition.??That takes capital supported by a willingness to take risks by nurturing and investing a company’s resources in AI projects that may or may not directly drive financial reward.

Paying for and justifying experimental research and development (R&D) for an existing company can be a challenge.??An approach well suited to fund projects when there is a new and evolving marketplace is to create a reinvestment-based R&D plan.??In its simplest form, a business allocates a percentage of profits or savings from traditional business case-based projects with predictable outcomes and Return on Investment (ROI) to fund experimental-based research and development.??That boring, but cost-saving accounting system enhancement to speed up the collection of invoices, can be used to fund generative AI-based R&D, which may lead to a breakthrough product or service without requiring any new additional funds.

4.?????Don’t overlook current enterprise applications, platforms, and systems

AI is often closer than most business leaders know.??Many software providers including Microsoft, Workday, SAP, Oracle, Google, and Salesforce have made significant AI investments in their product suites.??Companies should take some time to learn what has been released by their current software vendors and what is on their product release roadmaps.

A business may have undiscovered AI in-house already that can be put to immediate use to cost-effectively fast-track its AI enablement.??

Remember to review the AI offerings from all software vendors and not just the big names highlighted and check in regularly.

5.?????Think about more than efficiency and productivity

Many companies when developing a portfolio of potential AI business cases, focus on operations and technology-driven efficiency and productivity.??Optimizing revenue expansion is often overlooked.??

Include marketing, sales, and product teams in the brainstorming process.??Wins in these areas can have a major impact on a company’s top and bottom line.??A few potential ideas to focus on include:

• Lead generation
• Prospect scoring
• Intelligent scripting
• Demand prediction and forecasting
• Price optimization
• Personalized recommendations?
• Creative design
• Quality checks
• Market and prospect research
• Virtual sales and service assistants
• Automated response and follow up
• ?Language translation services

6.?????Set rules around how generative AI can be used today

Employees may already be using generative AI to perform their job functions today using unapproved company resources or their own devices.??Set the ground rules for approved tools and usage.??This is important to prevent Intellectual Property (IP) loss, data leakage, security issues, privacy violations, misrepresentation, inaccuracies, and copyright infringement.

To mitigate these risks, companies must establish clear guidelines, policies, and approval processes regarding the use of AI tools or non-approved devices. Employees should be educated about the potential risks and consequences of unauthorized usage and be encouraged to seek approval before using such tools. Companies may also consider implementing new security measures, access controls, and monitoring systems to prevent unauthorized AI tool usage.

7.?????Recognize generative AI is not quite perfect

Generative AI like humans is not perfect.??Data, algorithms, and models can produce an unintended bias or be factually incorrect.?

OpenAI’s ChatGPT chat interface highlights this point,?

“ChatGPT may produce inaccurate information about people, places, or facts.”

It is important to ensure solutions do not introduce or magnify biases in a discriminatory or unfair manner, violate privacy, and are ethical in their usage.??In today’s hyper-socialized and connected world, one misstep can have legal consequences and damage a brand or business.??

To help ensure that solutions carefully consider, comply with, and monitor the ethical, legal, and compliance landscape create an ongoing framework that includes the following:?

• Diverse data collection
• Fairness metrics and testing
• Bias mitigation techniques and sample review
• Transparency and explainability
• Continuous monitoring
• Ethics review
• Cross-organizational stakeholder involvement

8.?????Be flexible and keep an ear to the ground

Not too long ago conversations about generative AI did not exist beyond the confines of big tech firms. Most people never heard of OpenAI and ChatGPT.??A majority of corporate annual reports and strategic plans did not consider or address how AI and generative would impact their business, employees, customer, and competitors.??

Six months ago, that rapidly changed. We are still, by most expert accounts, very early in the AI innovation and adoption cycle. As generative AI evolves, closely monitor new developments, tools, laws, and government regulations. These all can have a major impact on the long-term success of a company’s AI program.

Appendix - Ideas in Action

Sample industry-specific business case example

OpenAI’s ChatGPT [4] was prompted to, “Please provide some ideas on generative AI business cases for a brokerage firm in the finance industry.”

Model: Default GPT-3.5 response

“Certainly! Here are some generative AI business case ideas specifically tailored for a brokerage firm in the finance industry:

1. Financial Report Generation: Develop a generative AI system that can automatically generate comprehensive financial reports based on relevant data inputs. This can streamline the process of producing reports for clients, saving time and improving accuracy.
2. Investment Portfolio Simulation: Create a generative AI model that simulates the performance of investment portfolios under various market conditions. This can help clients understand the potential risk and return of their portfolios and make informed investment decisions.
3. Automated Market Commentary: Utilize generative AI algorithms to automatically generate market commentary and analysis based on real-time market data. This can provide clients with timely insights, keeping them informed about market trends and investment opportunities.
4. Financial News Generation: Develop a generative AI system that generates financial news articles or summaries based on relevant market events, company earnings, and economic indicators. This can provide clients with curated news updates, keeping them informed about the latest developments.
5. AI-powered Trading Strategies: Utilize generative AI algorithms to develop novel trading strategies by analyzing historical market data, technical indicators, and news sentiment. This can enhance the firm's trading capabilities and potentially generate alpha for clients.
6. Personalized Financial Planning: Create a generative AI system that generates personalized financial plans for clients based on their financial goals, risk tolerance, and investment preferences. This can assist clients in developing tailored investment strategies and optimizing their financial decisions.
7. Fraud Detection and Risk Assessment: Employ generative AI models to detect patterns and anomalies in financial data, aiding in the identification of potential fraud or risk factors. This can enhance the firm's risk management capabilities and protect client assets.
8. Chatbot Assistance: Develop a generative AI-powered chatbot that assists clients with their investment-related queries, providing personalized recommendations, answering common questions, and offering support. This can improve customer service and engagement.
9. Sentiment Analysis: Utilize generative AI techniques to analyze social media data, news articles, and other textual sources to gauge market sentiment. This information can be used to identify emerging trends and sentiment shifts, informing investment strategies.
10. Portfolio Optimization: Create a generative AI system that optimizes investment portfolios based on client preferences, risk profiles, and market data. This can help clients achieve better diversification, risk management, and performance.

These generative AI business cases have the potential to enhance various aspects of a brokerage firm's operations, including research, client services, risk management, and decision-making processes. It's important to consider the availability of quality data, compliance with regulatory requirements, and the need for human oversight when implementing generative AI solutions in the finance industry.”

Model: Default GPT-4 response

“Sure, generative AI can be used in a variety of ways to add value to a brokerage firm. Here are some possible applications:

1. Automated Report Writing: Generative AI can be used to write financial reports, market analyses, and investment recommendations. It can take in structured data (like financial indicators) and generate well-structured, human-readable text. This can greatly increase productivity by freeing analysts to focus on higher-level tasks.
2. Client Communications: AI could handle routine customer communications, like answering frequently asked questions, guiding clients through website navigation, or even providing simple investment advice. This would free up customer service representatives to handle more complex inquiries.
3. Tailored Investment Advice: AI can be used to generate personalized investment advice for clients. Based on a client's risk tolerance, investment goals, and other factors, AI can generate a unique, tailored portfolio recommendation.
4. Risk Management: AI can generate predictive models for risk management. For example, it could use historical data to predict the likelihood of a client defaulting on a margin call or the risk of a specific trade.
5. Compliance and Regulatory Reporting: Brokerage firms are required to comply with numerous regulations and to provide regular reports to regulatory bodies. Generative AI can help automate these reporting tasks, ensuring accuracy and reducing the burden on staff.
6. Trade Order Generation: Generative AI models can be trained to understand market trends and indicators, and generate trading signals or even fully autonomous trading strategies.
7. Marketing and Sales: AI can generate personalized marketing content or sales pitches, tailored to the individual client's needs, preferences, and investment history.
8. Training and Simulation: Generative AI can be used to create realistic training scenarios for new brokers. By simulating various market conditions and client interactions, AI can help new brokers learn the ropes more quickly.
9. News Generation: Brokerage firms can use generative AI to write timely articles or news updates about the stock market, specific securities, or general economic conditions. These could be offered as a value-added service for clients.
10. Anomaly Detection: AI can be used to scan for anomalous patterns in transactional data, flagging potential fraudulent activity or unusual trading patterns that could indicate market manipulation.

Each of these applications could potentially add significant value to a brokerage firm, either by increasing efficiency, improving customer service, or offering new services to clients.”

References and Citations

1. OpenAI Website, https://openai.com/, June 26, 2023
2. Google Bard Website, https://bard.google.com/, June 26, 2023
3. Microsoft Bing Website, https://www.microsoft.com/en-us/bing?form=MA13FJ, June 26, 2023
4. OpenAI API, https://openai.com/product#made-for-developers, June 26, 2023
5. Google Generative AI Studio, https://cloud.google.com/generative-ai-studio, June 26, 2023
6. Microsoft’s Azure AI Service Website, https://azure.microsoft.com/en-us/solutions/ai/, June 26, 2023
7. OpenAI Chat GPT Website Prompt, https://openai.com/chatgpt, June 26, 2023

Cover image from Shutter Stock, , June 26, 2023

Disclaimer, Copyright and Trademark Statement

This article is provided for informational and educational purposes. It makes no warranties as to the claims, accuracy or fitness of information provided, referenced or cited. Use of the information, instructions and any examples contained in this work is at your own risk. There should be no implied endorsement of this article by any person or organization referenced.?

All trademarks, company, product and services names, images, descriptions, or public website content are property of their respective owner as source referenced. It is your responsibility to ensure that your use thereof complies with such licenses and/or right.