Chapter 5. Use Cases for an AI Project

Chapter 5. Use Cases for an AI Project

What are Use Cases?

The term “use cases” is the industry term for the various ways that LLMs and AI engines are used. Consumers and businesses use AI tools differently, and hence, have different use cases.

Consumers are using ChatGPT mainly for question answering, writing text, brainstorming, and as a companion. Businesses are using old-school predictive AI to give content feeds to their customers, and using generative AI for customer support chatbots and marketing content writing.

Use cases are continually expanding, and the onset of new “on-device” use of generative AI in AI phones and AI PCs will further grow the role of LLMs. The newer set of “agent” AI architectures are broadening the use cases to cases where the LLM takes an action on your behalf, such as sending an email or text.

Gen AI vs ML Use Cases

Generative AI is the new kid on the block. ML is the incumbent technology, already living the high life outside the DMZ.

They have different strengths and weaknesses. Hence, to understanding what projects to build with generative AI versus ML, it’s important to compare them.

Did you notice anything, or were you just skimming? ML and generative AI are quite complementary technologies. It’s not very likely that you want to replace an ML application with a version that uses an LLM. Instead, you want to look for new uses of the new AI technology that is very good at both understanding and writing natural language.

The use cases that ML is already used for, which is the same as the use cases you probably won’t use gen AI for, includes:

• Content recommendations
• Product recommendations
• Pattern detection in numeric data (e.g., quantitative trading analysis)
• Face recognition
• Cyber security (e.g., intrusion detection)
• Financial fraud monitoring
• Machine vision (e.g., autonomous vehicles)

However, you can combine gen AI and ML. For example, in the product recommendation space, the ML tools are best at choosing what to recommend, but an LLM is better at writing personalized content about that product when presenting it to the user in an encouraging way.

Longstanding AI Use Cases

AI has been around for years, and there are various things it’s always been good at, long before generative AI and LLMs came along. Some examples of things that computers could do before ChatGPT include:

• Categorization
• Summarization
• Machine language translation (i.e., foreign language translators)
• Voice recognition and understanding
• Transcription (voice-to-text)
• Voice generation (text-to-voice)
• OCR (optical character recognition, image-to-text)
• Content feed relevancy algorithms
• E-commerce product recommendations
• Sentiment analysis
• Named Entity Recognition (NER)
• Pattern detection in complex numerical data
• Facial recognition (e.g., your face is your password)

A lot of these items are technological features rather than use cases. There’s nothing wrong with aiming for any of the above ideas, if they fit with your business needs. The only point is that these are mostly old-school ML capabilities, rather than tasks that LLMs can do well.

Business Use Cases

There are many, many business projects ongoing that use generative AI capabilities. Some of the initial areas of success have included:

• Customer support chatbot
• Support staff internal scripts (generation)
• Writing content (various uses within an organization)
• Marketing content drafting/generation
• Product listings (e.g., e-commerce product descriptions, real estate listings, etc.)
• HR internal chatbot (employees searching policies)
• HR resume analysis
• Programming (e.g., coding copilots, debugging, etc.)
• Personal copilot (e.g., help staff productivity with email or document drafting)

The above list tends to be the quicker wins with generative AI. Some of these can also be done as a “buy” project rather than “build.” Furthermore, you’ll notice that most of these are staff uses rather than customer-facing applications, which makes sense because internal projects are easier than external ones.

Consumer Uses of Gen AI

Even if you’re looking for business AI projects, it can be helpful to consider what your customers are already doing with AI. This gives insight into how your customers might want to interact with your business in the future. Or possibly, you’re a technology startup writing AI apps for consumers, like me.

What are consumers using generative AI for:

• Generating things (text, video, audio, etc.)
• Companions
• Question & answer
• Internet search with summary (Google, Bing)
• Drafting text documents (e.g., emails, reports, letters, etc.)
• Outlining
• Proofreading, editing, and revising
• Searching long documents (or the internet)
• Summarization (e.g., PDF documents, web pages, long documents, legal briefs, etc.)
• Lists of ideas or topics
• Brainstorming
• Homework answering
• Homework drafting (for teachers)
• Chatting (who knew?)
• Recipes
• Researching products and services
• Travel research and planning
• Tax questions
• Financial questions
• Medical questions

What are they generating? (for better or worse)

• Creative writing (stories, poems, song lyrics, jokes, etc.)
• Non-fiction writing (e.g., business reports, homework, assignments, CVs/resumes, job application letters, etc.)
• Image creation (text-to-image)
• Video creation (text-to-video)
• Music creation (text-to-music)
• Song creation (text-to-lyrics)

No doubt, there’s some things missing from that list, but I’m sure you get the idea.

Software Engineering Use Cases

Programmers are also using code models for:

• Code auto-completion
• Code generation
• Debugging
• Documentation
• AI Game creation tools (levels, NPC characters, etc.)
• Data generation
• Test case generation

Actually, not really. We programmers just pretend to use AI tools to keep our boss happy, but we simply don’t need any help writing perfect code. Those time blocks we mark on our calendar as “debugging” are actually spent outside playing Ultimate Frisbee.

Instead, it’s all the other mundane stuff that we need help with: pull requests, commit comments, change logs, ad hoc scripts, status reports for the boss, emails to other easily-offended humans, sick notes when hung over, and asking favors of another team in the most flattering way.

Professional Use Cases

Professional consumers and small business owners are finding a myriad of uses for generative AI in their everyday lives and in their work. But it’s early days with adoption of the technology in a meaningful way, and there are still many areas where generative AI is not yet making any impact.

Professional business uses for individuals include:

• Drafting product descriptions and other content
• Email drafting (with personalization added)
• Blog posts and online content
• Advertising and marketing content creation
• Coding copilot with code completion and debugging (for programmers)
• Writing copilot (for writers)
• Drafting business reports

Some of the more specialty fields include:

• Researching complex issues (scanning document databases)
• Cybersecurity
• SEO content auto-generation
• Automating business processes
• Medical image analysis (radiologists)
• Medical notes and patient case summaries (doctors)
• Legal document analysis and summarization
• Medical research (e.g., drug discovery)
• Agriculture yield optimization

This is a long list and also a short list. The potential benefits of AI extend into many areas and the field is only getting started.

References

1. Rhiannon Williams, July 8, 2024, How to use AI to plan your next vacation: AI tools can be useful for everything from booking flights to translating menus, MIT Technology Review, https://www.technologyreview.com/2024/07/08/1094733/how-to-use-ai-to-plan-your-next-vacation/
2. Alistair Barr, May 16, 2024·, What’s the killer AI app for consumers? Google finally has a contender. https://www.yahoo.com/tech/whats-killer-ai-app-consumers-175056899.html
3. MV Financial, Jul. 20, 2024, The Search For The Killer AI App, https://seekingalpha.com/article/4705253-search-for-killer-ai-app
4. David Linthicum, Feb 13, 2024, 3 killer apps for cloud-based generative AI, https://www.infoworld.com/article/2335997/3-killer-apps-for-cloud-based-generative-ai.html
5. Olivia Moore, March 13, 2024, The Top 100 Gen AI Consumer Apps, https://a16z.com/100-gen-ai-apps/
6. Kate Knibbs, Sep 18, 2024, Most US Teens Use Generative AI. Most of Their Parents Don’t Know, Wired, https://www.wired.com/story/teens-generative-ai-use-schools-parents/
7. Stephanie Houde, Vera Liao, Jacquelyn Martino, Michael Muller, David Piorkowski, John Richards, Justin Weisz, Yunfeng Zhang, 2 Mar 2020, Business (mis)Use Cases of Generative AI, https://arxiv.org/abs/2003.07679
8. Henrique Centieiro & Bee Lee, Jun 26, 2024, Top 20 GPT-4o Use Cases That Actually Improve Your Everyday Life, https://medium.com/the-generator/top-20-gpt-4o-use-cases-that-actually-improve-your-everyday-life-c136f2c802d2
9. Rafe Fletcher, July 11, 2024, Finding AI’s Killer Use Case, https://www.linkedin.com/pulse/finding-ais-killer-use-case-rafe-fletcher-7welc/
10. Gene Rapoport, Sanjin Bicanic, Jue Wang, Richard Lichtenstein, Arjun Dutt, June 20, 2024, AI Survey: Four Themes Emerging: If 2023 was about experimentation, 2024 is all about results. Bain & Company, https://www.bain.com/insights/ai-survey-four-themes-emerging/
11. Nicholas Carlini, 2024-08-01, How I Use “AI”, https://nicholas.carlini.com/writing/2024/how-i-use-ai.html
12. Christopher Tao, Aug 2024, Do Not Use LLM or Generative AI For These Use Cases, https://pub.towardsai.net/do-not-use-llm-or-generative-ai-for-these-use-cases-a819ae2d9779
13. Harshita Katiyar, Aug 1, 2024, Artifacts: Top 10 Mindblowing uses of Claude 3.5 Sonent, https://medium.com/aimonks/artifacts-top-mindblowing-uses-of-claude-3-5-sonent-6830b2acfa4b
14. Joe McKendrick, March 28, 2024, This year’s top 8 use cases for AI, and what tech professionals need to support them, https://www.zdnet.com/article/this-years-top-8-use-cases-for-ai-and-what-tech-professionals-need-to-support-them/

• Online: Table of Contents • PDF: Free PDF book download • Buy: Generative AI Applications: Planning, Design and Implementation

The new Generative AI Applications book by Aussie AI co-founders: Deciding on your AI project Planning for success and safety Designs and LLM architectures Expediting development Implementation and deployment Get your copy from Amazon: Generative AI Applications