Everyday, more and more small businesses and startups are leveraging generative AI for their internal and customer-facing needs, with over 70% of gen AI’s billion dollar-unicorns using Google Cloud. As the technology continues moving at a rapid pace, with almost daily announcements of products and features, businesses are discovering new use cases and ways to leverage gen AI tools.

In Google Cloud monthly gen AI workshops, Google Cloud is meeting small businesses to provide them with a technical overview of our key gen AI products, AI techniques, and best practices. Our workshops are organized and led by a team of Specialist Customer Engineers, Monique Sullivan, Bruk Zewdie, Isaac Attuah, and myself. They ignite imagination, as together, we brainstorm and ideate with our customers to blend their creativity with Google Cloud’s hands-on experience deploying large models. Google Cloud workshop provides a collaborative environment to unearth innovative applications and use cases. Working with Google Cloud team, we explore industry use cases, confront and assess common challenges, and identify successful techniques in deployment strategies.

If you are a startup or small business looking for some inspiration on how to use gen AI, this article will explore some of the prominent themes and trends from Google Cloud most recent workshops—and explain how you can participate in a future edition.

Popular use cases

Customers across a range of industries are using the power of generative AI, from doctors helping patients, to businesses empowering employees and customers, to individuals bridging knowledge gaps or enhancing visual creativity. Below are some of the industry use cases that have sparked excitement in our recent sessions.

Healthcare:

  • Carry medical assistance in your pocket—Help doctors focus on patients rather than paperwork, with an AI physician’s assistant always at the ready.
  • Get a bespoke workout routine via an AI personal trainer that tailors its advice from a library of fitness videos and regimens.
  • Enable precision quality control – evaluating healthcare provider performance using statistical analysis of clinical data.

Employee Support:

  • Instill company wisdom with an HR chatbot answering policy questions 24/7, guiding employees with institutional knowledge.
  • Automate headshot quality by leveraging AI to ensure polished employee profile pics, optimized with studio-quality lighting and touch-ups.
  • Quickly find answers with a large language model(LLM)-generated knowledge base—a personalized, up-to-date employee handbook.

Customer Support:

  • Deliver instant, personalized support with chatbots that engage customers using natural conversations and contextual understanding.
  • Remember each customer via support bots that tap into full histories and preferences for tailored, seamless issue resolution.
  • Make better decisions with accurate forecasting at your fingertips, powered by an AI assistant that predicts outcomes using past data.

Visual-based:

  • Accelerate game development by leveraging AI to automatically generate graphics and visual assets.
  • Gain actionable insights by analyzing videos with AI to identify trends, movements, and anomalies.
  • Automate custom illustrations, using AI to create original drawings for children’s books to spark young imaginations.
  • Condense information into dynamic videos, transforming documents into shareable lessons with AI-generated animations.

Bridging Gaps:

  • Making data accessible—automatically translate complex financial reports into plain language for any audience.
  • Drive personalized recommendations, using AI to analyze assessment data and propose customized teaching strategies.

Successful prompting techniques:

In addition to use cases, workshop participants are also eager to understand how to more effectively interact with generative models— that is, how to prompt them.

Prompting is the art of guiding an LLM with your input so it provides your desired response. Throughout our workshops, our attendees share and unlock effective techniques like:

  • Establish the LLM’s expertise and role upfront in your prompt so it understands its persona.
  • Have the LLM restate prompts to ensure it’s correctly processed your input.
  • Use structured instructions for clarity: “summarize the input: [text.]”
  • Ask the LLM to craft specific prompts tailored for your needs.
  • Streamline prompts like Google searches – no fluff.
  • Organize the prompt with lists or dictionaries.
  • Insert “variables” into the prompt, allowing users to customize the prompt by just changing the variable words.
  • Prompting is a skill that mixes your personal creativity with your industry knowledge – try multiple approaches to get different results.