Enjoy this on-demand webinar presented by Promevo.

This webinar, originally recorded on September 26, 2023, walks you through how to use Google's new AI tools such as Duet AI and Bard (now collectively known as Gemini). John Pettit, Promevo's Chief Technology Officer, walks you through how to increase productivity by offloading tedious tasks to Google's AI tools. Then, he explains the important news to come out of Google Cloud's flagship event, Next, held in early September 2023.

Hailee Zapata: 

Welcome to our webinar on unlocking hyperproductivity with Bard and Google AI, and also revisiting Google Cloud Next 2023. 

Over the next 40 minutes, we're going to discuss a couple of things. John's going to talk about how to get the most out of Bard and GenAI, integrating GenAI with your Google solutions, a recap on Google Cloud Next 2023, and the key takeaways. We are going to have a couple of Q&A sessions throughout our presentation with John.

Now, I'd like to introduce myself. I'm your webinar host, Hailee Zapata. I am the Alliance marketing manager here at Promevo. Promevo is a premier Google Cloud partner that sells, services, and builds Google products. 

We are 100% Google-focused and specialize in ChromeOS, Google Cloud, Google Workspace, and our own proprietary Workspace management tool, gPanel. 

Speaking of gPanel, in November, we are going to host the first session of gPanel Office Hours, so make sure you reserve your spot! 

Alright, our presenter today is John Pettit. He is our CTO and CIO of Promevo. The first part of the presentation is on Bard and Gen AI, and John actually gave a cloud talk at Google Next with this presentation. So, you're in for a real treat.

John Pettit: 

Thanks, Hailee, for the warm introduction, and welcome, everybody. Thank you for coming to our talk today.

This talk is really more about why generative AI or “Gen AI” is important and why we're at an inflection point. I'm going to talk a bit about Bard and the Google AI options. Hopefully, we'll have some questions at the end where we can delve deeper into your questions, concerns, or projects you're interested in.

So, what is an inflection point? It's a changing point in time where everything before and what we thought about changes to something new. I believe we're at that inflection point with GenAI. 

If you look through previous inflection points in history, take a journey back in time, the agricultural revolution helped people free up more time as they were able to harvest food and produce more efficiently. It was a profound change in society and allowed us to build civilizations. 

We had the industrial revolution, where machines and factories transformed our production landscape. We created goods at an unprecedented scale. Things that took days or weeks could be done in hours, and the surge in productivity really laid the foundation for the modern world.

Computers and the internet, many of us have lived through that. I was lucky enough to be born at the dawn of the microcomputer age and have been part of that evolution. That brought a lot of automation for tasks that people used to do manually, through papers or other tools. 

Now they can do it through computers, allowing them to focus more on creativity and strategic work. We're familiar with the internet, the digitization of information. People used to have mnemonic techniques to memorize information, but now it's all externalized and easily searchable. 

We're all very familiar with Google and our ability to quickly find information, but now we have Gen AI coming into play. Gen AI is fundamentally different in a couple of ways. 

In the world of the internet, you have to think of the information you're searching for, find the right link, dig deep into the pages, explore the information, and then synthesize it to understand it and use it. Gen AI can change that. 

Generative AI, built upon large language models, allows you to quickly summarize information, synthesize it, and have a personal assistant with you at every step of the way. It changes the experience and the mindset of how you think about information.

My children grew up in a time with mobile phones, and they won't remember a time before that. I think the next generation, this GenAI generation, won't know a time without having that personal assistant with them and having a different approach to problem-solving. It's going to be an important aspect when you think about bringing changes to your organization with GenAI.

At its core, what is generative AI? It's a type of artificial intelligence or machine learning that can create content like text, images, music, videos, and code. It's powered by large language models that have been trained on massive amounts of information, providing a vast reservoir of knowledge to access and generate new content. You can even make it specific to your business, like Google does with medicine or call centers.

On the right, you can see examples of an image generated by AI and a music byte created by a music LM on Google. There are also code scripts and more. 

You can accelerate your abilities, think of it as a prosthesis for human creativity. You don't have to be a great writer to write well or a great communicator to communicate effectively. One of our clients is using it for English to English translation, explaining technical content to non-technical people with great success.

Some people wonder if this will kill jobs. I don't think it will. It will create new jobs, accelerate innovation, and open up opportunities for growth. It's going to be a game-changer.

Some of the things I've personally used Gen AI for are job requisitions, writing reviews, emails, root cause analysis, generating code, and creating scaffolding. It has reduced tasks that would have taken hours to minutes and days to hours, vastly improving productivity. Our team is using it for various tasks, and it makes a significant difference. However, introducing this into your organization should not be underestimated. It's going to change the way people approach problem-solving.

When you think about change, it's important to retrain your organization to get the most out of it. You can bring in experts or early adopters who can help spread knowledge, or find value in one group and let them become influencers within your organization. This is a significant shift, like the introduction of dual monitors. It will change how people work, communicate, and collaborate, leading to increased productivity.

How can you get the most out of it? You can generate text, create content, and iterate on it. It allows you to focus on the details of the content, not the structure. Summarizing information can speed up research and analysis. It's used in various parts of the organization, from sales to marketing to technical roles, improving productivity.

Automation of tasks is another key aspect, providing instructions that the AI can follow to generate reports, summarize data, and perform repetitive tasks. Combining generation, summarization, and automation is a powerful tool that can be integrated into various workflows.

So, what's the new speed of productivity? If you look at productivity change over the last few decades, it typically hovers around 2.7%, 1%, or 2%.

Historically, the growth rate has been around two to three percent. However, Goldman predicts that we'll see eight to nine percent growth from generative AI. They believe it will drive significant GDP growth through the productivity gains it offers in our day-to-day work lives. 

Studies from MIT show an increase in productivity by 37% and an 18% improvement in accuracy. Stanford observed a 14% boost in agent productivity with AI assistance. 

The benefits are real for those who have successfully integrated this technology into different workflows. It's crucial to understand how to bring this into your organizations, both through integrated tools like Duet AI and by giving people access to specific models or prompts that align with their workflow.

So, what are your options for integrating Gen AI? First, we have Vertex AI, Google's GCP platform that provides access to various AI models, including their proprietary models like Palm, a large language model. 

Vertex AI offers tools for image customization, search, conversational AI, and medical-specific AI models. You can use Vertex AI to create your own models, leverage their models, or conditionally train their models with additional information. This is useful for translating languages, writing content, and answering questions effectively. 

If you need guidance on implementing Vertex AI for your specific business needs, feel free to reach out.

Then there's AI for GCP, a solution specifically designed for Google Cloud. It offers functionality similar to Duet or Bard but tailored to your Google Cloud environment. It can help identify vulnerabilities, find errors, and assist with coding tasks. Think of it as a technical AI assistant for your skilled team members.

Finally, Duet AI for Google Workspace is designed for office workers and offers tools that make them more productive in their day-to-day tasks. It integrates with Docs, Sheets, Slides, and Meet, providing features like automatic text suggestions, grammar correction, style suggestions, information summarization, and more. This solution is now generally available, making it accessible to a broader audience.

These are the three main ways to approach integrating Gen AI into your organization. I'll pause here to see if there are any questions from the audience.

Hailee Zapata: 

Thank you, John. That was fantastic. We do have a few questions, and the most popular one is about creating policies around AI usage and whether it's necessary.

John Pettit: 

That's a great question. Yes, it's likely necessary to create policies regarding AI usage. When it comes to using consumer-based products like Bard or ChatGPT, there may be data sharing that's not allowed within your organization due to obligatory data sharing with these providers for model training. 

For instance, Samsung faced data leakage issues with ChatGPT. It's essential to ensure that your employees aren't using these tools with customer data or other sensitive information. Google offers privacy options, allowing you to opt out of data sharing while still using the models effectively.

If you're considering business-based solutions, they generally offer more privacy controls. At Promevo and other organizations, we've rolled out policies to guide employees on which tools are approved and what data sharing policies they must adhere to. 

If a specific group within your organization plans to use a tool with more freely shared data, they may need to accept additional terms related to acceptable use.

Hailee Zapata: 

Excellent. So, the next question is about concerns that people might overuse AI to the extent that they become less skilled in core human strengths, like problem-solving.

John Pettit: 

This is an interesting concern, and it has two sides. 

On one hand, some worry that excessive reliance on AI could weaken human skills. On the other hand, it could potentially make people stronger as they gain access to additional coaching and guidance they wouldn't otherwise have. 

I lean toward the optimistic side. I've seen that as people engage more with technology, they often become stronger at tasks rather than weaker. It ultimately depends on how someone uses these tools. 

If they just copy and paste results without contributing their creativity and insight, they might produce mediocre work. However, combining AI tools with their human strengths can lead to exceptional results.

Hailee Zapata: 

That's a great perspective. It's about using both tools effectively. Our next question is about the differences between Duet AI for Workspace and Bard extensions.

John Pettit: 

Bard is primarily a consumer-based product with some extensions. While it offers some powerful capabilities, it may not provide the level of data privacy and control needed for business usage. 

Duet AI for Workspace, on the other hand, is designed with businesses in mind and offers better data privacy. It's more tightly integrated into Google Workspace and offers greater potential for custom workflows. 

Bard is like a general chat system, whereas Duet is more business-oriented.

Hailee Zapata: 

That clarifies the distinction. Our final question for this section is about the Google AI that can help with Gmail responses.

John Pettit: 

To assist with Gmail responses, Duet AI for Google Workspace is your solution. It allows you to set up automatic writing and provides prompts to help you craft effective responses. 

This feature is available in the Trusted Tester program, and it's a valuable tool for improving your email productivity.

Hailee Zapata: 

Great. Thank you, John, for addressing these questions. If we didn't get to your question, don't worry. We'll provide responses in a follow-up blog post or reach out to you directly to ensure all your questions are answered. 

Now, I'll pass it back to John for our next topic, where we'll revisit Google Cloud Next 2023.

John Pettit: 

Alright, let's talk about Google Next 2023. I'm not sure how many of you managed to attend, but I found it to be a fantastic conference. It felt like there was a significant demand for in-person conferences again. 

There were between fifteen to eighteen thousand people who attended Next in person, marking a substantial moment of growth for Google in terms of their emphasis on AI and data.

So, what were the key takeaways from the event? It's clear that GenAI is of utmost importance. Google has placed a strong emphasis on GenAI, not only as a significant product addition but as a central focal point. This emphasis was evident in every announcement across Google's product ecosystem, spanning GCP and Workspace.

Vertex AI and Duet AI have seen significant enhancements and new offerings, making them tangible tools that businesses and developers can leverage. Whether you want to build new solutions on Vertex AI or utilize any of the Duet products, you have powerful options at your disposal.

One notable announcement was the general availability of Duette for Workspace. Many of us have been using this on a trusted tester program, and it's excellent to see it evolving into an official product. This will undoubtedly improve productivity and the quality of work within organizations.

Additionally, Google Workspace and GCP have seen improvements in Gen AI capabilities. With the introduction of GenAI App Builder, you can easily create your own applications powered by AI. Google has also been working on extending Vertex AI with more features, much like the Bard or ChatGPT extensions.

In the realm of data analytics, Duette AI is available in preview for Looker and BigQuery Studio, opening up opportunities to unlock insights from data more effectively. 

Google is committed to being open, as seen in the Looker semantic layer, which is now open to Tableau and Power BI. This means you can organize your data in Looker and make it accessible to other tools you may use.

Notably, Google's open cloud approach continues to evolve. You don't have to be tied to a single cloud provider; you can use the best technology available and expand towards the best solution. This open cloud philosophy gives you the flexibility to choose the right technology for your needs.

In summary, Google Next 2023 showcased incredible energy and enthusiasm for building next-generation applications. Many businesses are eager to evolve their cloud usage. With Google's open cloud approach and advancements in AI, it's a promising time for businesses and developers to embrace these new capabilities.

Now, if we delve into specific details, we find that Google Workspace is taking a significant leap forward. Duet AI for Google Workspace is available, offering a free trial. It can significantly enhance productivity across various productivity tools. And, if you're interested in Google Workspace, Duet AI is a game-changing tool.

Data analytics have also seen substantial improvements. Duette AI in Looker and BigQuery Studio is set to unlock valuable insights from your data. Google's commitment to being open is evident in the Looker semantic layer, which is now open to Tableau and Power BI, allowing you to share data across different tools effectively.

In the cloud, BigQuery offers more integrated machine learning models and features. Google also introduced vector and semantic search, which opens up exciting possibilities in machine learning. In the database world, they offer the capability to unify data from different sources, even across different cloud platforms, with BigQuery Omni and BigLake.

Overall, Google's focus on AI and machine learning is evident in its advancements in AI and data analytics. The commitment to open cloud solutions and the improvements in Google Workspace and GCP make it an exciting time for businesses and developers.

And with that, we can move on to your questions.

Hailee Zapata: 

I think you did an excellent job summarizing the conference. I was also present at Google Next, and the continuous stream of announcements made it an exciting event. So, thank you, John, for sharing your insights. 

Before we move on to the Q&A session, I want to direct your attention to our website. We encourage you to explore the wealth of information available there. 

On the right-hand side, you'll find information about our upcoming webinars. Many questions have poured in, particularly about Duet AI and its integration with Google Workspace. We've scheduled two webinars to address these topics. 

The first, titled "Brainstorm: Getting More Out of Your Google Workspace Management Tool," and the second, "Duette AI Is Great for Google Workspace." Details about these webinars will be included in the follow-up email, which will also contain the recording of today's session. 

Now, let's proceed to the Q&A session with John.

Here's an interesting question: “How does a non-technical person go about developing a customized language model for their business? For instance, if you own a tutoring company and want to create an AI model based on your teaching materials and methods to act as a virtual tutor for your students.”

John Pettit: 

The first step for a non-technical individual would be to consider using Vertex AI's Conversational Gen AI App Builder. This tool is designed for creating conversational AI applications and allows you to integrate your own source of information, such as teaching materials and methods. 

You can train the model to become specific to your business and answer questions related to your content. You can deploy this model on your website, through an API, or in various ways to offer virtual tutoring. It's a user-friendly approach that doesn't require extensive coding. 

If you require a more customized solution, you can explore further options within Vertex AI.

Hailee Zapata: 

Great! The next question focuses on the accuracy of AI-generated responses. How can one ensure that the information provided by AI models like ChatGPT and Bard is accurate, especially considering past instances where AI data was not reliable?

John Pettit:

Ensuring accuracy in AI-generated responses is essential. AI models like ChatGPT and Bard can sometimes produce incorrect or "hallucinated" responses. 

There are several ways to safeguard accuracy. First, you can compare the model's output to a known source of truth. If you have reliable data, use it to verify the model's responses. You can also create custom functions or additional models to check the AI's answers.

Another approach is to train the model on a source of truth to reduce hallucinations. This means the model will generate responses based on accurate and up-to-date information. Grounding your AI model in trusted data sources is key to ensuring accuracy.

Hailee Zapata: 

A question with numerous upvotes inquires about the cost of Duette AI as an add-on. It appears that adding Duette AI can almost double the per-user license cost. How can one justify this cost increase when seeking additional budget from the finance department?

John Pettit:

Justifying the cost increase for adding Duet AI requires emphasizing the productivity gains and influencers within your organization. It's a similar challenge to convincing organizations to invest in dual monitors in the past. 

At first, it might seem like an unnecessary expense, but when executives and influential employees experience the productivity boost, they'll push for its wider adoption. It often starts with higher-priced individuals benefiting from the tool, and then the question becomes, "Why don't we provide this for everyone?" 

ROI, whether perceived or real, will play a crucial role in justifying the budget increase. Highlight the real use cases that improve productivity in your company, and influencers will drive the adoption.

Hailee Zapata: 

Excellent advice! Now, let's address the question of security when implementing AI into a business. What security measures should be taken to ensure the safe deployment of AI solutions?

John Pettit:

Security in AI implementation involves multiple layers. Firstly, data security is paramount. Ensure that your data is encrypted in transit and at rest. 

Most professional and business tools in GCP offer the option to opt out of data sharing. Protect your data to avoid any breaches or compliance violations.

Additionally, consider intellectual property (IP) protection. Make sure your AI implementation doesn't infringe on others' IP. Have review processes in place to prevent copying or improper use.

Lastly, ensure the security of your AI system. Implement monitoring, firewalls, and security layers. Google and other cloud providers offer security measures to safeguard AI solutions. 

If you have specific security questions or concerns, it's advisable to have a conversation to explore the best approaches for your unique needs.

Hailee Zapata: 

And now, a final question regarding costs. Do different AI tools come with separate costs, or are they bundled together in a comprehensive pricing structure?

John Pettit:

The cost structure can vary. Specifically, for Duette AI, it's offered as a separate add-on for both Google Workspace and Google Cloud. Each of these comes with its own pricing. Custom solutions built on Vertex AI are consumption-based, so costs depend on your usage. 

It's essential to understand the pricing model associated with the specific tools or services you plan to use. For Duette AI, it's offered as a bundled SKU for Google Workspace and Google Cloud.

Hailee Zapata: 

That's a comprehensive answer. Thank you, John, for sharing your insights with us today, and thank you to all our attendees for joining us. You'll find John's LinkedIn contact information on the "Meet the Presenter" slide in the presentation deck. 

If you have further questions or want to continue the conversation, don't hesitate to reach out. Keep an eye out for the follow-up email containing the recording of this session and links to our upcoming webinars. Have a great day, everyone!