AI has sparked much controversy; it has been introduced into our daily lives but its use has been limited to certain personal tasks. Our head of technology has designed a bot that assists in the onboarding process. In this episode, we chat with him to discuss how he came to create this and what his vision is for the current and near-future landscape of AI. We also test the bot live!
Episode Highlights:
The buzz surrounding AI has reached a fever pitch, and it’s no wonder why. With its potential to revolutionise industries and streamline operations, AI is a game-changer. In our latest podcast episode, we delve into this fascinating world by exploring the creation and implementation of an AI assistant designed to enhance our company’s onboarding process. Joined by Carl, our Chief Technology Officer (CTO), we uncover the story behind this innovative project and its practical applications.
Carl’s journey into creating our AI assistant began serendipitously. He found himself with an unexpected eight-hour delay at the Brisbane airport, providing him with a rare opportunity to dive deep into the latest developments in AI. Carl used this time to explore PrivateGPT, a tool that had just been released, and decided to give it a go. Within those hours at the airport, he managed to set up a basic version of what would become our AI assistant.
He leveraged PrivateGPT, a virtual machine, and some Python scripting to create a prototype. By integrating OpenAI’s large language model, Carl drastically improved the speed and accuracy of the assistant. Upon returning to the office, he introduced this project during our IQP cycles, quarterly initiatives aimed at driving innovation within the company. This fortuitous timing allowed Carl to further develop the assistant into a robust tool that could significantly enhance our onboarding process.
AI has come a long way from the days of simple chatbots. Traditional chatbots operated on a set of predefined responses, often leading to a frustrating user experience. Today’s AI assistants, however, are powered by advanced natural language processing models, enabling them to understand and respond to a wide range of queries more naturally and accurately.
The key to creating a successful AI assistant lies in the quality of the data it accesses. As Carl emphasises, “Garbage in, garbage out.” Ensuring that the assistant has access to up-to-date, relevant, and well-organised data is crucial. For our AI assistant, this meant integrating various company documents, including policies, procedures, training materials, and our internal guidebook, “The Method.”
A fundamental step in developing our AI assistant was curating the data it would use. This involved sifting through company documents to eliminate outdated information and resolve inconsistencies. The goal was to provide a comprehensive and reliable repository of knowledge that the assistant could draw from to answer employee questions accurately.
One critical consideration was ownership and maintenance of this data. It’s essential to have designated personnel responsible for keeping the data current and ensuring its accuracy. Without this, the assistant could end up providing outdated or incorrect information, defeating its purpose.
To demonstrate the capabilities of our AI assistant, we conducted a live test during the podcast. Carl typed in several queries to show how the assistant handles different types of questions. For instance, when asked about the necessary meetings for managing people, the assistant provided detailed insights into daily huddles, weekly meetings, and more. This example highlighted the assistant’s ability to pull relevant information and present it in a structured manner.
Another test involved asking about the steps to take after a company-supplied Tesla crashes. The assistant quickly retrieved the relevant policy document, showcasing its utility in real-world scenarios where employees need quick access to specific information.
One common question is the difference between ChatGPT and OpenAI. While ChatGPT operates within a constrained environment, OpenAI allows for more flexibility by enabling API calls. This means we can build assistants that are not confined to a single platform and can be integrated into various applications, such as our Professional Services Automation (PSA) tool, ConnectWise.
By embedding the assistant into our PSA, employees can access it directly within their workflow, enhancing productivity and reducing the need to switch between different tools. This seamless integration is a significant advantage, making the assistant a valuable asset across different departments.
The future of AI in business looks incredibly promising. As Carl predicts, AI assistants will become increasingly embedded in everyday applications, offering more comprehensive support across various functions. One exciting development is the integration of visual and voice capabilities, moving beyond text-based interactions to create a more intuitive user experience.
Imagine being able to interact with the assistant via voice commands, making it accessible even on mobile devices. This advancement would further enhance productivity by allowing employees to access information and complete tasks on the go. The potential for AI to democratise knowledge within an organisation is immense, breaking down silos and ensuring everyone has access to the information they need.
While the benefits of AI assistants are clear, it’s important to address potential concerns. One significant challenge is ensuring the quality of the content generated by AI. As the technology becomes more widespread, the risk of average or low-quality content flooding the market increases. This makes it crucial for businesses to focus on maintaining high standards for the information their AI assistants provide.
Another concern is the ongoing maintenance and upkeep of the AI system. Regular updates and testing are necessary to ensure the assistant continues to deliver accurate and useful responses. This involves not only technical updates but also continuous improvement of the underlying data set.
One of the biggest questions businesses face when implementing AI is how to measure its impact. For our AI assistant, the primary metric is the reduction in time spent on onboarding new employees. By providing instant access to relevant information, the assistant can significantly cut down the time employees spend searching for answers, leading to substantial productivity gains.
In addition to onboarding, the assistant can be valuable in other areas such as customer service, technical support, and sales. For instance, integrating the assistant with a CRM system like HubSpot could streamline sales processes by providing quick access to customer interaction histories and other relevant data. This integration could save time and improve efficiency, ultimately contributing to the company’s bottom line.
As we look to the future, the possibilities for AI in business are vast. From enhancing employee onboarding to improving customer interactions, AI assistants have the potential to transform how we work. By continuing to refine and expand our AI capabilities, we can unlock new levels of productivity and efficiency, driving our business forward in an increasingly competitive landscape.
For those considering implementing AI within their own organisations, it’s essential to start with a solid foundation of high-quality data and a clear plan for maintaining and updating the system. With these elements in place, AI can be a powerful tool for driving innovation and achieving business goals.
If you have any questions about AI applications in your business or would like to explore how to get started, feel free to reach out. We’re always happy to share our insights and help others navigate the exciting world of AI.