Code meets imagination: Using generative AI to build Intelligent Products

Two approaches to working with generative AI

There are two main approaches when deploying generative AI within a business setting across a spectrum of opportunity domains. Much like you wouldn’t use a screwdriver to drive a nail into a board, each method has its place depending on the task or goal you’re looking to achieve.

The basic approach: Using off-the-shelf generative AI products

The most common way that people are dipping a toe in the generative AI waters is to use a pre-packaged, non-customizable, out-of-the-box product like ChatGPT, DALL-E, Bard, GitHub CoPilot, or Amazon CodeWhisperer as a work accelerator to help complete tasks.

This basic application of generative AI satisfies easy use cases with the direct, simple use of these tools. An organization doesn’t need to build its own large language model and has a low barrier to entry. All one needs to do is create a prompt to generate content to their specifications rapidly and easily.

These tools act as an assistant for human workflows and are not sophisticated enough for automation. They create models and algorithms to generate outputs (images/text/code/video) and work by learning patterns, features, and relationships within a given dataset.

Examples of the basic approaches to using generative AI include:

• Solving small-scale problems
• Conducting research such as unlocking advanced “search engine cases”
• Accelerating the coding process with automated software engineering assistants
• Summarizing content
• Assisting with creative writing, video production, and other artistic use cases

For this more basic use, generative AI can be implemented either on an individual or department-wide level, without interconnection among people and/or systems. This differentiation leads us to now explore the more advanced method of using generative AI in a way that can be leveraged by and provide support to the entire organization.

The advanced approach: Building Intelligent Products with generative AI

There’s using generative AI, and then there’s building with generative AI.

Out-of-the-box generative AI products may not be for everyone. Some initiatives aren’t well-suited, need domain adaptation, or may require more security. Building more customized Intelligent Products using generative AI is an option should you need additional tooling or want to go further.

Advanced implementations enable the application of generative AI-based Intelligent Products throughout the entire organization. Building Intelligent Products with generative AI helps you create customized solutions to integrate with applications, disparate inputs, and enterprise platforms such as customer relationship management or knowledge management systems.

Advanced AI solutions can work with inputted information, allowing generative AI algorithms to provide outputs and then operationalize the results, including classification, prediction, and personalization.

You can pull in internal and external data to augment your existing data sets or connect these systems to various outputs for enterprise-wide knowledge discovery and a single source of truth. Employees can surface and access information faster and benefit from automation with these intelligent products rapidly writing accurate drafts and materials that form the basis for marketing campaigns, personalized sales initiatives, contracts, and more.

Generative AI-based Intelligent Products can be extended to serve even more advanced use cases, such as providing next-gen chatbot experiences, enabling AI agents to perform actions, helping with automated custom software engineering, and much more.

Use cases for Intelligent Products built with generative AI

Generative AI-based Intelligent Products can be extended to serve even more advanced use cases, including tailored, highly personalized customer interactions and sophisticated operational improvements such as:

Smarter, more specific menus. Instead of providing generic navigation menus for a customer to sift through until they find a close enough match to meet their needs on your website or app, it’s now possible to bring greater sophistication and precision to this process. Site visitors can use a prompt to discover what they’re looking for and lead the conversational interface to hone in on their exact wishes. Going further, the Generative AI-based Intelligent Product can draw on past interactions and information stored in your systems to deliver an even more powerful, faster, and satisfying customer experience.

Imagine being able to get so granular as to ask the product to satisfy a request to help a family plan, budget for, and book a month-long vacation in a warm, but not too warm location, that includes travel, a hotel room, and a full itinerary with activities suitable to keep both a toddler and seven-year-old entertained. Oh, and add a few options for vetted childcare to that request.

Something that would take an agent considerable time and effort to accomplish can now be generated within seconds, delivering an exceptional customer experience.

Call center transcription + summaries. Companies are starting to use generative AI to transcribe and summarize conversations with call center agents. Even the most garbled, low-quality calls can be easily understood. This tedious, time-consuming, error-prone task for humans is handled with greater efficiency and dramatically accelerated with an Intelligent Product built using generative AI.

The technology makes short work of creating summaries of calls that agents can use to understand a customer’s entire purchase and support history. Immediate access to this information helps customer service staff provide top-notch care, leading to increased shopper loyalty. An Intelligent Product can process call summaries to generate instant recommendations to help with coaching and professional development.

A generative AI-based Intelligent Product can also make light work of scouring transcripts and summaries to identify customer sentiment. With greater accuracy, AI can process voice-of-the-customer (VoC) data with greater precision than humans to unearth vital trends, leading to improved customer experience, and increased revenue and operational agility.

The road to building an Intelligent Product using generative AI doesn’t need to be overwhelming. Start with a small test case, such as adding chatbot experiences to digital products or implementing features involving advanced actions such as personalization, classification, and prediction, and then explore additional areas where it can be applied.

If you’re looking for extra help navigating generative AI, wish to go further to build an Intelligent Product with generative AI capabilities, or want to evolve your existing Intelligent Product, we’re here to help. We have extensive knowledge of artificial intelligence and machine learning with experience crafting Intelligent Products for leading global companies.

Book a call to plan a workshop with our team of experts to gain an understanding of how an Intelligent Product built using generative AI can deliver positive business impacts.

[1] CIO Dive, Generative AI worth a shot despite risks, execs say, May 2023