Harnessing the Power of Gen AI: High-Impact Product Strategies for Forward-Thinking Companies

Published by: 1) Eric Xu, Sr. Director of AI Engineering @ Hubspot. Previously AI Engineer @ Meta & Director of ML @Reddit. 2) Matt Jaffe, AI Product Leader @ Upwork. Previously Product Leader @ Dropbox & Microsoft and CEO @ Applause. 3) Yasi Baiani, Co-Founder & Chief Product Officer @ Raya Advisory. Previously Product Leader at Fitbit, Teladoc, and Cleo and CEO & Founder @ ActivePepper.

In today's digital age, rapid advancements of transformative technologies like Generative Artificial Intelligence (gen AI) are reshaping the way businesses operate, innovate, and engage with customers. How end-users interact with products and their expectations are rapidly evolving. Per PwC research, artificial intelligence (AI) is projected to contribute $15.7 trillion to the global economy by 2030. Businesses have immense potential to utilize gen AI to meet customers where they are, solve their problems, drive innovation, and enhance efficiency.?

In this article, we'll explore some of the most high-impact yet practical ways businesses can leverage gen AI to solve customer problems, unlock new opportunities, and achieve sustainable growth. We believe any company can begin leveraging gen AI in one or more of the following ways as you seek to bring the benefits of gen AI to your customers and your organization:??

1. Conversational Assistants: Conversational AI technologies such as chatbots and virtual assistants are reshaping how companies interact with customers, employees, and stakeholders. By deeply understanding the intent of natural language and being able to respond in a similar manner, LLMs (large language models) can power chatbots, virtual assistants, and voice-enabled interfaces that enable businesses to substantially reduce the friction that may be involved in their products while automating customer service inquiries and providing personalized recommendations, advice, or feedback.

One such example is Klarna’s customer support assistant, built on top of OpenAI’s ChatGPT. Within its first month, Klarna’s chatbot was able to carry out nearly two-thirds of the company’s customer service chats, numbering at 2.3M, equivalent to nearly 700 agents. Customer satisfaction scores were in line with those of human agents but accuracy was higher, with a 25% decrease in repeat inquiries. Moreover, customers were able to resolve their questions in an average of 2 minutes, compared with 11 previously, in over 35 languages. All told the new assistant is expected to drive $40M in profit during the course of 2024.

It’s important to explore if your business can take advantage of implementing a virtual assistant to more effectively and efficiently interact with your customers, answer their questions, or address their issues. Be sure to pay particular attention, though, to response quality and accuracy by putting in place a robust system to avoid hallucinations – like promising customers a refund based on a non-existent policy!?

2. Vertical-Specific Virtual Agents: These conversational, AI-driven digital entities typically leverage a combination of advanced technologies, including natural language processing (NLP), machine learning (ML), deep learning, and domain-specific knowledge bases or ontologies, to understand user inputs, process information, and generate contextually relevant responses. They are trained and tailored to understand the nuances, terminology, and workflows relevant to their respective domains, enabling them to engage in natural language conversations, provide personalized assistance, and perform tasks specific to the industry they serve.?

Some areas that could most naturally benefit from vertical-specific virtual agents are healthcare, financial services, HR, and retail.?

One of the most innovative products and companies we’ve seen in this category is Hippocratic AI, which offers the safest artificial Health General Intelligence (HGI) that dramatically increases healthcare access and the scalability of healthcare professionals. While Hippocratic’s solution is based on an LLM they trained themselves – a costly endeavor – there are cheaper ways to customize LLM output, including through prompt engineering, fine-tuning, and retrieval augmented generation (RAG). OpenAI’s GPTs feature is a simple approach to such customization that practically anyone can undertake by themselves.

3. Personalized Customer Experiences: Gen AI enables the creation of highly personalized product experiences. By leveraging data, advanced algorithms, and deep learning techniques, companies can deliver hyper-personalized marketing messages, product recommendations, tailored content, and customer service interactions that resonate with each customer on a deeper level, driving engagement, loyalty, and lifetime value. These types of technologies have been implemented in many products such as TikTok, Netflix, and Spotify for years; now with gen AI the power of these types of products and experiences will be significantly amplified.?

If you haven’t already, it’s important to think about how to utilize your proprietary data and plug it into LLMs such as OpenAI ChatGPT or open-source foundational models to provide more personalized experiences for your customers. Spotify’s AI Playlist feature and Amazon Music’s new Maestro playlist maker are both recent examples of enabling customers to leverage gen AI to drive deep personalization.?

4. New Applications: Do you remember when the App Store started in 2008 and many companies were rushing to build a mobile application that was accessible via a smartphone? Most companies were trying to figure out how to create a mobile application and expand beyond a web surface.?

Today, we are in a similar place with AI. Companies such as Duolingo, Instacart, HubSpot, and Upwork, have already built new applications on top of existing Large Language Models (LLMs) such as GPT-4. To stay relative and competitive, other companies are racing to find ways to utilize gen AI and bring useful applications to their customers.??

Building applications on top of existing models – versus training your own model - is generally a faster and cheaper path to market. This will give your company an opportunity to build something and put it in front of customers for feedback and iteration while evaluating if and how to build and train proprietary models -- if that path makes sense for your company.

5. Workflow Automation: Many companies have ample opportunities to take advantage of workflow automation. This is true for all types of businesses but particularly relevant for those with heavy manual processes or those that require human involvement as part of their offering (e.g., healthcare, coaching businesses). Automating repetitive and manual tasks that are the core part of your offering could be one of the best ways to utilize gen AI to empower your customers.?

As an example, Upwork recently introduced a gen AI-based feature to facilitate the creation of new job posts by clients. Job posts are a fundamental component to hiring freelancers and are often time-consuming for clients to create. To address this, the company introduced an LLM-powered job post generator, capable of drafting an entire job post in minutes. Clients are now using this feature to post higher-quality jobs nearly 70% faster than before.

6. Operational Efficiencies: There are a substantial number of ways in which gen AI tools can be leveraged to drive operational efficiencies across different functions in your organization. From marketing (Jasper) to coding (GitHub Copilot) to UX design (UIzard), there’s likely a tool available to enhance the productivity of practically every function in your company. Given the pace of innovation in this space, our advice is to conduct internal pilots to identify the tools that work best for your company – just don’t forget to ensure you adhere to any infosec policies that might be in place.

Beyond function-specific tools, your organization may also have opportunities to leverage LLMs to automate internal processes, including those that were previously the sole domain of humans. One of our portfolio investments, a delivery company focused on facilitating online shopping returns, is currently leveraging the vision capabilities of GPT-4V to identify if packages have been delivered to the right address – and doing so at a substantially lower cost than their prior human-based solution.????

We are excited about the impact that AI is poised to have on everything we all do. We believe it will fundamentally change every product and function over time. As such, it’s important to get started in planning your AI strategy and implementing AI in your product in a way that makes the most sense for your business. This will give you a competitive edge while making your business more efficient and cost-effective.???

As you embark on your journey to explore ways to implement AI, make sure to utilize experts in the field to help prevent common mistakes and save time and money. Raya Advisory is focused on helping companies transform themselves by taking advantage of AI in their product and business. We look forward to hearing from you and supporting you in your AI journey.

Feel free to reach out to us at [email protected] for any questions or if you see a need in your organization for product and AI strategy.