Our product specialist, Pragadeesh Jayachandran, and product manager, Fiaz Ahamed, sat down for a discussion, where they examined how businesses can embrace automation and AI for customer experience. They also talked about the automation and AI trends, AI for business, and SalesIQ's AI roadmap for 2023. Here are some excerpts from the session:
Pragadeesh: As a PM who has to stay on top of things, what's your interest in automation and Al?
Fiaz Ahamed: The demand for automation has always been there. As far back as we can see in the past, persistent and growing demand for the mechanization of human labor has driven automation. From primitive calculators to modern aircrafts operated on autopilot, automation like Robotic Process Automation (RPA) has proliferated in every industry. It has become a necessary component of modern businesses.
AI, on the other hand, I owe to the exponential developments in the field over the last decade or so. The concept of formal reasoning and emulating human thought dates back to the 18th century, yet the progress made in AI in the last decade has been unprecedented. The widespread adoption of cloud computing, smartphones, and seamless internet access has made data readily accessible, which serves as the cornerstone of AI's growth. AI has become a buzzword and a topic of constant discussion, with new developments emerging every month, if not every week. Just look at the past few weeks: ChatGPT, BARD, and now Einstein GPT. Observing the endless potential of AI for business and the exciting possibilities it presents for the future is fascinating.
Pragadeesh: As you mentioned, a lot is happening in this space. What are the current AI for business trends that you're excited about?
Fiaz Ahamed: Undoubtedly, generative AI. It will be the focal point of attention for the next two to three years because so many generative models are released in a very short time. There was Open AI's DALL E, then Nvidia's GAN, Stable Diffusion, DALL E 2, and now three more.
A decade ago, the trend was to compete with these models by preparing datasets and building deep learning models from scratch. However, today's priorities have shifted, and no one wants to reinvent the wheel anymore. Companies are now looking to use these models to create custom end-user features and applications. Generative models, in particular, are expected to play a major role in this trend. They can automate tasks like content writing, text processing, programming, and debugging.
This trend has already begun to take shape, as evidenced by the success of features like Lensa's magic avatars. This feature leverages stable diffusion to generate animated images and has already garnered a substantial customer base in the social media industry, despite its recent release in November 2022. This demonstrates the ease with which companies can leverage these models and benefit from the advancements in generative AI.
In conclusion, we expect to see a continuation of this trend, with more and more SaaS companies benefiting from these developments in the future.
Pragadeesh: We're going to discuss AI for business here, but before we do, let me take a detour to ask a product-related question, as I want to clear the air for our customers. Yes, the automation and AI space is getting interesting. There are a lot of new things coming in. But we already have automation and AI features that cater to three major departments in an organization. How can our SalesIQ users utilize them to their fullest potential?
Fiaz Ahamed: If you examine the three departments that benefit from SalesIQ, namely marketing, sales, and customer support, you will see that they are the core of any business, regardless of size. A customer engagement tool has become essential for any business to promote and sell its product and provide support throughout the customer journey.
However, it's not just any customer engagement tool that is needed, but one that is adaptable across multiple platforms, including Facebook, WhatsApp, and Instagram. Customers require real-time support around the clock, and they don't want to wait for days or hours for an email response. As a result, a ubiquitous, real-time customer engagement solution is required.
SalesIQ can provide this. However, to maximize its value, especially for small and medium-sized businesses, it should be seen as a miniature CRM in addition to live chat and web tracking software. Our roadmap is designed to achieve this, allowing you to see where your prospects are coming from, track their behavior, interact with them, keep track of leads and contacts, drive a sales process, and provide support. To sum it up, SalesIQ is a 360-degree product.
Regarding features, in addition to live chat and visitor tracking, features such as the SDK, bots, and IM channel integrations have recently been more beneficial to our customers.
Pragadeesh: Now we obviously have to talk about the elephant in the room: ChatGPT. When new technology arrives, our expectations can make it impossible to see things objectively. The same can be said about ChatGPT. Within two months, the platform garnered more than 100 million users. How do you see this new technology? As a product manager, how do you envision using the AI for business and customer experience?
Fiaz Ahamed: Before presenting my vision of where ChatGPT will take us, I must acknowledge the remarkable contribution that OpenAI has made to the world. ChatGPT is trained on gpt-3, the 175 billion parameter language model that was released sometime in 2020. There was a lot of talk about it back then, but the real applications were never fully explored. This was mainly because, while it was the biggest language model at that time in terms of parameters and was good at answering questions, it was far from perfect. What we are seeing now is that beyond the data and parameters is the effort that went behind manually refining the model. OpenAI manually labeled the outputs where people say, "when this is the question, this should be the answer," and then ranked the outputs, which in turn improved the performance using reinforcement learning.
It answers questions, chitchats, programs, and empathizes. Not just that, I read that ChatGPT has also been trained in formal reasoning through codes, which is a huge effort on the supervised learning part. That's the reason we are discussing what we see now. A big kudos to them for this milestone!
Similarly, it writes essays and poems. But does it guarantee against plagiarism? After all, it is trained with historical data that was originally someone else's creation. So we cannot rule out the possibilities of plagiarism and redundancy.
So how should it be used? At this stage, ChatGPT should be thoughtfully and strategically placed inside products to augment human abilities and not completely replace them. That's all that should be the focus at this stage of AI evolution. Because at the end of the day, as the famous quote goes, you must learn to walk before you can run. AI is crawling right now. So the focus should be on building features using crawling AI, not the walking or running one.
Pragadeesh: Love that analogy, Fiaz. Well, the crawling AI is making the world go round. What's going to happen when it starts walking and running? Here's the topic all our users are waiting for. ChatGPT for business. What can our SalesIQ users expect from it? What problems can it solve for customers?
Fiaz Ahamed: We have been examining the potential and domains where ChatGPT can extend its assistance to SalesIQ users. Undoubtedly, the conversation module is one of the most apparent areas of application. In this context, our team has been exploring the possibilities of incorporating quick replies that encompass far more capabilities than the existing smart suggestions feature. Another salient attribute that we are looking to build is a fact checker. This feature will ensure that your operators do not give customers incorrect answers, enhancing customer service quality. Auto-tagging, and auto-complete—again, these are some of the features in the POC stages now.
Pragadeesh: It looks like customer support teams are excited to use ChatGPT, as AI for business has the potential to reduce workloads, save time, and impact customer experience. On that note, how will customer service change with ChatGPT and AI?
Fiaz Ahamed: Customer support is poised to undergo a significant transformation. Chatbots have already demonstrated their efficacy as a game-changing innovation in this arena. Our ambition is to elevate this technology to new heights by offering support for multiple languages and implementing reinforcement learning to support decision-making processes. The answer bot's release initially aimed to filter out repetitive inquiries. While it has performed well in this regard, we anticipate more exceptional performance and a broader scope soon.
Our intentions do not end with bots; we plan to integrate machine learning into the routing process and chat flow prediction. By scrutinizing patterns, these features aim to ensure that SalesIQ customers maximize their support team's potential and do not miss any opportunities. Thus, customer support is set to be one of our primary targets for our upcoming AI features.
Pragadeesh: I hope all the support executives attending this session are taking notes. Well, now we've come to the end of the session, and this is my last question. You answered this briefly, but I would like to know if we can expect more in this zone. Where is AI in SalesIQ's roadmap?
Fiaz Ahamed: I think I have given out enough already. But if we talk about the roadmap, one rule we follow whenever we come up with any feature or select which feature to take up is that we don't do something just because we can. As customer satisfaction is our primary objective, we prioritize their requirements above everything else. In the past, we have received numerous feature requests that were not possible to implement at the time, but with the emergence of advanced AI models, we can now explore these possibilities.
So, when ChatGPT came out, our initial plan was to go through our customer requirements in the past and see which of those problem statements could be addressed with this technology. Additionally, we consider the cost, time, and effort that customers will have to put in to use each feature. That has been one of the main components in our feature release checklist. So, considering that, I'm looking forward to what BARD AI can do. Being open-source aside, the amount of real-time data that Google has access to is day and night compared to others. We could have more options now that Microsoft is incorporating ChatGPT into the Bing browser.
Some of the areas where we plan to use AI:
• Knowledge base
• Triggers and routing
• Lead identification and conversion
You can expect an announcement from SalesIQ regarding all this soon.
If you'd like to learn more, schedule an exclusive demo now and check out what Zoho SalesIQ can do for your organization.