A Chat with Gen AI for a Stronger Chatbot
Martechvibe speaks to Rashid Khan, Chief Product Officer and Co-founder of Yellow.ai, about building the ideal conversation design tool to elevate a brand’s customer experience.
Topics
What to Read Next
- TripleLift Announces Partnership with Attain to Unify Context and Commerce
- Seedtag Partners with IRIS.TV to Expand Contextual Signals for CTV Targeting
- Magnite, Cognitiv Announce Deep Learning Integration for Real-Time Curation
- Dscout Announces Integration with HeyMarvin
- Comscore Launches Program-Level Capabilities within CCM
The global ecommerce market is expected to total $6.3 trillion in 2023. As retail continues to shift online, customers display a need for ease of purchase, which comes with speed and convenience. Marketers across the globe are using Generative AI to fast-track their sales and enhance customer support functions. However, there is a lot that goes into designing and implementing a conversational AI tool.
We spoke to Rashid Khan, Chief Product Officer and Cofounder of Yellow.ai, about the company’s Dynamic Conversation Designer to understand the application of AI in conversations and the technology that goes into building a conversational AI tool. Khan addressed the various challenges design and code teams encounter during the development process and shedded light on the impact of Generative AI on ecommerce.
Understanding conversational tools
Explaining Yellow.ai’s Dynamic Conversation Designer, Khan said that it is a conversation design tool with Generative AI that helps teams quickly and easily design chatbot conversations without writing any code. It enables the creation of development flows automatically from design flows, meaning there is no need for developers to start from scratch, which saves significant development time and effort. Khan lists the key features of the Dynamic Conversation Designer:
- Design chat and voice conversations without any coding or training.
- Visualise and interact with the designs.
- Share designs for quick feedback.
- Auto-sync of flows in real time between design and development.
Training the models
“Yellow.ai DynamicNLP allows businesses to bypass the tedious, complex, and error-prone process of model training, whereby we give a pre-trained model for common utterances for a myriad of use cases, including customer service and support,” said Khan.
With the tool, the company is reducing the need for training utterances by providing industry-specific pre-trained models that include the major intents and entities relevant to that industry. “For custom use cases, users can add and train utterances on our platform. We use Generative AI to provide utterance suggestions for bot training and testing these custom intents. We use different pipelines for generating the training and testing utterance suggestions.”
Leveraging dynamic conversations for customer service
Brands can create engaging and personalised conversations with dynamic designs, enabling them to deliver enhanced customer experiences, drive more conversions, and boost sales. Khan said it fulfils the need for enterprises to deploy an end-to-end conversational automation platform wherein they don’t need to invest in siloed tools or extensive programs to train new resources. “Leveraging it, we were recently able to deliver a full-fledged chatbot implementation in two hours, a significant improvement from the previous requirement of at least two days for design and development in silos.”
Generative AI eliminates silos
Talking about a common challenge, Khan said, currently, designers create static flowchart diagrams and hand them over to the developer for implementation. The problem with this approach is that stakeholders are not able to visualise how the conversation flows from the flowchart diagrams, and they inevitably ask for changes to the flow once the bot is built out by the developer. These iterations end up delaying the go-live date.
“With Dynamic Conversation Designer, customers get a working bot right from the design product, and all changes made by the developer in Studio reflect back in the design. We are seeing that customers are able to reach a consensus on the conversation flows much earlier, and the time to go live has been reduced by 50%.”
Overcoming barriers in dialogue flow chatbots
Conversation design is the starting point for developing conversational AI and plays a critical role in delivering superior customer experiences, said Khan. “However, the final outcome of conversation design is subpar, as it is done using flowchart-based tools, which results in clunky and complicated designs.”
This slows down the development process, as flowchart designs are impossible to visualise. Khan said that developers must recreate these complex designs within a chatbot-building platform, leading to a repetitive and tedious process. He listed key pain points that enhanced conversational design tools must address:
- Difficult-to-use design tools
- Lack of standardisation
- Repetitive development efforts
- High maintenance
Yellow.ai Dynamic Conversation Designer addresses these problems through:
- Streamlined Workflow
- Reduced Training Costs
- Standardisation
- Increased Productivity
How is Generative AI permeating marketing functions?
Modern marketers must develop a combination of technical and soft skills to keep up with the changing industry dynamics due to Generative AI. This will help generate better insights, develop more effective campaigns, and deliver more personalised experiences for their customers.
For instance, said Khan, they would need to learn how to use AI tools and platforms to generate insights, which can help them create marketing strategies. This includes knowledge of tools like predictive analytics, chatbots, and recommendation engines. “Having said that, human creativity will still be the key aspect that will help them stand apart in the market by developing compelling content and campaigns that resonate with their target audience.”
Talking about the launch of ChatGPT, Khan said that there has certainly been more awareness across the globe, with more enterprises becoming receptive to leveraging Generative AI in general. “They are open to deploying GPT 3 for multiple use cases across business functions, from customer support to content creation.” Further, he anticipated expansion of its usage across newer use cases, such as conversational marketing, that enable more natural and intuitive interactions with customers.
Generative AI’s impact on ecommerce
Generative AI is providing new ways to personalise customer experiences, optimise operations, and create innovative products. Khan listed GeneratveAI’s advantages for enabling seamless ecommerce:
- Customer service and support: Providing instant and automated responses to a wide spectrum of customer inquiries.
- Hyper-individualised marketing campaigns: We will see support and marketing interactions occur autonomously and continually improve conversions with continuously improving text variants.
- Personalisation: Analysing customer data and generating personalised product recommendations, offers, and custom product designs based on individual preferences.
- Concepts for product design: By analysing customer feedback and market trends it can help generate new product concepts and designs.
Generative AI has a bright potential in ecommerce. Besides personalisation, content like product descriptions, social media posts, and blog articles can use innovation with the help of AI. “Marketers can input parameters, such as tone, style, and topic, and the AI will generate content that meets those requirements,” concluded Khan.
