Dinesh Sharma is the Co-Founder and CTO of AskSid.ai, a company that develops a conversational AI solution that helps enterprises scale their sales and customer support operations.
In this interview, Mr. Sharma shares insights on how conversational AI can increase sales by bridging the product knowledge gap between the customer and the enterprise.
This interview has been featured in the Conversational AI Initiative 2020.
What is your background and how was AskSid started? What is AskSid‘s mission?
AskSid.ai came into certainty from the real-life experience of Dolly wife of Sanjoy Roy, CEO & Co-Founder of AskSid.ai. A few years back, Dolly, an expectant mother was starting a job in the Netherlands and was looking for a pair of warm leggings. While logging into one of the famous webshops, she was overwhelmed with over 400 options of warm leggings to choose from, but she had questions about the products that require product expertise e.g. “Is this safe to wear during pregnancy?”, “Will it hide my veins & scars?” etc. She couldn’t get the answers from the webshop or customer support, confused, she decided to go to the same brand’s store and get a sales associate to help her make a decision. Finally, she purchased a pair of leggings.
Which led to Sanjoy thinking. He saw that when there is a knowledge gap between the buyer and the product, the shoppers tend to ask questions, which she doesn’t get easily on webshops. While the brands that make the products have answers to all possible questions.
Knowing that AI has got the answers to this precise problem led Sanjoy, who was heading the Mindtree’s digital business in APAC with 21+ years of experience in technology business, start AskSid.ai along with his colleague Dinesh Sharma, – CTO & Co-Founder of AskSid.ai having 21+ years of global experience in architecting and deploying enterprise-scale products and solutions for Retail, Consumer Goods, Finance, and Travel domains.
AskSid’s mission is “to organize and enrich product knowledge in a way that makes shopping easy and convenient. Inspiring every consumer to buy”
Walk us through AskSid’s products, and what problems they solve.
AskSid.ai is a Full Stack conversational vertical AI solution for Retail & Consumer good brands, where it owns the experience layer (chatbot), the underlying AI/ML models that power the chatbot, and the domain-specific product data that powers these proprietary AI/MLmodels. The solutions consist of 6 modular components
- Cognitive workbench to ingest the brand’s product & catalog data to generate and associate product attributes, Question and Answer pairs, and their utterances
- Shopping assistant bot that can handle multi-turn conversations in multiple languages across web, Facebook, skype, slack, Whatsapp etc.
- Live chat hand-off to escalate the conversations to human agents for complex questions and based on configurable trigger moments
- Replay conversations helping agents to know the context of the customer beforehand for quicker resolution
- Vertical AI for Retail that comes preloaded with retail-specific intents, named entity for retail with its retail domain ontology making any brand onboarding possible within 4 to 6 weeks while delivering multi-turn conversations
- Insight & Analytics delivers actionable insights on customer preferences & demand signals with a real-time KPI dashboards
Even if we call the call center of our favorite brand, most often the agent is clueless on product questions. On the other hand, the brand who is making and selling the product knows everything about the product and therefore has all the answers. Now, can technology be used to bridge this gap?
Whenever we shop something online we all get questions about the product but we do not get the answers easily. All we have in the online catalog are some images of the product and a 3-4 lines of the product description, which is often inadequate for us to decide BUY vs NO BUY. Even if we call the call center of our favorite brand, most often the agent is clueless on product questions. On the other hand, the brand who is making and selling the product knows everything about the product and therefore has all the answers. Now, can technology be used to bridge this gap?
The validation of the above problem lies in a recent Forrester report which states that “53% of customers will abandon an online purchase if they can’t find a quick answer to their questions”. Added to that, Amazon too has taken note of this consumer need and recently on its product pages, Amazon has started the new “Customer Q&A” feature – where a customer can log product questions and then Amazon helps in crowdsourcing the answer from other customers or sellers.
What are the biggest challenges that you are currently facing at AskSid?
The conversational AI market is too crowded at the moment, prospects are overwhelmed with the number of options available and lost in figuring which is the right solution for their enterprise and use case. The biggest challenge at the moment is getting the enterprises to understand the difference between a Full-stack Vertical AI vs General-Purpose conversational AI.
How do you measure the performance of your products?
The performance or the value of our solution is measured on 3 broad parameters that we promise to deliver.
- Accelerated Conversions by simplifying the shopper’s journey. For example, to our customer a luxury fashion brand in Europe, we have delivered an influenced order rate of 11% compared to their website conversion of 1.5%.
- Enriched product catalog by discovering new questions from shoppers. For example, for a fortune 500 paints company, we have discovered 10,000+ product questions in a period of 6 months.
- Instant Cost savings and ROI by automating customer service queries
Tell us about the AskSid team.
AskSid team, we call it Sidlings, is 12 members strong. The team is a tight-knit, talented and passionate group with deep expertise in AI, ML & DL technologies with a shared vision of creating a world-class conversational AI solution. All of our Sidlings are unique individuals who are united by a set of core values that apply to everything we do.
What is your experience with voice-enabled interfaces?
We have implemented a few very interesting use cases in our Lab but yet to hit the market with an offering. We are working with our customers to find use cases where the voice-enabled interface can help their buyers.
Voice is expected to become an interaction medium of choice for the consumers, which might result in the search behaviour of consumers changing in the way they discover the brand’s product, the queries they will have on the product before they buy or post-sales when using the products.
What opportunities do voice-enabled interfaces bring for enterprises?
Right now, with just a handful of simple scenarios, it’s evident that voice assistants are influencing our connected world. With the launch of Amazon Transcribe & Google making its assistance open to developers will help in creations of multiple use-cases augmented with artificial intelligence, enabling major advancements in the years to come. With this, it is up to solutions providers and brands to leverage voice to solve a specific problem area and provide value to their customers. Voice is expected to become an interaction medium of choice for the consumers, which might result in the search behaviour of consumers changing in the way they discover the brand’s product, the queries they will have on the product before they buy or post-sales when using the products.
How can enterprises get started with their conversational ai strategy?
My suggestion would be to start with a pilot for an identified pain area with well-defined outcomes and scale at the shortest possible timelines. In today’s world getting started with a conversational AI has become quite easy with numerous offerings from platforms available to build self-service solutions to specialists solutions in solving a very specific use case. It is important to fail-fast in order to find the important business area, where enterprises can derive maximum value from Conversational interface and interactions.
And choosing the right solution is imperative to the success of the strategy, and some of the must-have evaluation criteria are
- Whether the solutions is a full-stack one? Providing the flexibility to handle the experience, data ingestion to the measurement metrics.
- Whether the shortlisted use case warrants an industry trained AI/ML models or general-purpose ones?
- How deep is its intent library? Broader and deeper pre-built intent library means handling multi-turn conversations that will enhance customer experience
- How much of data & training on the data is required for the solution to start delivering the results?
Answers to the above considerations can lead the enterprises to a well-evaluated start and path to a successful pilot.
The bots might seem like a tool that belongs only in an online world, but don’t be tricked, they are transforming into an indispensable asset in physical stores as well.
How can enterprises prepare for the changes in the industry brought by Conversational AI technology?
Enterprise can no longer shrug off conversational AI technology as a mere cost-saving tool meant to automate the repeatable processes.
The bots might seem like a tool that belongs only in an online world, but don’t be tricked, they are transforming into an indispensable asset in physical stores as well. Imagine conversational AI in stores providing instant help to a shopper: this can enhance the brand experience and elevate a shopper’s outlook toward the brand.
Gartner in its recent report says that 25% of global brands are going to integrate virtual customer assistant technology by 2020, up from 2% in 2017. Hence to outpace the competition, enterprises need to acknowledge the advancements made in conversational AI technology.
- Acknowledge & Identify business use cases that generate value to the shoppers, experiment with the use case with a pilot in a short time frame of 4 to 6 weeks.
- Rapid rollouts for a complex use-case would be possible only with the right solution provider who has pre-trained models and proprietary data that can be adapted to an enterprise’s unique use case.
- And not the least, having the right conversational metrics in place to measure the success of the pilot before rolling out the solution across channels & categories.
What are the biggest opportunities for enterprises to leverage Conversational AI?
The 3 biggest opportunities that I have seen in enterprises who had successful conversational AI deployments are:
- conversational AI assistants have provided timely, accurate and tailored experiences that resulted in accelerating the conversions that are much higher compared to their traditional webshops.
- the enterprise has achieved a better understanding of its customers from the insights discovered from customer’s free typed messages and queries, which often leads to a multi-faceted and cross-functional impact on marketing, new product innovation, and promotions.
- And finally, the cost savings that are achieved by the automation of repeatable questions which doesn’t need the valuable time of a customer support agent.
How can enterprises leverage Conversational AI to make their customers happier and more satisfied?
Enterprise can leverage vertical focused Conversational AI across their digital mediums or in their stores to excite their shoppers and remove the friction points.
The enterprises can automate and resolve the customer service problems and direct the more complex issues to your customer support agent. For one of our customers in Europe, we have delivered automation equivalent to adding 8 extra agents at 20% the cost.
The data collected by the conversational AI solution can then be used to create personalized communication with customers, improving customer relations, and the overall service experience.
It is proved that Conversational AI can increase engagement with its personalized customer service capabilities – something that even the best videos and websites simply cannot do.
Conversational AI will need to move towards Vertical AI solutions as compared to horizontal/general AI, which can be trained and adapted to the enterprise’s business very quickly.
Which Conversational AI-related technology trend do you think will have the biggest impact in your industry in the coming years?
Vertical AI is all about solving specific business problems. It is important for the AI that powers the experience is able to understand the brand’s business. An AI that understands brand products, business, offerings can drive effective conversions and deliver a better experience.
In my view, conversational AI will need to move towards Vertical AI solutions as compared to horizontal/general AI, which can be trained and adapted to the enterprise’s business very quickly.
The solutions will need to enable enterprises to test multiple use cases, fail fast and fail-safe to arrive at the business problem that is best solved through Conversation AI.
Also, I believe that Conversational AI will become the NEW APP which will deliver dynamic, on-demand, and 1:1 experience to consumers. This APP will not be restricted by 3-click or Menu or filters driven experience But ‘serving the need by understanding the question and delivering the precise response’. I believe the future of information/knowledge dissemination and interactions will change drastically from static to dynamic conversation. The future of interactions belongs to Conversational APPS through text, voice, and image exchanges.
Editor Update, April, 2020: We reached out to Dinesh for a comment on the trends of Conversational AI, the Q&A follows.
What have been the most relevant developments, breakthroughs and advancements in Conversational AI in your industry in 2019, and what do you foresee for 2020?
The way I see it, first wave of Conversational AI was to enable businesses to interact with their customers on Social Channels. In 2019, the businesses started finding the value of Conversational AI on their Websites. Every second website that we visit today has a virtual assistant with varied degree of support. The websites that had Live Chat capability got Conversational AI to handle most the frequently asked questions and transferring the control to customer service agents for potential leads.
Conversation AI has moved from cost saving solution to revenue generation medium. Recent developments and contributions from Google, Microsoft, and Facebook in Natural Language understanding, the Conversation AI has become more contextual, more understanding, and able to support multi-turn conversations. It has progressed significantly from simple drag-drop to deeper vertical conversational AI solutions.
In 2020, Conversational AI will move faster towards domain/vertical specialization where
1) channels: a) it will continue to listen to users (text, voice, image) on digital channels like website, mobile, and social channels b) it will move forward from digital channels to physical stores helping walk-in consumers and staff alike
2) understand users need accurately by applying Conversation AI’s domain knowledge specialised to the specifics of business’s offering
3) generate the best and seamless responses
4) Moving forward from a siloed Conversational AI to complete solution that will source or deliver information and insights to appropriate enterprise systems
I believe we will see transformational business cases realized through Conversation AI in Y2020 moving forward from basic FAQ / Survey type simpler ones.
