Sonakshi Nathani, co-founder of AI tech company Bik.ai shares insights on the future of the company while speaking about how GenAI-based implementations are helping e-commerce evolve.
New Delhi: Sonakshi Nathani, an entrepreneur at the forefront of AI in e-commerce, is the co-founder and CEO of Manifest AI, a shopping assistant used by over 800 brands. With a BTech in Computer Science from the International Institute of Information Technology Hyderabad (IIITH) and experience as a software developer at Microsoft, Nathani focuses on building AI solutions to enhance e-commerce efficiency through her other company Bik.AI, which she founded in 2019.
Bik.AI was developed to fill gaps in existing marketing tools that were not designed for e-commerce. Understanding the challenges of lean D2C teams operating under tight margins, the platform provides a unified, cost-effective solution that integrates chat support and marketing.
Starting with WhatsApp and later expanding to Instagram, Bik.AI evolved into a comprehensive platform that now includes email, SMS, and web push notifications. Driven by its integrated retention and retargeting strategies, brands typically see a return on investment (ROI) of 23-25 times.
The company achieved impressive growth, scaling from $0 to $2.5 million in revenue since inception. About 50-60% of its revenue comes from markets outside India. However, last year, the company generated $1.8 million in revenue from the Indian market.
“We are aiming for 5 to 7 times growth in the next 5-7 months and are targeting a revenue of $10 million by December 2025. This is a significant goal for us, and we believe that with the right strategies in place, we can achieve it as we scale our customer base and expand globally,” added Nathani.
In an interaction with IndiaRetailing, Nathani shares insights on the company’s future and how the e-commerce story in India is booming with more and more GenAI-based implementation.s
Edited Excerpts
What notable brand successes have you observed with your solutions?
In 1.5 years, we’ve partnered with over 300 brands, yielding high ROI for them through our comprehensive approach in conversational and traditional marketing channels.
Can you share Bik.AI’s growth journey?
Starting from a vision of unified, e-commerce-centric marketing, Bik.AI has captured 20% of the market within three years. This growth was driven by strategic expansions across platforms and user-friendly, high-impact solutions that don’t require extensive training or resources to deploy.
How is your customer base evolving?
India’s consumption landscape is growing rapidly, with a stronger emphasis on digital engagement. Our customer base reflects this trend, as brands increasingly prioritize conversational AI to meet evolving consumer expectations and enhance their online experiences.
How does your AI solution contribute to customer experience and personalization for brands?
Manifest.AI and Bik.AI support 600 brands, maintaining a 98% CSAT (customer satisfaction) score by providing accurate, personalized answers in areas like healthcare and skincare. For instance, users receive tailored product recommendations based on their BMI (Body Mass Index) or skin issues. Immediate, natural-language interactions in English or Hindi improve customer engagement, boosting conversion rates by 50-60% from day one.
How do you train your Large Language Modules (LLMs) to handle diverse and complex customer queries in e-commerce?
We build on OpenAI’s models, enhancing them with proprietary algorithms for e-commerce. For smaller stores (annual revenue under $100 million), we use pre-set algorithms to deliver real-time responses for 18 common e-commerce intents. These stores can launch within 3-5 days using automated data extraction from platforms like Shopify or API integrations for others. For larger, complex operations, we fine-tune models to meet specific brand needs for greater accuracy and scalability.
How do you ensure data security and customer privacy for retail brands and their customers?
Data protection is critical for establishing trust. We implement stringent security protocols, including advanced encryption methods to keep data safe. User interactions are anonymized, ensuring that personal queries remain confidential. We comply with major data protection regulations such as GDPR (General Data Protection Regulation) to reassure brands and their customers that their data is securely handled and privacy is maintained.
Do you see data security as a significant challenge for your company?
Not really. These challenges have been largely resolved. One advantage is that enterprises were among the first to adopt generative AI, which made compliance easier. To ensure data security, we either use data as reference points without storing it or, for larger clients like Decathlon, we sign non-disclosure agreements and process data on their servers or cloud infrastructure to maintain full control and privacy.
Does an interconnected flow of questions help you in training LLMs?
Yes, interconnected question flows are key for engagement and enhancing customer experience. For example, if a user asks, “Does this shampoo work for dry scalp?” the AI can respond and anticipate follow-ups like, “Will it prevent dandruff?” This predictive approach creates a seamless experience, guiding users naturally through their decision-making. LLMs trained to recognize and build on the context of initial queries make this dynamic engagement possible, enriching personalization and maintaining user interaction.
Is there human intervention involved in managing the AI responses?
Yes, human intervention is crucial, especially during the initial deployment phase. For the first month, teams actively monitor AI interactions using conversation heat maps to assess how users engage with the system. This helps identify issues like unsatisfactory responses or unexpected user behaviour, allowing for prompt adjustments and fine-tuning through a dedicated dashboard.
As the AI gains experience and learns from user data, human involvement decreases, transitioning to ‘auto mode’ where only significant updates (like new product additions) require input. Performance metrics, such as conversion rates and CSAT scores, are continually monitored to ensure effectiveness, with human oversight reintroduced if anomalies arise.
Are there any plans to incorporate voice-based searches or image recognition features in the future?
Current efforts prioritise enhancing AI accuracy, reducing response times to 1.5 seconds to 2 seconds, and broadening capabilities to manage complex queries like competitive pricing and contextual follow-ups. While voice-based searches and image recognition are promising future features, our main focus is on refining the models and optimizing their performance. We aim to make the shopping experience more interactive, transitioning from static storefronts to dynamic AI-drivequestionson and answers that directly support purchase decisions. Voice-based functionality will be revisited as LLM adoption matures.
How does the company address legacy business owners’ reluctance to adopt new technologies like LLM?
We use real customer stories and real-world examples to help them understand the benefits. When we present data on how these technologies have helped other businesses succeed, it often helps bridge the gap for older founders who may be more hesitant. Their focus is usually on results, so when we demonstrate the potential impact through case studies, they are more likely to be open to adopting new solutions.
Do you regularly engage with technology leaders to gather insights?
We gather insights from three main sources: The perspectives of brand owners or chief technology officers on how they see LLMs evolving, the feedback from customers using these technologies to improve their shopping experience, and data from brands already onboarded to our platform. By combining these insights, we create our product roadmap. This is crucial, especially since we are operating in such a new category, where traditional competitor benchmarking is not always applicable. Additionally, trade media help us understand industry trends and where the market is heading, making it easier to align with the needs of these leaders.
How are WhatsApp and Instagram automation helping boost brand sales?
Automation on WhatsApp and Instagram boosts sales by enhancing engagement and streamlining the customer journey. On Instagram, automation helps brands respond to comments and reactions, guiding followers through the sales funnel and enabling in-app purchases via automated direct messages. WhatsApp automation offers personalized shopping experiences and efficient customer support, handling queries about orders and refunds, quickly. This boosts customer satisfaction, simplifies the buying process, and increases sales through continuous engagement.
Does the pop-up mechanism help drive conversations and conversions?
Yes, pop-ups play a key role in sparking conversations that lead to conversions. Gamified features like scratch cards or spin-the-wheel pop-ups engage visitors and encourage them to share their phone numbers. This triggers an AI-driven WhatsApp journey with personalized messages and offers, like discount codes, motivating immediate purchases. If a user doesn’t convert right away, the system nurtures the lead with tailored product recommendations, boosting engagement and increasing conversion chances. This strategy drives 60-70% of conversions.
How are you helping brands reduce Return to Origin (RTO) rates?
To reduce RTO, especially for cash on delivery orders, we focus on two key strategies. First, we push customers to make upfront payments by offering additional discounts and nudging them through WhatsApp, Instagram, and email. This confirmation before the order is shipped helps reduce RTO significantly. Second, we address the root causes of returns, particularly size and product mismatches, which are a big issue with cash on delivery. By integrating Manifest AI on the storefront, we can provide personalized size recommendations in real-time.
Do the shortening delivery cycles of Quick Commerce affect your work?
No. We focus on enhancing customer experience and building brand loyalty beyond the purchase. Our priority is pre-commerce and post-purchase interactions like product info, warranty support, and continuous engagement via WhatsApp, adding value without interfering with delivery timelines. We aim to enrich the user journey after purchase rather than impacting the delivery process.
For instance, toy brands use QR codes to guide customers to an engaging WhatsApp journey powered by LLMs, offering educational content.
In electronics, QR codes connect customers to WhatsApp for easy warranty access, simplifying registration and claims. This approach streamlines processes and boosts customer loyalty through continuous engagement.
What is the company’s five-year plan?
In the next five years, our goal is to be a publicly listed company. We see immense potential in scaling globally, and we believe our ability to build an engineering team in India, combined with our product’s appeal, gives us a strong competitive advantage. Indian brands are highly receptive to new technologies, which provides us with a great testing ground for innovation. We plan to leverage this market while expanding our footprint internationally.
