SUMMARY
QX Lab AI has launched a chatbot, Ask QX, which serves in over 100 languages, including 12 Indian languages
Currently, the startup fosters a network of more than 14 Mn users and has set its eyes on expanding its geographies to the GCC and MENA markets
The platform is currently in the beta phase but already showcases distinct behaviour, incorporating ethical dilemmas and other features
As generative AI becomes increasingly influential worldwide, numerous tech giants and startups are entering the field by introducing new AI chatbots. These chatbots boast capabilities such as answering questions, generating copy, drafting emails, engaging in conversations, and explaining code in various programming languages.
Notable players in this segment include OpenAI’s ChatGPT and Google’s Gemini. However, it’s worth noting that many of these AI chatbots are primarily designed to interact in English, potentially limiting their accessibility.
To counter this, QX Lab AI launched its Ask QX chatbot last month. The chatbot is available in over 100 languages, including 12 Indian languages.
“We didn’t start with large language models. We began with language models in general. There are more than 10,000 language models, with GPT being just one of them. So, we began our research in this area. Over the past eight years, technology has advanced significantly,” Tilakraj Parmar, the founder & CEO of QX Lab AI, said.
Founded in 2018 by Tilakraj Parmar, Arjun Prasad and Tathagat Prakash, QX Lab AI is an artificial GenAI company, which aims to advance AI capabilities by enhancing human-AI interactions. What forms the core of the company’s vision is its commitment to making AI accessible to people from all walks of life.
Currently, the startup fosters a network of more than 14 Mn users and has set its eyes on expanding its geographies to the Gulf Cooperation Council (GCC) and Middle East and North Africa (MENA) markets.
How Is Ask QX Chatbot Different From Its Rivals?
Unlike ChatGPT and Gemini, the platform utilises neurologically trained algorithms in over 100 languages, employing a hybrid model consisting of 30% Large Language Model (LLM) and 70% Neural Network Architecture. This approach has made it cost-effective. Although the subscription model is yet to be introduced, the premium plan will be priced at $1.
“Our proprietary architecture allows us to save 70% of the company’s costs. We have applied for patents for all our technology. This proprietary algorithm is a significant advantage for us, enabling access to this technology at a very affordable price. If you look at our competitors, they are charging premium plans in the range of $25-39,” Parmar said.
Besides the price point, there are factors which make their hybrid model different. The platform is currently in the beta phase but already showcases distinct behaviour, incorporating ethical dilemmas and other features.
“We are 70% neural network based and only 30% LLM based. Therefore, with 70% of our architecture relying on neurons, there will be a reduction in computational power. While we cannot definitively claim sustainability due to our reliance on computational power, we have strived to adopt the most sustainable approach possible. We acknowledge that our computational power is not yet at an optimal level. However, we have made efforts to minimise our server’s energy consumption in developing our product,” Parmar said.
Currently, Ask QX offers three distinct modes designed to cater to the diverse needs of its users — ‘Creative’, ‘Standard’, and ‘Professional’. Standard Mode is designed for daily queries and tasks such as quick help with homework, email drafting, travel planning, and daily schedule management.
Creative Mode steps in when someone is seeking inspiration or innovative solutions. This mode is ideal for breaking through creative blocks, brainstorming ideas for marketing campaigns,
writing, or any artistic endeavours.
Furthermore, the Professional Mode offers expert advice and detailed, precise responses tailored to professional needs. This mode provides the industry-specific insights and expert knowledge necessary to navigate complex challenges in professional plans.
But What’s The Monetisation Plan?
The startup has not yet entered the revenue generation phase. Founders want to start charging users once the product is ready. Currently, the focus is on product development. However, they will launch the subscription plan priced at $1 this year itself.
“Until we feel the product is ready, we won’t even mention the $1 scheme because we believe in always delivering the product first. So, we want to have it fully ready,” Arjun Prasad, cofounder and chief strategy officer of QX Lab AI said.
Currently, on the Ask QX chatbot, a premium plan is mentioned. However, when a user clicks on it, it shows “Join the queue. We’ll inform you via email as soon as you’re approved to become our premium member,” indicating a possible phased launch of the premium plan.
Parmar said the company is working on introducing additional features, such as archiving and pinning conversations, to its premium plan. The idea of pinning conversations for users will particularly help students and researchers.
The founder said that features like pinning and archiving conversations are crucial for enhancing user experience.
QX Lab AI In The Indigenous GenAI Battleground
As of now, the startup wants to go beyond the B2C model and is looking to launch a B2B product in the second quarter of 2024. For this, the company is considering partnering with enterprises such as healthcare, and education, among others.
Another reason for QX Lab AI to enter the B2B space is that the regional language market for GenAI is getting crowded at a very fast pace. Many startups have embraced LLMs crucial for GenAI applications, specifically trained in Indian languages.
A surge of India-focused LLMs is sweeping the nation. With CoRover.ai’s BharatGPT supporting 14 Indian languages, and Tech Mahindra’s announcement of The Indus Project, an open-source LLM focussed on 40 Hindi dialects in its initial phase, Parmar and his team will have to stay ahead in the GenAI game if they want to remain relevant in the industry that is projected to cross the $17 Bn mark by 2030.
Earlier, Prime Minister Narendra Modi unveiled Bhashini, an AI-based language translation tool for Indian languages, enabling real-time translation. Therefore, QX Lab AI may have to broaden its playbook beyond just offering a GenAI chatbot for Indic languages playbook. What will help the startup stay relevant in the Indian GenAI battleground will be its ability to deliver accurate and swift responses, minimise hallucinations, and offer innovative features.
