Data-centred services have been on the rise in Singapore and around the Asia Pacific over the past few years. Clubbed together with machine learning and artificial intelligence, big data has risen to become one of the most promising sectors in the region. Businesses today are able to leverage data analytics for growth and scalability.
One of the most crucial and biggest aspects of data is audio, and automatic speech recognition solutions based on AI and machine learning plus data analytics are changing the audio data industry in the 4.0 era. On a global level, the speech and voice recognition market is expected to grow at a CAGR of 17.2% from 2019 to 2025 to reach $26.8 billion by 2025.
However, when it comes to enterprise ASR solutions, there are still some major gaps and scope for improvement. AI-based ASR transcription in the enterprise has an accuracy of just 60 to 70 per cent as compared to 80 to 90 per cent in consumer applications like Alexa and Siri. Furthermore, the processing speed of the ASR engines tends to be a lot slower than what is required in enterprise use cases where large volumes of audio data need to be transcribed and long turnaround times make it inefficient. Data privacy and integrity is yet another problem faced by the enterprise ASR sector where there is a lack of clarity on segregation, storage, usage, and security of data.
As such, Singapore-based AI Communis is helping address these three core challenges so businesses can benefit from their audio data and increase productivity.
Addressing the core challenges faced by the enterprise ASR industry
AI Communis provides customised engines that provide better accuracy that can go as high as up to 85+ per cent. They are able to achieve this because each ASR engine is custom made to suit the need of the specific client enterprise in that engine learns from the client’s unique audio data, understands the industry and company jargon, and thus is able to provide a more accurate transcription. They do not follow a one-size-fits-all model.
“When I first met with my co-founder, we spoke about the future of AI and ASR. We quickly understood that to realise our vision, we need to have proprietary technology and that is why we are very focused on our core technology,” shares Nobuhiko Suzuki founder and CEO of AI Communis who comes with over 18 years of experience in banking as well as corporate strategy.
Furthermore, AI Communis’s engines are much faster by up to 20 times, transcribing a one-hour file in just 60 seconds. Big tech and legacy companies take about an average of 1200 seconds to transcribe a one-hour audio file.
Another key highlight is that the startup has full ownership of the entire technology life cycle and thus complete control of the data, which ensures that there is no cross-contamination of data plus the enterprise has complete visibility into the data’s storage, access, and control.
Data privacy and security is one of the major challenges faced by startups, SMEs and big corporations alike in Singapore. This is even more pertinent today amidst a dramatic shift towards digitalisation following the coronavirus pandemic and subsequent lockdowns. While businesses are being encouraged to leverage big data and embrace digital transformation for survival, growth, and scalability, the risks of cybersecurity is one of the main threats faced by companies. With AI Communis, businesses need not worry about data contamination and other risks because they have dedicated virtual private clouds ensuring maximum data integrity and security.
“We own the underlying AI technology and that is what separates us from the rest of the players — having the ownership of the entire tech stack. This also helps us ensure maximum data integrity and privacy,” says Devanjan Sinha, VP of Operations at AI Communis who comes with 12+ years of experience in equity research and investments.
Additionally, AI Communis’s ASR engines are able to support three different languages, including English (UK, US, and Japanese. Singapore Standard English and Bahasa are under development. They offer real-time transcription and are able to offer on-premise support.
Improving efficiency in compliance monitoring at Financial Institutions and other sectors
Financial institutions, such as banks and fintech startups make sales calls all day long and they have to comply with a host of regulations and guidelines. It can be tedious for a small team of compliance professionals to monitor all sales conversations which can be around 100 to 200 hours per day. A single slip can lead to risks of fines, lengthy customer remediation processes and loss of brand credibility. This is where AI Communis’s ASR engines can help: they transcribe the conversations, then run their algorithm to pick up problematic keywords and flag them, reducing time and effort, and increasing productivity.
In addition to the finance sector, the core technology of AI Communis’s ASR solution has a wide range of use cases. They can help enterprises in different sectors, including real estate, logistics, education, governments, as well as corporates.
Startups and SMEs looking to enhance productivity by making the most of their audio data can engage AI Communis’s SaaS model via APIs to leverage their customized ASR solutions. Whereas, for large enterprises that have more robust tech infrastructure, AI Communis can to license their core engines.
With a keen focus on banks and financial institutions, such as insurance companies and fintech startups, AI Communis is looking to onboard clients and collaborate with partners in Singapore and across Southeast Asia. To learn more about them or to connect with them, visit https://e27.co/startups/ai-communis-pte-ltd/
This article is produced by the e27 team, sponsored by JETRO
We can share your story at e27, too. Engage the Southeast Asian tech ecosystem by bringing your story to the world. Visit us at e27.co/advertise to get started.
The post AI Communis: The first B2B ASR-based solution provider in SEA with local language adaptation appeared first on e27.