Verbit posted on 30 Nov 2020
"Transcribing audio is time-consuming, often tedious and as a result, humans are prone to make mistakes when they do it. AI machines excel in precisely these kinds of scenarios," said Tom Livne, CEO of Verbit.
Today is an exciting day for me, as I’m celebrating two years with Verbit. It seems to be the best time to write an article about the company. When I first entered the Verbit office in Tel Aviv, I knew that I was going to make a career here at the company. I knew little about transcription, but I felt the impact and recognized the potential of Verbit early on.
Transcription dates back to ancient Egypt. During the height of the Pharaohs, transcriptionists were important figures of the courts and were known as scribes. The scribes were exempted from physical labor and taxation. The word “transcription” was first used in the modern world in 1598 by Joh Florio as an entry in his Italian/English dictionary.
On average, we speak 150-170 words per minute, which is five times greater than what we can type. That’s why it’s typically a manual process, taking at least 5 hours to transcribe a one-hour file. It can take much longer if there are multiple speakers, difficult audio quality, or accented speech.
A few interesting market numbers: The transcription market in the US alone was valued at $19.8 billion USD in 2019 and is anticipated to expand by 6.1% from 2020 to 2027. The fact is that organizations across the globe generate large volumes of data every day that can be effectively used to obtain valuable insights.
The importance of accuracy in reporting on data is significant, as these numbers are often used to make critical business decisions. Data is often quite sensitive, so security is an important factor that should not be overlooked by a transcription company when transcribing the data of its clients.
Often, a business or organization will use transcriptions for research projects, classes, webinars, legal proceedings, data analysis, blog posts, website content, search engine optimization (SEO) and anything else that requires converting speech to text.
With that said, these are the 10 reasons why Verbit is effectively revolutionizing the world of transcription:
#1. Proprietary speech recognition technology. Verbit builds its own speech recognition engine in-house. The engine uses three models: The acoustic model: reduces background noise and echoes to cancel out factors that reduce the audio quality. The linguistic model: identifies specific terminology, recognizes different accents and dialects, and differentiates speakers. The contextual events model: incorporates current events, news and relevant updates. Verbit uses the latest advances in Deep Learning, Neural Networks, and Natural Language Processing to train its algorithm, making it more accurate over time.
#2. Domain-focused. Verbit’s smart approach to disruption in the transcription market is focused on the vertical and being an expert in it. Verbit operates in 4 domains: Higher education, legal (court reporting), media and enterprise. This approach allows Verbit to understand the needs and challenges of those domains and provide customized and tailor-made solutions to address all needs.
#3. Consistent accuracy (Hybrid model). Verbit’s in-house, AI-enabled speech recognition technology is 90% accurate on its own. To reach its guaranteed 99.9%+ accuracy, the Verbit solution incorporates a community of 22,000 human transcribers to correct errors and make any necessary revisions. Verbit has a unique scoring method and quality assurance processes for the transcribers to make sure they’re doing their jobs properly.
#4. Competitive pricing model. One of the key factors of Verbit's success is its hybrid model of artificial + human intelligence. Automating the majority of the transcription and captioning process saves money, as it presents a reduced need to invest resources in training and salaries. The ability of Verbit’s solution to achieve a high level of accuracy means less manual work is required for correction, resulting in less spend.
#5. Security. Verbit conducts a variety of audits to ensure continuous compliance with industry-standard best practices: HIPAA, SOC 2 and GDPR. All connections to Verbit are encrypted using SSL, and any attempt to connect over HTTP is redirected to HTTPS. All customer data (including recordings and transcripts) is encrypted. Verbit utilizes relational, as well as NoSQL databases as managed services hosted by AWS. Verbit customers may configure a data retention duration, and customer data is purged from Verbit’s systems in line with contract termination upon request.
#6. Well funded. Verbit is a VC-backed company. In the last week, Verbit raised $60 million USD in its Series C round with a total investment of $125 million USD since the inception. Verbit is the most well-funded company in the transcription and captioning world. Some of Verbit’s investors include Sapphire Ventures, Stripes, Vertex and Viola Ventures.
#7. The 'white-glove' approach. Unlike other transcription and captioning vendors that operate as a B2C, Verbit has a different approach. The company is an enterprise-focused company that allows Verbit to be customer-centric and deploy many resources for customer success. There's a unique onboarding process for every customer that combines training, support and consultancy.
#8. Scalability. Due to COVID19, there's a much higher demand for transcription and captioning tools, and a lot of other companies cannot handle it. Verbit has a massive community of 22k+ transcribers and live captioners from 30 countries around the world. Verbit’s solution increases an organization’s operational capacity by allowing unlimited volume, leading to significantly shorter turnaround times and higher profit margins.
#9. One-stop-shop. An important offering for Verbit’s clients is the capability to use a hybrid platform for everything. Verbit provides real-time captioning, closed captioning and transcription solutions at scale. A lot of companies that used multiple transcription vendors face challenges in terms of accuracy, speed, scale and the quality of the services. Verbit’s AI-generated transcripts also offer differentiated value in that they are actionable. Users can search these transcripts in real-time for keywords, take notes, highlight within them, and take advantage of additional interactive features.
#10. Culture. Verbit’s culture is one of its most powerful assets. The workplace culture is based on six main elements- Collaboration: Contribute your unique talents to the joint effort. Success comes when working together. Winners: One way or another, you will get there. Excellence: Enable people to be the best version of themselves. Ownership: Lead by taking action. Own by committing to results. Innovation: Be curious and ask questions. Make an impact. The joy of creation: Enjoy the ride while doing your best.
In conclusion, Verbit's vision is to provide a technology that fuels a world where all media shared in classrooms, workplaces and online platforms are inclusive. Verbit is dedicated to providing universities and businesses with the tools they need to make all video and media accessible to all individuals.
I hope you enjoyed my 2-year anniversary article for Verbit.
Exciting times at our portfolio company Innoviz Technologies !
Innoviz today announced a SPAC merger to go public at a price of $1.4B, an important milestone in the company's amazing journey to make autonomous and safer driving a reality in the future.
Congratulations to Omer Keilaf, Oren Rosenzweig, Oren Buskila and Eldar Cegla! You are an inspiration to us.
We are truly excited to be a part of this amazing journey, and are looking forward to what 2021 will bring.
Congrats Michael Gabay, Daniel Gabay and the unstoppable Trigo team.