Indian Data Annotation Startup Playment Acquired By TELUS International

Bengaluru-based data annotation startup Playment has been acquired by Vancouver-based digital consultancy TELUS International for an undisclosed amount, the company said in a statement today.

Playment was launched in November 2015 by former Flipkart employees- IIT-Kharagpur alumni Siddharth Mall and Akshay Lal, along with IIT-Guwahati alumnus Ajinkya Malasane. It offers software solutions to build high-quality visual data labels- a process known as data annotation. The startup works with more than 200 machine learning teams across global companies. Playment’s GT Studio offers ML-assisted 2D and 3D labeling tools.

Globally, most companies outsource data labelling jobs to dedicated data annotators. Data labelling takes up the bulk of data scientists’ time, which could otherwise have been devoted to building algorithms. As in case of technology outsourcing opportunities, India has emerged as a leading data labelling destination globally.

“The (computer vision) technology has become increasingly valuable to tech firms and the wider market as brands seek to uncover solutions for scenarios that typically require complex human judgement from visual and spatial patterns to solve nuanced business challenges, and build innovative, market-leading smart products and services,” said Jeff Puritt, president and CEO of TELUS International.

Playment has raised $2.5 Mn funding till date. Its last funding was in 2017 when it raised $1.6 Mn as part of its seed funding led by YCombinator, Sparkland Capital and Silicon Valley angel investors such as Ryan Petersen, Max Altman and others. Playment was also part of YCombinator’s Winter 2017 batch and it was part of the Google launchpad accelerator. Earlier in 2016, Playment had raised $700K from Elevation Capital as part of its seed funding round. 

The startup has been profitable since April 2020, claims the company. So what is it about Playment’s unique offering that made it the target of this acquisition?

Playment Product and Pivots

While artificial intelligence drives a lot of solutions globally, little is understood about how it works. Playment particularly deals with AI and machine learning software of the visual kind. Suppose an automobile manufacturer wants to build self-driven cars. It will need software that recognises objects and visual cues that the vehicle driver normally sees. However, before the machine learns that it has to avoid some visual elements- like people and animals- and respond to elements like traffic signals, the software has to know what these elements are. This is where Playment’s solutions come in. 

The startup, 70% of whose clients are automakers, helps the self learning software of its clients identify objects.  In 2015, the team started with providing cataloging and content moderation services to large marketplaces such as Flipkart, Lazada, Paytm, Ola and others. “At Flipkart, we used to address a lot of repetitive cataloguing requirements. Especially because a lot of sellers did not know how to categorise their products properly. This led us to build the Playment solution,” said Malasane speaking to Inc42. 

To make this feasible, Playment technology divided the big chunk of work into small micro-tasks like lego blocks and qualified users on its platform solve the entire puzzle. For instance, sellers often bulk upload apparel catalogues on ecommerce platforms. A person on the ecommerce platform has to inspect each and every item submitted before it is published to ensure that the products are labelled in a way that buyers can find easily. Playment simplified it allowing products to be identified by machine learning. 

In the initial years, Playment used to work with crowd-sourced data annotators who would do the labelling work. “Since 2019 we shifted to a hybrid model that includes crowd-sourcing but majorly partner with over 10,000 annotators across secure business process outsourcing (BPO) companies in India to offer the same solution keeping in mind that a lot of these client projects are highly confidential,” said Malasane. This demand for data annotation solutions also helped the company become profitable in 2020.

In India it competes with data annotation companies like iMerit and Infolks. Globally, it competes with the likes of Scale AI and Appen. It was in the process of looking out for partners and investors in the business that Playment founders engaged with TELUS leading to the eventual acquisition. All 85 of the startup’s employees, including founders, will join TELUS post acquisition.

Data Annotation And Computer Vision Opportunities

The global data annotation tools market size was estimated at $321.45 Mn in 2020, according to a report by Expert Market Research. The data annotation tool market is projected to record a CAGR of 27% between 2021 and 2026. The market growth is largely attributed to the increased adoption of software for image data annotation in the automotive, retail, and healthcare sectors.

Currently, Playment services a host of international clients including Samsung, German automotive tools giant ZF, US-based self driving solution company Nuro and Daimler AG among others. TELUS’s largest external client globally is Google. “With TELUS in the driving seat we hope to reach a larger client base globally. We are also aggressively hiring for these opportunities,” says Malasane.