Jarvis+ has continued development of its decentralized conversation-as-a-service platform into 2019, releasing weekly reports on its progress. Jarvis+ hopes to enable natural language interactions with blockchain and smart contracts.

Two examples of Jarvis+ services are its group quiz and daily check-in games that are played on the Telegram chat platform. The games make use of Jarvis+ technology by asking participants questions, which if answered correctly are rewarded in JAR tokens automatically through a smart contract.

At the start of the year, Jarvis+ allowed its flagship community-engagement bot, Grouplus, to be accessible via a free trial. Grouplus is described by Jarvis+ as an “AI assistant” that can perform topical and public sentiment analysis as well as manage community ecosystems.

Grouplus can be deployed on most chat services including Skype, WeChat, Discord, QQ and Telegram. It’s also reported to be capable of intelligent conversation with threaded responses for querying token prices and automated customer service. A complete list of Grouplus’s features can be read here.

By the end of January, Jarvis+ had completed push notifications for its application and entered internal testing for its virtual KOL. Some improvements were also claimed to have been made to its FAQ bot, which is another system under development.

In February, Jarvis+ reported a general improvement to its WeChat integration as well as the rollout of push notifications for English and Chinese versions.

The team also attended the NEO DevCon event that was held in Seattle from February 16th to February 17th. In its summation of the event, Jarvis+ described it as a “forward-looking premium summit for developers, showcasing various decentralized applications and top contributors in the ecosystem.”

March saw the release of Jarvis+’s 30th weekly report, as well as a slew of development updates. The team added “emotion recognition based on deep neutral networks” as part of its AI training as well as the ability to detect and classify various types of spam in Chinese.

In the platform’s most recent update, more work was reported on the spam detection algorithm, with Jarvis+ stating: “Spam classification has been deployed online, using deep neural networks for continuous learning and training. Currently supporting Chinese, it can identify eight categories of spam, including advertising, abuse, porn, politics, violence, and others.”

Emotion recognition was also completed, along with some improvements to its WeChat functionality.

The group quiz and daily check-in games remain ongoing.