AI E-Mail Assistants Streamline Meeting Scheduling

Emailing back and forth to book meetings can be a huge waste of time, especially when people are in different time zones, don’t know each other’s availability, or schedules change, etc.

In a Doodle survey of 1500 professionals, respondents needed an average of 4.75 hours to arrange nine weekly meetings with seven parties. A personal assistant could handle that, but for that would be cost prohibitive for most. People need an easy and affordable solution to scheduling meetings by email.

With all the progress being made in Artificial Intelligence (AI) and Natural Language Processing (NLP), how about building an AI powered virtual assistant? The idea of intelligent personal assistants and AI scheduling appeared more than years ago, but early research never took off: AI algorithms were not strong; there was a lack of quality training data; and the whole scheduling problem remained too complex. It took years, but now several promising AI meeting scheduling startups and products are emerging.

Industry leaders

Clara Labs – designed for enterprise

Clara Labs is a San Francisco startup that created the industry-leading Clara AI assistant, which can automatically schedule meeting times and places based on attendants’ preferences. Users simply “Cc” Clara on meeting invites, and she sorts it out. If Clara can’t, one of the company’s human staff members will step in. Clara Labs raised $12 million in two funding rounds, with a US$7 million series A led by Basis Set Ventures and Slack Fund in July 2017.

X.ai – maximize automation, lower pricing

The New York based startup’s AI personal assistants Amy and Andrew have meeting scheduling workflows similar to Clara Labs. X.ai however offers a free basic plan, with features similar to Clara’s US$99/mo essential package. X.ai was founded in 2014 and has attracted US$32.3 million in funding.

Calendar.help – a Microsoft product, takes care of business

The Calendar.help project started years ago, and enables Microsoft digital assistant Cortana to arrange meetings and schedule tasks. In 2016, Microsoft also acquired Genee, an end-to-end scheduling tool that integrates with calendar apps and email providers. As a Microsoft product this solution can naturally integrate with Office 365. Calendar.help values performance over speed and automation, with a workflow that brings both algorithms and human assistants into the scheduling loop.

Julie Desk, Evie –competitors from Paris and Singapore

Julie Desk is a French startup that has raised US$3.69 million, while Evie.ai is a Singapore based startup. Both were founded in 2014. Their AI email assistants Julie and Evie are X.ai and Clara Labs’ overseas competition.

Fin – more powerful than email schedulers

Fin is a San Francisco based startup founded in 2015. Its powerful personal assistant leverages AI techniques in speech recognition and natural language processing with human assistants, and can call, text, email, schedule, research, book, and even make purchases.

Limitations

Will AI and robots take over from secretaries and personal assistants? Not yet. “Despite advances in natural language processing, extracting information from free-text is still error-prone and algorithms often make mistakes that seem trivial to humans”, says a Microsoft researcher. This is why most of the startups we surveyed still involve humans in their AI-powered products.

Assistant systems are also not yet scaling up. Clara Labs’ median virtual assistant email response time is a sluggish 44 minutes, which raises concerns about how the system will perform if hundreds of thousands of people use it at same time. Moreover, vendor contracts and policy and procedure standards need to be developed to address security and privacy issues. We can expect to see improvements in AI assistants’ control functions and automation activities to address these operational risks.

Looking Ahead

There is no doubt that future email assistant systems will have fewer humans in the loop, process messages faster, and be more accurate. They may also integrate with more platforms, such as team collaboration tool Slack or social media app WeChat. Tomorrow’s conversational AI assistants will be more functional and push deeper into enterprise. They may also become stenographers, make meeting contents searchable, follow up on action items, or highlight key moments of a telephone call. This may sounds futuristic, but researchers are already working on it!

Reference

https://www.aaai.org/ojs/index.php/aimagazine/article/view/2039/1932
https://hbr.org/2017/07/how-we-built-a-virtual-scheduling-assistant-at-microsoft
http://delivery.acm.org/10.1145/3030000/3025780/p2382-cranshaw.pdf?ip=207.237.117.131&id=3025780&acc=OA&key=4D4702B0C3E38B35%2E4D4702B0C3E38B35%2E4D4702B0C3E38B35%2E7FDF4EA53E57CC84&CFID=965890573&CFTOKEN=57692965&__acm__=1501293677_f40c3a29cd5a4e6bd7825e2f6977eab7
https://www.pwc.com/us/en/financial-services/research-institute/assets/pwc-top-financial-services-issues-2017.pdf
https://techcrunch.com/2017/07/19/claraseriesa/
https://blog.claralabs.com/claras-promise-bringing-transparency-to-conversational-intelligence-dfce8e1af66d
https://www.forbes.com/forbes/welcome/?toURL=https://www.forbes.com/sites/clareoconnor/2017/04/18/clara-labs-wants-to-save-your-from-your-inbox-with-cyborg-assistants/&refURL=&referrer=#fcaf2b174ca2
http://www.nanalyze.com/2017/03/ai-personal-assistants-secretary-duties/