man on his cellphone reading email
PHOTO: Anete Lūsiņa

Steph Locke, CEO of Nightingale HQ, has 11 interviews scheduled for the next day in her Outlook account. A message arrives: does she need to book some time for herself as well? Well, actually she does, but Locke knows that is not going to happen. So she tells the messenger not to bother with it. 

Locke uses Outlook and this particular feature is part of Microsoft Insights — and an example of how artificial intelligence (AI) has crept into our email usage. It also suggests how AI is poised to go even deeper into email as various research initiatives get underway. The larger theme will be the same for the most part — extracting information about tasks that need to be completed from the emails and placing them in a list. The future of AI in email, though, also includes integration into other messaging channels as well as other related tasks, such as transcriptions of meetings and the automatic inclusion of documents. 

The Integrative Productivity Experience

“We are going to see a natural extension of email into applications such as Slack, overlaid with AI,” Locke said. “It will become less obvious what application you are in and much smoother to move between company tasks, or adding emails to ad-hoc meetings. We will see tasks emerging from email that are automatically being placed in to-do lists, with reminders pinging people in teams on Slack.” These applications will be helpful in other ways too, she continued, such as automatically sending documents that are referenced in a meeting to be read by the team. “We will definitely see this in the next few years.”

Google, for its part, is focusing on bringing its email platform back to a centrally-oriented one as it removes its legacy apps. Locke said Google is at somewhat of a crossroads as it still has to determine whether it wants to invest in hangouts and further AI integration or whether it aligns with applications like Slack to compete against Microsoft on integrative productivity experience. 

Microsoft is also pushing forward by adding voice into email and general productivity apps, such as meeting transcriptions, Locke said. “Special devices could be used to detect who we are and extract what we said into to-do items. Or it could just send us the transcriptions.” 

Related Article: Natural Language Processing Is Hitting Its Stride

Natural Language Processing and Email

There are other use cases Locke foresees, such as the ability to send a group email, which would be scanned beforehand for inclusive language. “That will be the next level — as we are transcribing meetings and identifying attributes in the conversation, AI will suggest tips on using inclusive language to improve the way people are interacting.” 

Voice — fueled by advances in natural language processing (NLP) — will also be a key driver in AI’s further push into email, according to Gillian McCann, co-founder and head of cloud engineering & AI at Workgrid Software. “After all, email is basically about having a conversation with someone. It is possible to replace email in some ways with collaborative and AI-based technology,” she said. 

One simple example is a voice assistant, which can provide a summary of what the email says. Another is to use voice to create the email in the first place. There are, of course, variations of this on the market, but in McCann’s vision the functionality is seamless and not so glitch-prone as it is today. “Reading comprehension for NLP has seen great advances recently and it will become part of our everyday experience with email,” she said.

The Corporate Realm

There are some corporate email solutions that are moving at a faster pace and different direction than what is happening with the mainstream providers. Messagepoint, for example uses AI to identify elements within an email that are out of compliance with a brand’s standards, according to Patrick Kehoe, EVP of product management. Screening for restricted terms like acronyms, discontinued brand names, or jargon are some examples, he said. “Further, it can help to identify inconsistencies in the phone number, date or time formats, use of contractions, proper application of copyright or trademarks, etc.”