As a designer, developer and digital marketer for Mazepress, David Alexander uses personal assistants to handle basic tasks to improve his productivity. Some are quick and simple uses such as checking time zones that his clients are in and converting currencies. Other use cases are more advanced. For example, he asks for daily sales reports and other details from his Shopify stores.

Alexander admits these uses are fairly basic, but he is looking forward to the day when personal assistants start to mature into something smarter and more proactive. In the future, Alexander predicts that personal assistants like Amazon Alexa and Google Home will become more efficient and be able to handle multi-step requests like creating invoices, adding tasks to your enterprise resource planner or client information to your CRM. “AI will undoubtedly aid in the development of these tools as large amounts of usage data can be analyzed and then used to improve the functionality of voice or personal assistants.”

In some ways the future is already here, said Andy Abramson CEO of Comunicano. The market is already seeing machine learning and natural language processing in the workplace with services that can schedule appointments, or Google’s prepared replies to email, he said. These are basically entry level, first gen examples where repetitive tasks are handled by the machine. “Where all this is going is the movement from simple task based ‘assistants’ to neural network based reasoning, where the AI based assistant actually reasons the way ‘you’ would.

Following are some examples of this next level of personal assistants in action.

Related Article: Intelligent Virtual Assistants: Improving, But Still Not AI

Sifting Through Thousands Of Pitches

A journalist receives hundreds of pitches from publicists asking him to consider writing about their client. The journalist knows from his or her years of experience what a good pitch is and what makes for a good story for his readers, Abramson said. “Using a neural [network], only the stories that will likely be of interest will come to your attention, while the others are replied to politely with a decline,” he said. “This neural assistant does in seconds what takes hours each day of sifting and sorting, qualifying the facts, etc.”

Assistants That Act On Your Behalf

The scenario that Abramson described could, in fact, be applied to a number of situations. Personal assistants will augment the work of a professional in many different ways, said Swamy Gowda, co-founder at Wizergos. It could provide help in debugging an application or even a machine — Gowda envisions a day when blue collar workers will have their own assistants as well — where the engineer or mechanic is looking for a bug or flaw and the assistant is able to spot anomalies that could guide him in his analysis.

A personal assistant, one backed by deep AI, could also help in making decisions. These will understand the reasoning behind decisions made over the years, and would try to guide user to a successful outcome.

Learning Opportunities

Gowda also sees the day when an assistant is able to talk to other assistants to help you. “Think about the possibility that your assistant is doing your salary negotiations with your company's bot,” he said.

Related Article: Customer Service Heaven Is Just a Virtual Personal Assistant Away

Relegating Complex Tasks To A Voice Command

Voice-enabled chatbots aimed at the business user in which complex functions can be performed are on the horizon, according to Skuid CTO Mike Duensing. “These chatbots could have applications across a myriad of departments,” he said. Instead of digging through spreadsheets, a salesperson could simply ask their voice assistant something as complicated as "Search opportunities by accounts valued over $5 million," or, "Calculate project revenue in Q3," and come to their client meeting equipped with the proper business intelligence.

IT staff could use a voice assistant to monitor remote facilities and perform diagnostic checks. Marketing staff could assess the performance of a campaign by asking for click-through rate and conversions from a particular campaign. Initial implementations are still superficial, he said, “but soon we can expect employees to employ a voice assistant tailored to their specific role within a company that can deliver the business intelligence they need.”

Voice Assistants That Use Speech To Text, AI

There are emerging solutions that are focused on voice assistants that use Speech to Text (STT) to capture the text of meeting conversations, said Dave Damer, founder and CEO of Testfire Labs. “STT accuracy has made great strides over the last two years thanks to the massive increases in compute power and improvements in deep learning techniques,” he said. Unfortunately, he also noted that there are still many challenges that thwart truly automated and accurate transcription in meetings, such as people cross talking or background noise or accents.

Still, the first commercially available personal assistants aimed at meetings have begun to appear on the market such as Cisco’s Spark Assistant, while other companies such as, Voicera and are rolling out assistants that can transcribe, summarize and capture key data items, Damer said. “AI-based meeting systems also promise universal language translation, identification of speakers, and emotional analysis,” he said. “Natural Language Classification has been in use for several years but its capabilities are evolving quickly. These systems can not only detect sentiment and tone but can categorize emotions and intents and deliver insight on the moods and feelings of the speaker.”