Conversation with Dennis Mortensen: What the Future Holds with AI Agents - Banner
Tweet about this on TwitterShare on FacebookShare on Google+Share on LinkedInDigg thisShare on StumbleUpon

During the April 6th SalesTech event, Dennis Mortensen – CEO and Founder of x.ai – presented his views on the prospective use of virtual AI agents and the future it holds. At present, Mortensen and his company have their goals set on getting their personal assistant agents – Amy and Andrew– to as many potential users as possible.

Founded in 2014, x.ai “magically schedules meetings” for its users. The AI operates over email through a conventional human-like conversation. The personal assistants free up people’s time by dealing with the tedious dialogue ping pong that comes along with scheduling meetings. We caught up with Mortensen for more insight on Amy and the overall potentialities of virtual AI agents.

How can you assure global citizens that virtual AI assistants won’t be used to pass over personal or confidential information? With online hacking growing exponentially, should businesses and national security perceive AI assistants a threat to privacy in any way?

The idea suggesting more expensive equals a better experience and stronger security is not necessarily the case in the world of intelligent agents. If you take Amy (our AI personal assistant who schedules meetings for you) as the least expensive option and put her on a continuum with the most expensive option (a human personal assistant), then you see a surprising reversal.

Using Amy costs about $39/month, whereas a human personal assistant can cost anywhere from $800 (for someone who works virtually), to $7500 (for a full-time employee). Blended solutions start at $200 per month.

We would argue that the more expensive options are actually the far greater privacy risks. We’ve all made the Reply-All mistake, which could expose potentially sensitive information to anyone on the thread. We’ve also done the auto-complete error, where we send the email to the wrong person. A human personal assistant might accidentally expose information to the wrong people, and social engineering is notoriously effective.

All of these potential leaks can be engineered against in an AI agent. For example, Amy will never share any information about our customers except the available slots they are offering. They are immune to social engineering, and they don’t gossip.

We find this reassuring, as we’re likely to move towards a future with more, not fewer intelligent agents.

Are virtual assistants suggesting job cuts in the corresponding field? What measures are tech companies, who are developing virtual assistants, taking to safeguard certain positions — like those of personal assistants?

There are a vanishingly small number of people who get their own human personal assistants these days. AI assistants like Amy and Andrew can democratize access to services that were otherwise only available to a select few.

Scheduling meetings, managing your receipts, and repetitive sales tasks are among the chores we do on our own every day. We do these tasks without the help of others. They’re not core to our jobs and distract us from the real work we’ve been hired to do. But vertical AI agents are starting to take on some of these tasks. Amy and Andrew, for example, are able to take over the meeting scheduling process entirely and can give knowledge workers back hours in lost productivity.

We don’t see ourselves as replacing human assistants. We rather see ourselves as affording knowledge workers who don’t have a human assistant access to one aspect of that role by having AI manage their calendar.

Contemplating the rise in the competition of AI and its related products in the contemporary tech market. What would be the most reasonable cost of a virtual assistant with features similar to x.ai?

Amy and Andrew from x.ai cost $39 to $59 per month per user. We wanted to find a price point that was affordable because we (and our customers) see x.ai as part of the suite of tools that any functioning professional needs to operate. Our pricing reflects that, and as a result, we’re the most cost effective option in the assistant market.

On the customer’s end, the ROI is dramatic. Our pricing is hard to argue with when an individual or business owner sees how much time they’re wasting on meeting scheduling, and how much they could be saving with x.ai.

X.ai-ROI

What is the percentage of companies willing to take the risk of experimenting with relatively unknown products opposed to the more reliable and tested established virtual assistants?

Some companies will always be willing to adopt new technology early as a way to gain competitive advantage. We don’t see this as any different really. Today, intelligent AI agents are appearing on the market as the new “in” technology. However, we think that agents will be the norm soon enough.

Which extended features should virtual assistants have for larger companies compared to personal assistant bots?

In the future, we believe there will be a marketplace of highly specialized agents. Each focused on a single task which we call vertical AI. Some of them will be very specific to assist in a certain role or industry. Others (like Amy from x.ai) will cross roles and industries. And we believe that eventually, these vertical AI agents will integrate with AI assistants like Siri and Alexa.

Now that AI agents are working alongside humans, who is at fault when something goes wrong?

If we imagine AI agents more akin to employees or team-mates than apps, our relationship with them will need to be dynamic and two-sided. When something goes wrong, we’ll need to be curious and understand what went wrong and then how we can ensure it doesn’t happen again. When a teammate makes a mistake, we don’t throw them under the bus to a client. Instead, we apologize, figure out how to fix the mistake, and strive to learn something from it. The same should go for AI.

When we, at x.ai research make “mistakes,” the majority are either made by Amy’s boss or the guest. Assuming the machine is always wrong is no more helpful than assuming it’s always right. As we become more accustomed to working alongside AI assistants, we’ll learn to extend the same forgiveness to them as we would to humans.

Nearly 90 million US knowledge workers schedule roughly 10 billion meetings a year. Not one of those workers adds something meaningful to the scheduling process itself. In fact, most of us aren’t very good at scheduling meetings. Sometimes we forget to follow up, or pick a location that may work well for us but not so well for our guest.

An AI personal assistant can do better than humans. By taking over the job of scheduling your meetings from start to finish, Amy saves you time and aggravation. Most importantly, because Amy and Andrew are purely AI, we’re excited to make them accessible to all 90 million of those knowledge workers.

Read our next blog ....