Artificial Intelligence is getting pretty wild, y’all. Google and Uber are both working on developing AI systems with self-doubt, the University of Cambridge added a “Superintelligence” modification to popular computer game Civilization, and Japanese scientists can basically read minds with deep neural networks now.

Artificial Intelligence (AI) broadly covers the idea of machines and technology carrying out “smart” tasks. AI is driven by machine learning (ML), which allows devices to analyze data and learn through pattern recognition.

AI’s potential is widespread, from personal assistants like Siri and Alexa, to services like Pandora and Netflix. Utilizing machine learning (ML) software, these services apply algorithms to data sets to analyze and learn user preferences.

Whenever you like a movie or show on Netflix, you get suggestions of what you may like based on previous reactions, watching history, and Netflix’s extensive dataset. Machine learning does the analysis work, while Netflix as a service is considered something that uses AI.

Many companies use AI and ML to evaluate and manage data. In 2016, $20-30 billion was spent worldwide on AI. Of this, ninety percent went to research and development, which speaks to global interest in improving and increasing AI technology.

As the amount of worldwide data increases, AI and ML can help manage information and deliver insights across a variety of industries, including retail, real estate, education, energy, manufacturing, and so many others.

Sales can particularly benefit from AI since it reduces the manual labor of researching prospects and qualifying leads. With AI, sales teams can determine when to engage prospects, and which information will be most relevant.

Additionally, AI provides insight into which content is doing well so sales teams can better optimize high-performing strategies. In turn, this can improve engagement based on insights instead of intuition to increase close rates.

Close analysis of data doesn’t have to be a tedious administrative task with AI and ML. By finding out what your customers need based on close data analysis, you can create targeted, personalized solutions.

Plus, AI can help reduce lost sales by evaluating product availability, and implement dynamic pricing along and demand forecasting.

In terms of customer support for sales, you can already easily implement chatbots that use machine learning to answer frequently asked questions and generate leads.

We’re not exactly at Westworld levels of automation yet, but the future is leaning towards AI. Those in the sales industry can greatly benefit from implementing artificial intelligence solutions to save time and increase productivity for anyone who’s still human on the team.

And now for a neat graphic to digest:

If you’re already running everything in your business though Slack, you might want to keep an eye on Ottspott, a startup currently registering early adopters in beta.

Ottspott is a phone system that runs inside Slack. You don’t even have to own a telephone set – you can make and receive all of you calls through your computer browser, without leaving Slack.

No coding or technical skills are required. Sign up takes less than a minute, and your entire team is integrated into the system – no need to invite team members or have them sign up individually. You simply select a phone number from a list of 9,000 cities in 40 countries, and Slack takes care of the rest. Included are major tech cities such as Dublin, Amsterdam, London, San Francisco, and New York. Ottspott is a great tool for global businesses that want to keep local phone numbers for their customers.

Ottspott can help you with internal communications, as well as calling clients and customers.

You can label calls for efficiency (for example “urgent” or “sales”), and you can have calls automatically forwarded to the appropriate member of your team. Your Gmail contacts are integrated with Ottspott to provide caller ID. You can also create folders of contacts to share with your team. Ottspott can even facilitate conference calls using Slack’s slash commands.

Ottspott notifies you instantly when you receive or miss a call, or when you get a new voicemail. You can then click-to-call from these notifications or from within voicemail, so you don’t need to dial the number.

You can also use OttSpott’s analytic metrics to measure your sales team’s phone performance.

And what if you’re away from your computer? No problem. Ottspott has a built-in voicemail system, and can also forward calls to your cell phone or landline.

In yet another instance of the machines winning, Google recently released a demo of Google Duplex — an automated calling suite — which showcased the software completing calls to schedule appointments on behalf of a user:

Duplex, which will soon be bundled into Google Assistant, sounds uncannily natural; when completing a phone call, the AI can handle and react to unclear instructions such as long pauses, deviations from the conversation’s topic, being placed on hold, and being asked to repeat itself. While the main two calls released by Google only show Duplex operating in two venues (a restaurant and a hair salon) Google plans to implement Duplex across multiple platforms eventually.

That means you may start getting calls from Google – we’ll get to that in a bit…

Having the conversation sound as natural as possible was a key point for Google. Since most conversations with AI assistants tend to feel jarring and forced — especially from the AI side — it was clearly important for Duplex to feel as inviting and human as possible. This is evident from Google’s inclusion of various hesitations (e.g., “um”) and variations in the language used by Duplex.

Timing is another critical component of Duplex’s mannerisms.

While many AI assistants have uniform timing between specific conversational segments (such as sentences), Duplex pauses almost intimately, and its reactions to new information sound realistic enough. Between Duplex’s timing and the “flawed” mannerisms mentioned above, the AI represents a tremendous step forward for the human-facing side of AI.

Keep in mind that the AI currently has some limitations regarding its conversational abilities; it seems that Google’s strategy was more based around fleshing out a few specific scenarios and expanding the AI’s conversational options within them than allowing the AI to run wild with limited conversational depth. Eventually, though, Duplex will most likely be much more capable than it is now.

For example, as of now, a Google Assistant user might feasibly ask Duplex to schedule a restaurant reservation or inquire about busy hours. However, future renditions of Duplex may comprise tasks such as scheduling a vehicle repair, ordering take-out, calling an Uber, and more. Like calling to set up a house showing, even though you already have a button for that on your site, can do it via email or app, and pay for a service to manage all of this (the consumer cares about their convenience, not yours).

So what happens when your phone rings and it’s a robot?!

For now, Google Duplex isn’t selling leads, they’re simply launching the beta test as a scheduler, which we all know will become more complex in the future. But let’s say someone tries this during beta test: “Hey Google, call Ron Thompson at Century 21 in Dallas and schedule a home tour for tomorrow at 2pm.”

What is Ron going to do when a robot calls and they don’t know who the client is, hasn’t prequalified them, doesn’t know where they want to take tours, and also doesn’t understand he’s talking to a robot? The call is going to fail and Ron isn’t going to get the lead.

Perhaps the client will move on to the next person, or perhaps they’ll understand that their request takes more human intelligence than a robot can provide.

But more importantly, how should you react when you get your first call from a robot? Just behave normally. Speak naturally and normally, and if you can help, do it, but if you can’t, make it clear why. The information won’t necessarily make it to the client, but calls are recorded and AI learns through these instances over time.

Don’t speak slowly as if you’re talking into a speech-to-text app, just speak normally and answer questions clearly for now. You may not set that appointment because it’s unsafe to do so without pre-qualifications, but you can set the appointment with the robot and immediately send prequal questions to the clients via text. That’s a win-win.