Jack in the Inbox: 3 Ways AI Will Cure Modern Email Woes

Robert Lorenzen

That scrambled, overwhelmed feeling you carry around every day? That’s TMI, a text abbreviation that has become a legitimate medical condition, with symptoms ranging from TL;DR and ICYMI to AFK. Technology that makes life easier often ends up creating more work. It’s just how it works sometimes.

For a barometer of how barraged we are, look no further than email. Email is an ideal measure for how extraordinarily over-informationed we are. As of this year, people send 269 billion emails per day. Half of the world’s population uses email. You personally get at least a quarter of that in your inbox.

From a consumer standpoint, it’s pretty easy to eyeball and filter. Retailer offers and charities blare their intentions right to the trash bin, newsletters from print publications you no longer subscribe to give you a pang of guilt on their way out the door. In other words, it’s pretty easy to achieve Inbox Zero on your personal account.

At work, it’s a different story. There’s more noise to filter, especially if you’re a salesperson fielding 100-plus messages a day, looking for leads. You need to be ready when someone’s in a mood to buy, but that intention will get buried beneath the hemming and hawing of other prospects, managers setting up meetings, notes from coworkers about the office fridge. That modern condition, TMI, flares up.

Ghostwriter in the Machine

Software equipped with artificial intelligence can sort through your inbox like a personal assistant. Another way to say it is that it’s like a medical professional triaging your messages. With Natural Language Processing (NLP), an AI system can hunt through your emails for not just keywords but actual context. It can filter, based on importance.

What’s key with business emails is that there’s a time cost to this triage. You won’t know who’s requested pricing until you go through the messages, and that important one could be buried dozens down. If 74 percent of buyers pick the seller who provides value first, you’ve got to be zeroed in on the email that expresses interest in doing business.

We’ll soon see AIs that can identify and emails based on different categories, and even streamline responses or automatically respond for you. An AI-powered inbox will serve in three different capacities:

1. Prioritization

You need help sorting the various intentions of those messages, so AI will help by focusing your efforts where they’ll be most impactful. Let’s say someone requests a meeting, your AI will be able to analyze the sentiment of that message, recognize the intention, and put that message in a separate box. It’ll have categories for prospects who request pricing, or mention competitors, or are ready to buy. It’ll specifically look for those inflection points in the sales cycle in which speed of response means the difference between pushing toward a sale and losing the deal altogether.

2. Summarization

Implicit in the design of this system is the reality that no one’s reading more than four lines down. So while it’s very helpful that important emails are corralled in one place, the AI system needs to get to the point for each. You now know which ones are important; NLP will tell you why. Your “smart inbox” will scan through a long email, parse the meaning, and then give you a prompt up top: it’s a request for meeting, or pricing. But there’s always a next step. The inbox will stamp an icon on those emails you need to follow up on.

3. Recommendation

The clock’s still ticking. It’s time and labor-intensive to formulate a response to even the priority emails, so an AI-powered inbox would recommend next steps from a set of built-in responses. Click on the email from the prospect who asked for a meeting and your calendar will automatically open, allowing you to highlight your availability and fire it off. When the prospect picks one, you’re alerted. What can take a half-dozen exchanges is dealt with in a few seconds.

A meeting request is straightforward enough, but what if the prospect mentions a competitor? You have a decision to make: do you respond with a plug for your own product, try to get them on the phone, suggest a meeting? The AI is connected to the rest of the team and can see what has worked for other reps in previous scenarios. Your inbox will then suggest the next best action. If there are five possible responses, your AI will tell you which is most likely to move you to the next step in the sales process. It won’t take over, it won’t make the decision for you, but it’ll be sort of a weather report: 30% chance of rain, 60% chance of closing. The system will operate based on global insight that you won’t have on your own.

And that’s the fourth step: by capturing all these emails and their various responses and feeding them back into the CRM, the AI gets smarter. It learns how to improve its responses, what works and doesn’t. It makes the whole history of those interactions available to each salesperson so that everyone benefits.

AI is by no means a cure for TMI, but by cutting through the noise and prioritizing the important stuff, it will treat the symptoms, at least.

Robert Lorenzen is the senior manager of product marketing at Salesforce Sales Cloud.