Everything you need to know about sales automation

May 08, 2019 by Jessica Munday


What is sales automation?

Sales automation is when an organization uses technology to automate sales processes through static roles. For example, converting leads into the next stage in the CRM based on triggers that occur elsewhere like sending out certain documents through email.

Sales is expensive and inefficient

The most expensive department for many businesses is its sales team. Some companies spend up to half their revenue on sales alone. Salesforce, for example, spends 53 percent of its $10 billion in revenue on its sales team.

This becomes a major problem when reps are inefficient, which is unfortunately fairly common:

  • 67% of sales reps miss annual quota
  • 58% of deals are stalled due to lack of value communicated about the product
  • 90% of marketing content goes unused by sales reps
  • Sales reps spend more than 8 hours a week searching for information
  • 30% of leads drop out of the pipeline
  • 59% of their time is spent on administrative tasks
  • Sales reps perform 300 CRM updates every week
  • Sales reps send 600 emails every week
  • Sales reps spend 20 hours a week writing emails
  • 40% of those emails are repetitive

By nature, sales people are extroverts. Referring to the Myers Briggs personality test, most reps are ENFJ types which means they enjoy working with people but are data and process adverse.

Yet, sales leaders have created a repetitive sales process and force them to become data-entry robots.

AI sales automation is the solution

IDC predicts that between 2017 and 2021, AI-powered CRM activities will boost revenue by $1.1 trillion. And this research is solely focused on sales process analysis like “next-step” coaching and lead scoring; it did not include sales automation capabilities that can reduce repetitive tasks. Given that sales automation wasn’t considered in this calculation, it can be assumed that its impact will not only boost revenue beyond IDC’s predicted amount but also exceed expectations about overall efficiency.

AI assistants are already widespread in the personal lives of individuals. According to Accenture research, 46 percent of U.S. consumers currently use a “voice-enabled digital assistant.” And by 2021 there will be more AI assistants than humans on earth.

Our obsession with AI assistants will soon bleed into our work lives and we predict that in 10 years every information worker will have a business AI assistant to support them in completing their daily tasks.

Background and breakdown of AI

To give a proper definition, artificial intelligence is when a machine exhibits human-like intelligence in approaching a problem. This is achieved by training it with data to follow the same processes humans go through to solve a problem.

Just to be clear, we have not achieved artificial general intelligence; this would be the robocops and Terminator-like AI that many are afraid of. But we have achieved artificial specific intelligence, meaning technology that can perform very specific tasks.

Knowledge-based AI has been around the longest. These machines have a decision-tree that’s made up of “if-this-then-that” rules which dictate which outcome the AI can generate. The shortcoming with this type of AI is that assembling these decision-trees often require large teams and take many years to build.

Within the last 10 to 20 years there’s been major developments in machine learning (ML), specifically in traditional ML. These types of systems are built with specific algorithms to solve very specific problems. Some are good at speech recognition, others at predicting which ad people will click.

Branching off of ML innovations is deep learning, which is where the biggest and most recent breakthroughs in the AI space have taken place. Deep learning models have neural networks that mimic the human brain in the way that they are interconnected and can fire between neurons. These give deep learning systems enough storage and computing capabilities to run much larger networks that can be layered on top of each other.

One example is facial recognition. There are multiple data points that make up a person’s face and each computation layer divides up those features. One layer may be edges, another may be shapes and the last focused on high-level features. By interconnecting all of these features from the input layer, it can create a certain output.

Historically, machine learning capabilities tapered out at a certain threshold of data. No matter how much data the machine was trained with, it wouldn’t get any better at computations. The opposite is true with deep learning models as they continue to improve the more data they are given. These developments made it possible for deep learning models to make significant leaps in several areas such as natural language processing (NLP).

Intelligence tools for sales all rely on these types of learning models to automate and augment specific aspects of the sales process. And while sales AI can be extremely helpful, it is also is a big buzzword. We suggest taking the time for a deeper dive and double-clicking on their models before purchasing a sales AI tool.

Sales AI market overview

We’ve identified six key areas for AI to improve the sales process.

1. Sales automation

This is where Automation Hero falls. Sales automation is when a machine or tool can perform a function with minimal human involvement. AI sales automation handles repetitive daily tasks without a sales rep having to waste their own time.

Something incredibly important to mention is that we’re far away from completely automating the sales rep. The buying experience is the biggest brand differentiator and sales will always need the human touch.

2. Predictive engagement

Predictive engagement focuses on sales AI tools that can help guide reps in the right direction. Think “next-step” analytics that coach them on what to do with a certain lead/prospect next. This could be suggesting they send certain pieces of content or reminding them to follow up on a certain day and time.

3. Predictive prospecting

Essentially predictive prospecting helps sort through opportunities and leads. Prospecting tools can automate the lead scoring process and find more potential customers.

One of the biggest struggles for salespeople is the prospecting phase as interest and qualification vary between leads. AI uses data to bring relevant leads to the attention of sales reps and assist in prioritization.

4. Sales process analytics

While these tools focus on more traditional analytics, they still bring a ton of value to sales organizations by analyzing rep behavior. The tools then generate reports based on KPIs or offer suggestions to improve processes. Basically, these keep sales leaders in tune with what activities their reps are focusing on and assist in optimizing the sales process.

5. Voice/Text analytics

These tools observe and report on sales reps’ conversations with customers to improve communication skills and increase conversion rates.

One downside to these tools is that the customer legally needs to be informed that the conversation is being recorded, which may deter some opportunities.

6. Chatbots

Chatbots are designed to replicate human conversation using NLP and are traditionally used to help with customer service or prequalification.

However, it’s important to note that chatbots aim to replace parts of the sales rep process.

The time is now

There’s a cycle of innovation that occurs in waves called the Schumpeterian Cycle of Innovation and Entrepreneurship. This theory discusses the waves of technological innovation and explains that as time goes on these waves are compressing.

For example, the first wave of innovation (water power and textile irons) lasted 60 years, while the fourth wave, the latest innovation cycle, was only about 40 years.

According to this theory, early adopters of the latest technology have a significant advantage over those that wait to implement. The peak of the wave displays the point at which a company would be ahead, and the valley represent when failing to implement would be a disadvantage. As the waves contract at an increasing rate, it becomes even more imperative to adopt these new technologies and do so quickly.

AI innovation is accelerating at an even faster pace than previous waves. The stakes are substantially higher for companies to implement AI into their business processes. Companies are already adopting these technologies at a high speed and there’s a winner-takes-all situation here. Those that invest in the right opportunities early on will far outcompete others in their market. Automation Hero predicts that this wave of AI implementation will be about six years.

AI Sales Automation examples

Below are three real-world use-cases for sales automation that can bring real value to the sales organization.

1. Cross and upsell.

These types of models can help identify triggers that could lead customers to upgrade or purchase complementary products.

Let’s use insurance as an example. If a customer recently bought a home, a rep could offer them home insurance in addition to the life insurance policy they already have. If someone just welcomed a new addition to their family, a rep may want to review the current family policy and upgrade it.

2. Intent detection and sales automation.

Intent detection relies on NLP and deep learning to understand the intent of written communication and then automate the response or action.

For example, an email could be forwarded to the correct department or a predefined response to a customer inquiry could be drafted on behalf of the sales rep, which could dramatically accelerate turn-around times.

3. Dark data extraction.

Many organizations have unstructured and structured data (e.g. email or Excel spreadsheets) stored in decentralized systems. With dark data extraction, the important information is pulled from the various docs and sorted into the correct reporting system.

This is the case for many sales organizations as sales reps utilize spreadsheets and PDF documents instead of logging customer data into their CRM system. The result is significant amounts of dark data that could not be utilized for impactful business decisions.

Suggested Next Reading