Artificial Intelligence. Machine learning. In 2016, these buzzwords were some of the most often used terms by vendors, but misunderstood by buyers. So what do these terms actually mean? While many use the two terms interchangeably, there are key differences.
Artificial Intelligence: At its core, AI is about creating machines that think like humans do. Today, that mainly has manifested itself in computer software that is able to automate simple tasks humans are good at. AI is a broad term to describe the technology of which buzzwords like machine learning, deep learning, and natural language processing are a subset of.
Machine Learning: Machine learning is a subset of AI. At its most basic, machine learning is the practice of using algorithms to parse data, learn from it, and then make a determination or prediction about something. The key here is the algorithm's ability to learn and change on its own without additional programming.
Every year, Gartner releases their hype cycle for new, emerging technologies. At the very tip of the curve or the “peak of inflated expectations,” lies machine learning. That means in 2017, AI and machine learning may enter the “trough of disillusionment” where people quickly realize that many companies may be talking about artificial intelligence, but few are able to deliver on their promises.
We’ve also seen our fair share of SaaS sales vendors touting their use of AI and machine learning within their products. So, how do you understand what’s real amongst all the hype? And what additional value can AI-enabled technologies bring for your sales team in 2017?
Use cases where AI has proven to deliver business value for sales teams
At its full potential, AI has the ability to transform the way your sales reps work and increase their efficiency. And the stats show there are productivity gains to be made. According to a study done by Accenture, only 34% of your reps time is spent selling and 57% of executives listed increasing sales effectiveness as one of the top 3 objectives in the next year.
AI-enabled technologies allow your reps to focus on the essence of what they are good at, having the best conversations they can have with your prospects. Instead of having to worry about the mundane tasks like identifying which prospects to work on, AI-enabled technologies will tell your reps which prospects to focus on or surface relevant knowledge for your reps automatically based on the conversations they are having.
While still a nascent technology, AI has become incorporated into enterprise software and there are already use cases in sales where it has seen success. AI-enabled technologies can potentially add value to your sales teams if they meet these criteria:
- Narrow domain: the best products using AI are doing so to automate a specific business problem. For example, 6sense uses AI to surface new leads and opportunities that are the most likely to close. They are solving a specific problem which is reducing the amount of time it takes your reps to find new prospects.
- Proprietary, unique data: Without unique data, even the most sophisticated machine learning algorithms are useless. Gong.io uses natural language processing to analyze your sales rep’s phone conversations and machine learning to uncover insights that improve how your sales team communicates with prospects. With access to potentially hundreds of hours of unique phone conversations, Gong’s machine learning algorithms can continually improve the insights they deliver to your business.
- Lives in your team’s workflow: To take advantage of the latest AI-enabled software your company brings in, it must be adopted by your sales team. The easiest way to drive adoption is to buy software that seamlessly integrates into your team’s workflow. X.ai is a personal assistant who is powered by AI. There’s no app or log-in info you need, all users have to do is simply CC email@example.com to have the bot schedule meetings for you.
Data, not algorithms are the true IP for AI-enabled technologies
At the heart of AI is data. It is the engine that fuels improvements in machine learning algorithms. Interestingly enough, the algorithms themselves may have little stand-alone value. Some of the largest companies in the world like Google, Microsoft, IBM, and Amazon realize this and have open sourced their machine learning algorithms. So, as we mentioned earlier, having access to unique, proprietary data is how companies using AI-enabled technologies will gain a competitive advantage.
For platforms like Google, Facebook, or Salesforce, collecting data is easy. So what kinds of data acquisition strategies can smaller startup employ? One opportunity not often talked about is leveraging technologies like browser extensions or chatbots to enable data collection outside of your native app. Companies with stand-alone apps limit their data collection abilities because they can only acquire data when users interact with their product. Since extensions live on top of your browser they have access to data (with appropriate permissions enabled) across a user’s entire journey on the web.
Breaking through the vendor BS
Since everyone is talking about AI, how do you cut through the noise and learn what the AI capabilities of a vendor truly are?
We’ve compiled a list of questions that you can use to understand whether the vendors you are evaluating are blowing smoke or actually understand how AI will impact your business:
Where does your training data come from and how do you use it?
While you shouldn’t expect vendors to spill all their secrets, it is a red flag if a vendor refuses to share where they gather their training data from. A vendor should be willing to share the internal and external signals they use to help train their algorithms, why they chose them over others, and how using these signals add value to your business.
How much training data does your algorithm need to produce trusted results?
Be wary of vendors who brush this question off or dismiss the importance of having enough data. Machine learning algorithms can only produce trusted results when they have a sufficient amount of training data. For example, for predictive lead scoring tools that rely on win/loss data, that could mean you need at least a year’s worth of data for algorithms to work properly. It’s in your team’s best interest to delay implementing an AI-enabled solution until you have enough data to work with.
How will your product scale with our growth and improve as it gathers more training data?
As you amass more data and grow, it’s important to understand how a vendor’s machine learning algorithms will scale with your growth. That means learning about how models are updated and how often they update them. Ideally, these models are personalized for your company’s specific needs and retrained when needed. This question should also give you a good gauge to understand whether the vendor has worked with other companies in your vertical.
Distinguishing between vendors who claim to use AI and vendors who can actually add value to your business with AI is difficult. Armed with this checklist, we hope you will feel prepared to properly evaluate vendors touting their AI capabilities and better understand how AI-enabled technologies can add value to your sales team in 2017.