Salespeople: How to Talk About Machine Learning

Josh Newall
Josh Newall

Published:

The hype around artificial intelligence (AI) is borderline manic, and no one in tech can stop talking about it. It also has everyone pitching it as part of their product line. When it comes to worldwide adoption, AI is still new. And the monsoon of marketing and sales campaigns touting AI-driven products has made it difficult for potential buyers to decide what’s fluff and what’s impactful.

As a Salesperson, we’re often the first to demonstrate technology to individual buyers. It’s tempting to throw around terms like AI, Machine Learning, Algorithms, and Deep Learning -- after all, everyone is, right? Don’t do it. Pushing magic may get the initial sale, but it also gets a quick churn.

Start by understanding the basics

There’s a lot that can technically be considered artificial intelligence. There are also varying definitions -- but most relate to machines performing cognitive functions. Many things can fall under that definition, from very rudimentary calculations to predicting the exact millisecond a multi-billion-dollar trade should be executed to maximize revenue gain.

There’s also a whole lot of fluff (and opportunity). As a salesperson, you’ll be incredibly valuable to a buyer if you can start by helping them clear that fluff away.

Get acquainted with AI and have a solid 30,000 foot understanding of the AI landscape. I’ll give you a 200,000-foot view in this post, but I suggest digging in more elsewhere.

Artificial Intelligence: Machines performing cognitive functions

Machine Learning: A subset of AI -- self-learning algorithms that detect patterns and anomalies. They get better as they consume more data (but that improvement will eventually level off, and more data will not lead to an increase in performance). You don’t necessarily need massive computational power to do this.

Deep Learning: A subset and more advanced stage of Machine Learning. This is the true arena of big data, and most companies you’re selling to won’t be here (at least not yet) because they lack the sheer amount of data necessary to make it worthwhile.

Deep learning is also best applied to problems centering around perception and language. To keep this simple, think of the concept of machine learning -- but when it comes to deep learning, there’s no ‘ceiling’ of improvement based on data ingestion.

It continues to get better, and the only limitation is your ability to feed it more data. Here, you need massive computational power.

I would focus your time on Machine Learning — it will most likely have the quickest impact on your prospect’s business. Keep in mind, Machine Learning is a version of AI, just as Deep Learning is. Deep Learning will undoubtedly become more accessible over time, but for today, Machine Learning will likely be most impactful based on the amount of data/investment required.

Understand the business implications of Machine Learning

Machine Learning does two things: It senses and predicts. This applies to many things, such as:

  • Detecting anomalies to prevent fraud
  • Identifying high-level attributes to predict which leads have a higher likelihood of closing
  • Discovering issues when manufacturing products
  • Optimizing product distribution to meet specific market demand
  • Personalizing content to specific people
  • More efficiently managing just about any type of organizational risk
  • Predicting churn or lifetime value of clients

This list can go on, and the whole point is to show when it comes to Machine Learning, there are few limitations to where it can have an impact. Which brings me to my next point …

Set expectations and start simple

It’s easy to promise the world to someone. It’s especially easy to promise the world when it comes to Machine Learning. Why? Because there are few limitations, and you can tell your prospects its impact will be utterly transformative -- and you wouldn’t be lying.

In theory, these algorithms offer to vastly improve current processes and align your organization for future opportunity.

In theory.

So, the question is, what’s the one thing stopping that theory from happening? Your customers’ data. If they don’t have a large amount of valuable proprietary data, then that super awesome machine learning algorithm isn’t going to have much of an impact.

Machine Learning heavily relies on clean, accurate data.

Set those expectations with your prospect. If you can help them build a foundation that will allow them to take better advantage of Machine Learning, then you’re not just pitching a product, you’re adding a ton of value to their business.

Learn your prospect’s data strategy, and build a use case around it

As a Salesperson selling Machine Learning technology, you’ll likely learn how people use that data every day. There’s probably no one better positioned to consult on data strategies than someone who has insight into how dozens of other companies do it.

Take those learnings, better understand the ins and outs of data collection, cleaning, and management, and help your prospects and clients apply that knowledge.

Understand data strategy

Quick tip: Use Kaggle to research different machine learning competitions. These competitions will help you understand what business problems people are trying to solve with data and what types of data they’re using.

Heading into calls with your prospects with a better understanding of data strategy will help you sell products that are driven by Machine Learning. You’ll be better equipped to see the different ways your product can impact their business.

Limitations will rarely fall on the end of the actual Machine Learning algorithm. Instead, it falls on the customer’s ability to supply valuable proprietary data to put into the algorithm.

So, instead of promising your prospect huge ROI from Machine Learning technology, start by understanding their resources. You might have amazing technology and an itch to pitch your highest-level products, but if your prospect doesn’t have the data, it’s all a wash.

Prospect doesn’t have a data strategy? Help them build one

Here’s the thing -- creating a data strategy shouldn’t be too hard at first. For many companies, it’s as simple as adopting a CRM and collecting basic data on your prospects and customers. You could start with engagement data, product usage data, demographic data, competitive intelligence, or transactional data.

There are plenty of free tools available that enable the collection of this data. Start by helping your prospects identify which data points will be more valuable than others in relation to what they’re trying to solve for.

Help them formulate a solid strategy by identifying data that will have a big impact on their venture into Machine Learning. And show them how that strategy will lead to your technology, giving them better results.

Refrain from promising the world, and help your prospects understand that machine learning is only as transformative as the data you can put into it. It might lead to a longer sales cycle and more educational calls, but, in the end, it will also lead to a happier customer that genuinely values not only your product but your advice as well.

This piece first appeared on Medium and was republished here with the author's permission. 

HubSpot CRM

Topics: Sales Trends

Related Articles

CRM software from startup to enterprise.

GET FREE CRM

Powerful and easy-to-use sales software that drives productivity, enables customer connection, and supports growing sales orgs

START FREE OR GET A DEMO