Home>Blog > Surfing the AI hype
AI

Surfing the AI hype

By -

Watercolor image created by Kjpargeter - Freepik.com

Every time a new hype appears people are trying to surf the wave and sell "old wine in new bottles". This article in HBR is a clear example of someone trying to surf the AI hype.

"How AI Is Changing Sales"

The examples given in this article have nothing to do with AI although in some cases Machine Learning algorithms will add value. They are all examples of (advanced) sales analytics that involve forecasting, predictive & prescriptive analytics and dashboarding that have been around for years. They are all examples of data driven decision making in sales that a lot of mature & modern sales departments are already (partially)leveraging.

I'll go through the examples used in the article 1 by 1

Price Optimization

"Today, an AI algorithm could tell you what the ideal discount rate should be for a proposal"

Price optimization is a combination of predictive modeling & optimization algorithms that work together to offer you a recommendation. This is also called prescriptive analytics. It has been in use in sales for decades. Think about airlines, hotels and others that are using yield models.

Does machine learning algorithms add value here? Definitely, because the better your prediction the better the result.

Does it make it AI? IMHO: NO!

Forecasting

"Using an AI algorithm, managers are now able to predict with a high degree of accuracy next quarter’s revenue"

The author mixes up 2 different types of forecasting: sales forecasting and revenue forecasting. Revenue forecasting is a problem of forecasting an aggregate (all revenue in a period): a time series problem. Sales forecasting focuses on individual deals and those aggregated will deliver a forecast: a cross sectional

Revenue forecasting has been around for a long time. Time series analysis for forecasting revenue can be traced back to the early 20th century. It has been extensively used by finance departments, marketing departments, investors, analysts, economists,.... New and better algorithms will improve the forecasts but they will not help sales leadership. It will only help their senior management to set their overall targets.

Sales forecasting is actually a prediction problem: which deals will close when and for what amount? Using predictive models a sales manager could benefit from this to identify deals that are likely to close in a quarter based on the data and therefore having a benchmark for is sales people's estimates. It will help focus the efforts. This type of pipeline analysis is rare but more advanced sales departments definitely use this data driven approach to help them better manage their results. Machine Learning algorithms can definitely help do improve these predictions and therefore the aggregate sales forecast.

Does machine learning algorithms add value here? Definitely, because the better your prediction/forecast the better the result.

Does it make it AI? IMHO: NO!

Upselling and Cross-Selling

"you can use an AI algorithm to help identify which of your existing clients are more likely to buy a better version of what they currently own (up-sell) and/or which are most likely to want a new product offering altogether (cross-sell)"

What is being described here is a 'recommendation engine' or a 'next best action model'. These types of algorithms have been in use for a long time. Think of Amazon, Netflix and others as examples these days but these types of recommendations were already generated in the 1980's en 1990's. Today's technology allows us to integrate them more tightly in our interactions with customers.

Recommendations are part of the predictive analytics approach. Machine Learning algorithms are definitely useful and heavily used. The increase in data and computing power have allowed for much more detailed and powerful algorithms but they are not new,

Does machine learning algorithms add value here? Definitely, because the better your recommendations the better the result.

Does it make it AI? IMHO: NO!

Lead Scoring

"With AI, the algorithm can rank the opportunities or leads in the pipeline according to their chances of closing successfully."

This use of predictive analytics is similar to the sales forecasting. Using a predictive model to determine if a lead will turn into an opportunity that is likely to close. For advanced and mature sales departments this is standard practice that is usually handled at the intersection between sales and marketing. MAchine Learning models are heavily used here and add a lot of value.

The author seems to imply that with this approach you will make decisions based on complete data compared to incomplete data in the traditional approach. Let me assure you: your data will ALWAYS be incomplete.

Does machine learning algorithms add value here? Definitely, because the better your prediction the better the result.

Does it make it AI? IMHO: NO!

Managing for Performance

"Using AI, sales managers can now use dashboards to visually see which salespeople are likely to hit their quotas along with which outstanding deals stand a good chance of being closed."

Building a dashboard for sales managers to use for decision making is the opposite of AI. It is providing input for decision making by Human Intelligence. At best it is using the results from predictive and prescriptive modeling (including very possibly the output from some Machine Learning models) to augment human decision making.

Does machine learning algorithms add value here? Probably in some of the indicators that are being displayed on the dashboard

Does it make it AI? IMHO: ABSOLUTELY NOT!

Top Stories