robot with a small smile. looks so innocent
PHOTO: Franck V.

“In the future, there will be two kinds of jobs. Either you will manage the machines or the machines will manage you. So decide today which you want to be.”

I couldn’t ask for a more attention-grabbing or clearer call to action. It came at the end of a presentation at MarketingProfs’ B2B Marketing Forum in San Francisco earlier this month.

The presenter was Christopher Penn and his session was titled “The Present and Future of B2B Analytics and AI.” Penn is co-founder and chief innovator at Trust Insights, a consulting firm that helps clients with marketing, artificial intelligence (AI) and analytics.

Building an In-House Competency in AI

Penn’s presentation showed how organizations can build an in-house competency in AI.

After the presentation, I thought, “Can’t we outsource AI competency to the vendors incorporating AI into their products?” After all, that is what we did at my last job. We were a software company of 100 employees, with a marketing team of four.

Related Article: 5 Tips for Attracting AI Talent

Predictive Lead Scoring for Marketing

We licensed predictive lead scoring software to segment sales-ready leads (those that should go into a salesperson’s queue) from leads that required more nurturing from marketing programs.

The software utilized AI and machine learning: It profiled our existing customers and predicted which leads would be most likely to convert to customers. As marketers, we appreciated the benefit of machine learning but understood little about the academic side: the math, statistics and algorithms, etc.

Following the conference, I followed up with Penn via email and he provided additional insights about the benefits of having AI competency in-house.

Digital Transformation and the Role AI Plays

Penn pointed me to an article titled “How to Achieve Digital Transformation,” by Bill Schmarzo, who is now chief technology officer for IoT and analytics at Hitachi Vantara, but published the article when he was CTO at Dell EMC’s global services big data practice.

Regarding digital transformation, Penn notes, “One of the key pieces that is so important is the idea of data and models and assets and digital assets becoming part of your business model.”

In his article, Schmarzo refers to digital transformation as the coupling of real-time data with modern technologies “to enhance products, processes and business decision-making with customer, product and operational insights.”

One of the modern technologies Schmarzo cites is artificial intelligence. Here’s a diagram from the article:

digital transformation

According to Penn, if your organization has no interest in digital transformation, then you can choose to do nothing with AI, or you can purchase AI-based solutions from vendors. He calls the latter option “offloading” — it’s similar to the way you offload management of a server farm to cloud providers.

If you believe digital transformation is a strategic imperative, however, “then you want to bring those disciplines and that AI expertise in-house. You want to know what’s in your model, you want to have fine, granular control over that model, and you want to be able to repurpose that model, perhaps even sell that model,” Penn says.

Related Article: 4 Keys to Improving Your Analytics Team

An Analogy for In-House AI Competency

Penn compares in-house AI competency to cooking. If you follow each literal step in a recipe, you’ll get whatever the recipe calls for.

However, he notes that “if you understand the science behind the recipe, and why some ingredients go together and why this pan is better than that pan, you can level up that recipe and eventually transcend that recipe into making your own dish that’s inspired by that, but it really is your own.”

When you purchase a vendor’s AI-based product, you get what the vendor provides out of the box. But you’re now reliant on the vendor for those capabilities. If you build the knowledge that forms the basis for that product’s AI capabilities, then you can better understand what the vendor is doing and potentially build your own down the road.

“When it comes time to renew or re-evaluate vendors, you’ll be on a much more firm footing to be able to say, ‘No, I want this, this and this, because this is what is standard or this is what is expected,’” Penn explains.

Going back to my example with predictive lead scoring, it may have been useful for me to find preliminary online courses to learn about the machine learning implemented by our vendor. The idea would not be to reverse-engineer the vendor’s product, but to become more conversant about the methodology. In addition, I could apply what I learned to other disciplines within marketing.

Related Article: Democratization of AI Development Is Beginning

Approaches to Learning

Along these lines, Penn offers the following suggestions on how to prepare your career for AI:

  1. Develop multidisciplinary skills.
  2. Learn to think like a machine (i.e., algorithmically).
  3. Develop skills for machine oversight (i.e., the ability to define tasks for machines and manage their output).
  4. Be outcome-focused.

To illustrate the idea of being outcome-focused, Penn offers this example:

deep learning with Keras cheatsheet

For a year and a half, Penn learned to code on “Deep Learning With Keras.” He made it through a small portion of the pictured cheat sheet.

Then he went to an IBM conference where they said, “We just made all of this drag-and-drop.”

According to Penn, “Now, instead of needing to know how to code, you just need to know what the Lego blocks do and drag and drop them in the right order to get the outcome you want. So be outcome-focused.”

Here’s a look at the drag-and-drop user interface, from Penn’s slide presentation:


Related Article: How to Keep Your Skills Sharp and Your Career Vibrant

A Human (Me!) Learning About Machine Learning

In addition to getting an opportunity to learn, one of the benefits of going to conferences is having your preconceived notions challenged.

I went to B2B Forum thinking that midsize companies need not build AI competencies in-house. I left, thanks to Christopher Penn, realizing the benefits of developing in-house AI competencies.

Dear machine: Tell me where to get started.