Category: AI

Generative AI, Machine Learning, etc

Posted on

Causal Artificial Intelligence, with John Thompson

Casual AI with John Thompson

Episode 206

There’s no denying that ChatGPT and other GenerativeAI’s do  amazing things.

Extrapolating how far they’ve come in 3 years, many can get carried away with thinking GenerativeAI will lead to machines reaching General and even Super Intelligence. We’re impressed by how clever they sound, and we’re tempted to believe that they’ll chew through problems just like the most expert humans do. 

But according to many AI experts, this isn’t what’s going to happen.  

The difference between what GenerativeAI can do and what humans can do is actually quite stark. Everything that it gives you has to be proofed and fact-checked. 

The reason why is embedded in how they work. It uses a LLM to crawl the vast repository of human writing and multimedia on the web. It gobbles them up and chops them all up until they’re word salad. When you give it a prompt, it measures what words it’s usually seen accompanying your words, then spits back what usually comes next in those sequences.  The output IS very impressive, so impressive that when one of these was being tested in 2022 by a Google Engineer with a Masters in Computer Science named Blake Lemoine, became convinced that he was talking with an intelligence that he characterized as having sentience. He spoke to Newsweek about it, saying:  

“During my conversations with the chatbot, some of which I published on my blog, I came to the conclusion that the AI could be sentient due to the emotions that it expressed reliably and in the right context. It wasn’t just spouting words.” 

All the same, GenerativeAI shouldn’t be confused with what humans do. Take a published scientific article written by a human. How they would have started is not by hammering their keyboard until all the words came out, they likely started by asking a “what if”, building a hypothesis that makes inferences about something,  and they would have chained this together with reasoning by  others, leading to experimentation, which proved/disproved the original thought. The output of all that is what’s written in the article. Although GenerativeAI seems smart, you would too if you skipped all the cognitive steps that had happened prior to the finished work.

This doesn’t mean General Artificial Intelligence is doomed. It means there’s more than one branch of AI – each is good at solving different kinds of problems. One branch called Causal AI doesn’t just look for patterns, but instead figures out what causes things to happen  by building a model of something in the real world. That  distinguishes it from GenerativeAI, and it’s what enables this type of AI to  recommend decisions that rival the smartest humans. The types of decisions extend into business areas like marketing, making things run more efficiently, and delivering more value and ROI.

My guest is the Global Head of AI at (EY) Ernst & Young, having also been an analytics executive at Gartner and CSL Behring and graduating from DePaul with an MBA. 

He has written five  books. His 2024 book is about the branch of AI technology we don’t hear very much about, Causal AI. So let’s go to Chicago now to speak with John Thompson.

 

Chapter Timestamps

0:00:00 Intro

00:04:36 Welcome John

00:09:05 drawbacks with current Generative AI

00:16:09 problems causal AI is a good fit for

00:22:47 Way Generative AI can help with causal

00:26:50 PSA

00:28:08 How DAGs help in modeling

00:38:36 what is Causal Discovery

00:47:52 contacting John; checking out his books

People/Products/Concepts Mentioned in Show

John is on LinkedIn

John Thompson has been on Funnel Reboot twice previously:

Episode 136

Episode 181

Causal Diagramming tools:

https://www.dagitty.net/

https://cbdrh.shinyapps.io/daggle/

 



Listen to episode

Components of Causal AI modeling process.
Posted on

Marketing more efficiently with AI, with Rich Brooks

Rich brooks

Episode 205

 

Rich Brooks is founder and president of flyte new media, a digital agency in Portland, Maine.  He founded The Agents of Change a weekly podcast that has over 550 episodes. He is a nationally recognized speaker on using digital channels like search, social media and mobile for marketing to your audience. Rich also hosts the Agents of Change conference which takes place October 9th and 10th both virtually and in his hometown of Portland, Maine.

 

Timestamps/Chapters

0:00:00 Intro

00:02:49 welcome Rich

00:08:56 using GPT to make text seo-friendly

00:17:32 blending generative text with your own content

00:22:47 expanding to image & video

00:27:11 PSA

00:27:45 managing projects and events with AI

00:38:36 when to use a human vs aGPT

00:47:52 info on Rich. his podcast & his conference

People/Products/Concepts Mentioned in Show

The Agents of Change Conference, Oct 9-10, 2024

Rich’s Podcast – The Agents of Change

Rich on LinkedIn

Rich on Twitter/X



Listen to episode

Posted on

Building AI out of Data, with Yash Gad

yash gad

Episode 171

AI won’t end up being one thing, it will be present in many little applications – hopefully that will help us in our marketing. But what kind of AIs do we want? Are we looking at the ingredients that go into them? 

Those are the kinds of questions innovations our guest considers as he makes AI models for healthcare and the retail marketing sectors.

Yash Gad is a data scientist, education advocate, and foodie. Founder and CEO of RingerSciences and Chief Data Scientist of Next Practices Group. He earned his PhD from University of Illinois Urbana-Champaign in Computational Biology, Neuroscience &, Biophysics and received his undergraduate degree from Johns Hopkins. He joins me from Austin TX. 

People/Products/Concepts Mentioned in Show

Yash’s company: https://www.ringersciences.com/about

Yash also founded a consultancy, here’s the Next Practices Group’s site and their social feed.

Yash on Twitter

Frederick Winslow Taylor

Boston Dynamics

Tesla

Posted on

Episode 117: Marketing Artificial Intelligence, with Paul Roetzer

Marketing Artificial Intelligence

Paul Roetzer graduated with a journalism degree from the E.W. Scripps School at Ohio University and a few years afterwards he founded Ready North (formerly PR 20/20). In 2016 he founded the Marketing AI Institute. The idea for such an organization came from what Paul saw when AI began impacting his agency. He thought the only way marketers like him could work alongside AI would be by better understanding its capabilities. 

Part of their vision of educating marketers is through an annual event, and in 2019 they held their inaugural Marketing AI Conference. MAICON was on pause during lockdowns, but it came back in 2022.

In 2022, He and co-author Mike Kaput published the book we’re talking about, Marketing Artificial Intelligence. The book draws on years of research and dozens of interviews with AI marketers, executives, engineers, and entrepreneurs. He has also authored The Marketing Performance Blueprint (2014) and The Marketing Agency Blueprint (2012). Through his podcast and as a conference speaker, Paul makes AI approachable and actionable for marketers. 

He and his family live in Cleveland, Ohio. 

People, Products and Concepts in the Show:

Posted on

Episode 84: Data Science Wizardry with Richard Fergie & Brett Serjeantson

Do you own a crystal ball to predict the future? No? Well, I have good news for you. You don’t need one. 

When you feed lots of historicals into a machine, it doesn’t have to know what those numbers mean in business terms. Thanks to sophisticated data models, machines can determine the trends behind your numbers, including a pretty accurate prediction of what the next numbers in the series will be. 

That’s not all there is to data science though. Our guests believe it’s what takes our analytical abilities up the evolutionary ladder, from what happened, to how it happened, to how it might happen.

This episode talks about how cloud computing, which  crunches these numbers, has changed data science, doing everything that high-end servers used to do, at a fraction of the cost. It also covers the skillsets of the people needed to help machines work. To learn about how you can start implementing data science in your marketing, I’m joined by two people who are well versed in the field.  

Richard Fergie lives in the UK. He showed a penchant for math, which led him to Oxford University where he took his bachelors in mathematics. After spending time living in South America, he moved home and took a liking to the quant-heavy world of pay per click marketing. He has taken the 10 years of experience he got consulting on digital marketing and web analytics projects to found Forecast Forge, a quick and accurate forecasting tool that runs on Machine Learning.

Brett was an early SaaS pioneer and helped change the PR industry by creating one of the first companies to develop a real-time media monitoring and analysis platform that was able to successfully generate business intelligence from both traditional and social media.

Brett would lead the company as both the CEO and chief architect and developer of the platform as well as being awarded 2 patents.

Brett led his company through several significant transitions, including recognizing the need for big data and machine learning capabilities very early on as well as recognizing the importance of social media. 

Brett also identified the need for both journalist and publication data from the aspect of creating new markets as well as leveraging the data to further improve the analytics.

Brett’s company, MediaMiser, would be acquired in 2014.

Brett would later go on to improve his own analytical capabilities by achieving certifications in both data science and AI from the University of Waterloo and the University of Toronto.

Brett also has a bachelor’s degree from Western University and a diploma in PR and communications from Algonquin College.

Let’s go talk to Richard Fergie & Brett Serjeantson.

People/Products/Concepts Mentioned in Show

What you need to doTools that can do it
Visualize large datasetsGoogle Colab/Jupyter, Matplotlib 
Find trends, use predictions to decide next actionPomegranate
Do regression analysisPandas (built on NumPy)
unclassifiedProphet, PyTorch, TensorFlow

Episode Reboot. 

Check out https://www.forecastforge.com/learning/

Posted on

Episode 78: How AI Helps Marketers, with Cathy McPhillips

Lots of experts have said that AI is transformational to business. Call me cynical, but when you look at how the average marketing team functions in 2022, it doesn’t seem like much of our work has been transformed. My guest today, a mainstream marketer who hadn’t been influenced by  all the hype, will share her appraisal of AI’s practical uses.

Cathy McPhillips is a longtime Cleveland-based marketer who has worked in corporate, agency-side, as well as owning a small business. She worked for years with the Content Marketing Institute which was led by Joe Pulizzi (who was on Episode 56 of the show). Nearly a year ago, she joined the Marketing AI institute as chief growth officer, but admittedly with very little hands-on experience with AI. 

Brought on to grow the relatively new institute in a hurry, Cathy’s gotten up to speed on how to use these tools. She’ll tell us her story, including how the way we traditionally do things has to change in order to get the benefits of AI. Cathy lists many resources and tools, which have been compiled into a table that’s in the show notes.   I hope by listening to Cathy, you’ll try some of them out, and come away as excited as she is about how AI can be a superpower for today’s marketing.

People/Products/Concepts Mentioned in Show

Episode Reboot:

Tools to help contentE.G. PersadoDRIFTDescriptMarketMuseWriting social posts, messages (Copy.ai, persado) Personalizing landing pagesAnalyzing what’s said on social about our brand
Tools to help mktg activitiesE.G. Rasa.ioSprinklrOptimizing ads on campaigns to get most leads for least cost-summarizing/visualizing our web visitors and prospectsUnderstand which prospects have intent to buy, are open to contactData cleanup or other repeated tasks, learning from watching you

Also check out our other AI-related episodes:

Posted on

Episode 57: How AI Levels the Marketing Playing Field

If every part of your customer acquisition can be measured, you’ll figure out how to do it profitably. That premise has driven why digital marketing, and especially Pay-per-click (PPC) is managed by experienced humans. These professionals scour through data for the relationship between a company’s ads and the buyers actions; once found, budgets get shifted to achieve that optimal effect.  

A wrench has been thrown into our acquisition dreams by the ad platform titans: Google, Facebook and Microsoft (who own LinkedIn). Thanks to major AI investments they have made in the last five years, they’ve been able to automate much of the work that marketing professionals have done. In tandem with implementing their ‘smart’ software that runs autonomously, they have been restricting a marketer’s ability to manually control campaigns. 

The platforms believe their AI is smart enough to run marketing, so we can either be passive, letting them spend our money as they see fit, or we can choose to give them navigational assistance while they drive. The point is, you should have a game plan that works with the platforms’ AI. One that, over time, will generate the leads you need at the best possible acquisition cost.

I believe listening to this episode will give you that plan. It covers:

  • How marketing has become more computationally complex than humans can handle
  • What was in it for the platforms to automate PPC marketing
  • Stages of maturity for dealing with data, ending with predictive analytics
  • Why you shouldn’t fight ad platform automation, but instead use your business data to train algorithms how to market you more effectively
  • How you should integrate your in-house systems and apply data science to uncover insights

People/Products/Concepts Mentioned in Show

Paper estimating how much data optimized advertising requires, authored by Randall A. Lewis of Google; Justin M. Rao of Microsoft: “A calibrated statistical argument shows that the required sample size for an experiment to generate informative confidence intervals is typically in excess of ten million person-weeks”

Quote by Chuck Heamann & Ken Burbary in “Digital Marketing Analytics”:  “If you think about all the tools we have talked about…you see that there is one common denominator: You do not own any of the data. Herein lies what we think is the biggest revolution coming to digital analytics..companies will be building internal repositories for this data.”

Episode Reboot 

Go talk to a coworker who uses statistical measurement, to understand how the efficiency it achieves in other fields can be applied to marketing.

Posted on

Episode 47: AI For Marketers by Chris Penn – Summer Books

Marketers know that Artificial Intelligence is being integrated into our work, but many are unsure how they can apply it to their daily work. Our guest is Christopher S. Penn, the author of AI for Marketers: A Primer and Introduction, just out in its 3rd edition.

Our guest is an authority on analytics, digital marketing, and marketing technology. A recognized thought leader, best-selling author, and keynote speaker. He has been named by IBM as a Champion in IBM Analytics.

Chris is a cofounder and Chief Data Scientist of Trust Insights, a Boston -based digital analytics firm. He is co-host of the Marketing Over Coffee podcast. He has also run the marketing for a series of startups in the financial services, SaaS software, and public relations industries.

People, products and concepts mentioned in the episode:

Seven steps of AI Maturity:

  1. Data Foundation
  2. Measurement & analytics
  3. Insights & Research
  4. Process Automation
  5. Data Science
  6. Machine learning
  7. AI-Powered
reproduced with permission

Chris’ Social Profiles