Cutting Through the Hype and Harnessing the Potential of AI
Editor’s note: Automated speech-to-text transcription, edited lightly for clarity.
Throughout 2023 SurveyMonkey conducted a few surveys with people from different professions across the U.S. They also surveyed 400 marketers and researchers about their usage, hopes and concerns with AI.
In their webinar on December 13, 2203, they not only shared some of the results from these surveys, but how professionals are using AI today in the workplace.
SurveyMonkey’s Divya Shroff, senior product marketing manager, and Zoe Padgett, research scientist will walk you through all the insights gleaned from their surveys.
Watch the full video or read the transcript below.
Webinar transcript
Joe Rydholm:
Hi everybody and welcome to our webinar “AI and Marketing: Cut through the Hype and Harness the Potential.” I'm Quirk’s Editor, Joe Rydholm.
Before we get started, let's quickly go over the ways you can participate in today's discussion. You can use the chat tab to interact with other attendees during the session and you can use the Q&A tab to submit questions to the presenters during the session. We'll answer as many as we have time for during the Q&A portion.
Our session today is presented by SurveyMonkey, enjoy the webinar.
Zoe Padgett
Hi everyone. Welcome to our webinar, “AI and Marketing: How to Cut through the Hype and Harness AI's Potential.”
I'm Zoe Paget and I'm a Research Scientist here at SurveyMonkey.
Divya Shroff
And I'm Divya Shroff. I'm a Senior Product Marketing Manager for AI and Enterprise here at SurveyMonkey.
Zoe Paget
We've all heard a lot about AI, especially since the launch of ChatGPT. AI is creating a lot of excitement and conversation, but it's also being met with some skepticism and some people are even seeing it as sort of an ominous presence that's hovering in the background and about to pounce on us. But down in the center of that conversation, I think you'll see where people are really trying to figure out the future with AI. And for anyone who works with a creative team, there's this meme about creatives being replaced. So maybe we don't have to worry so much.
One of the benefits that I have of working at SurveyMonkey is that my team basically has endless potential to field surveys. So, if we're curious about anything, we can pull together a survey and find out what people think about it.
Obviously, this year that has meant really trying to keep a finger on the pulse of AI and capturing authentic input from people in, what we call, the “gen pop” or the general population about how they're seeing AI show up in their everyday lives.
So, to set the stage a little bit, here's a chart showing the percentage of people across the U.S. by generation who expect AI to have a major impact on their lives in the next five years.
We can see that 59% overall, say they think AI will have a major impact on their lives, and that number really doesn't differ much across generations, 59% of Gen Z, 57% of Millennials, 62% of Gen X and 58% of Boomers. Overall, only 10% of people think that AI will have no impact on their lives at all.
People are really preparing themselves for some big changes in their lives over the next few years and seem to be expecting some major breakthroughs.
So, in April and then again in September, we asked how often do you hear people talking about AI?
There really is sustained type about AI. In April, we had 37% say they were hearing about AI daily or weekly, and that number was 36% in September. So really no change.
I think these numbers show us that the chatter around AI that we were seeing this winter and spring around the ChatGPT release, it hasn't died down and AI is still top of mind for many Americans. So, there's still plenty of opportunities to join the conversation.
Just looking at some of those demographic differences, we do see large differences by age in how much people trust AI and how much AI will affect their own futures.
On the left, we can see that the youngest generations have more trust in the accuracy of AI generated content and as age increases, distrust increases as well.
Then on the right side, we can see again as age increases concern about the future of what AI will bring increases and excitement about AI actually decreases.
We've seen from our survey results, so far, that AI is really being talked about a lot in conversation and we know that people really care about AI and how it's going to impact their work.
Let's go ahead and dig into more data on that topic.
The data I'll share in these next few slides all come from our CNBC SurveyMonkey Workforce Survey where we poll more than 8,000 workers across the U.S. twice per year.
This one was done in May and one of our focus areas was on AI. You can see here, represented by that dark green bar, that overall, 43% of workers expect their jobs to change significantly in the next five years due to disruption from AI. And every field above that line has a higher-than-average expectation for disruption due to AI with advertising, marketing and technology at the very top. At the very bottom we have personal services and government who don't think that they'll be impacted too much by AI.
Of course, the big concern is, ‘Will AI make our jobs obsolete?’
We asked people how worried they are that AI will soon make their jobs obsolete. Overall, just 24% of workers said that they were very or somewhat worried about AI making their jobs obsolete. Those numbers are higher for lower income workers, workers of color and younger workers. But still overall the majority are not concerned that AI is going to make their jobs obsolete.
So, more broadly, whether people have to use AI for their specific job or not, we know that AI is making inroads in the workplace already.
One area where AI is already used and used somewhat widely is in HR. AI can help work through resumes, identify low performers and helps to schedule shifts. However, most workers of all industries and fields say that they're uncomfortable with AI being used to assist with these standard HR tasks.
And, getting back to that trust issue, 42% say that if their company did make greater use of AI in HR, they would start to trust HR less.
Okay, that's it for work. Now I'll get into what you probably came here for, some specifics around how marketers are using AI at work.
As of July, we're seeing that 36% of marketers say that they're using AI daily or weekly for work, and only 5% say they've never used AI for work.
79% say that AI is becoming more important, that it's more important this year than it was last year, and marketing leaders seem to agree. 85% of marketers said that their leadership has invested in AI training.
Main things that marketers are excited about when it comes to AI are that it will allow them to focus on high level tasks and automate those more repetitive tasks. 60% of marketers said they were excited about those things along with enabling faster decision-making. They're also pretty excited about reaching new audiences and markets, new experiences for customers and more personalized offerings for customers.
We also asked about where AI is showing up in marketing organizations. And 25% say AI is only used within standalone tools. Those are things like chatbots and e-mail automation. While 34% say it's only used for backend integrations like product recommendations, digital ad buying and predictive analytics. And another 35% say it's used in both. In all, 60% are using standalone tools and 69% are using AI on the backend.
This one's pretty amazing. 73% of marketers say that AI helps them do their job better. And 63% think that AI can actually do better than human marketers when it comes to gaining insights from the data.
They're more torn on whether AI or humans can do a better job recommending products. AI has a slight edge there and whether AI or humans can do better at understanding their organization's target audiences. And that one's a pretty even split.
It's important to note here, I think, that we're talking to marketers, not any detractors of marketing.
There are some barriers that marketers face when it comes to using AI tools.
The main one is a lack of knowledge to effectively use AI. 41% say that's a challenge for them or their team when it comes to using AI.
Some other challenges are integrating AI within legacy workflows, a lack of data or data quality to actually feed into the AI applications and the fact that AI is unable to fully replace their workflows. It seems like marketing leaders are already preparing to address that knowledge gap. That's a barrier for integrating AI.
85% of marketers say that their leadership has already invested in AI training programs. 30% say they've invested in paid training programs, 25% say they've invested in free training programs and 31% say both paid and free training programs. So overall that's 61% that have had paid training, which is a pretty big number.
We've already covered a few fields and shown how AI's impacting the general public and the workplace as well as marketers.
So, now I want to talk a little bit about how AI is impacting market research. In these next few slides, I'm going to walk through a study that we fielded in June where we surveyed market research professionals. This is a little bit more niche, but I think it's a topic that will definitely hit home for many of you and maybe some of you are market researchers.
We found that a quarter of market research professionals say they now use AI on a daily or weekly basis to do their jobs. So, a little less than marketers, but still a significant portion. And on the right, you can see that the main tasks are data analysis, data visualization and developing data-driven recommendations, which makes sense because AI is very data-oriented.
Here are two key findings from that study. Most market researchers say that they have both support from their leadership at a theoretical level to use more AI powered market research tools and that they have literal monetary support to do so that their company is actually likely to pay for AI enabled market research solutions.
That's why the AI market is so hot right now. Companies are maybe cutting back on their budgets, but they're still willing to spend on these AI powered solutions.
The results at the bottom, I think, show why they're willing to do that. Most market researchers say that using AI enabled market research would make their insights less biased, more accurate and faster.
This slide shows how much market researchers are really clamoring for AI powered tools. 54% say that it's important for a survey platform to integrate AI, and that's very important for us here at SurveyMonkey. A majority say that it would be useful for them to have AI powered tools both in the survey design process as well as on the analysis side. The front and back end of the research process basically.
We asked about specific tasks that a market researcher might be responsible for and how much they think that AI will be involved in those tasks over the next five years. 38% expect AI to pretty much take over the data analysis task in the next five years, and that's the area where market researchers expect the most disruption.
Data visualizations are about the same as data analysis, writing survey questions and brand tracking a little less than data analysis and then developing data, data-driven recommendations, localization and concept testing are the lower ones.
In general, perhaps because of the results that we just shared, market researchers are actually pretty optimistic about the effect that AI will have on their jobs and on the industry overall. Almost four in 10 say that AI will help more than it hurts them personally.
The same is true when we asked about whether it'll help or hurt their company as well as the broader market research industry. And in their own words, here's what some market researchers are most excited about when it comes to AI:
- “The accuracy and reliability of the data.”
- “The opportunity to decrease biased opinions.”
- “Spotting trends that they might've missed previously.”
- “Having customizable and personal data.”
The things that they're concerned about are:
- “Fraud.”
- “The data being purposely skewed.”
- “Inaccurate results due to bad AI programming.”
- “Failure to account for intangibles in research like emotions."
- “Privacy.”
Divya Shroff:
Now let's turn to a fun topic, which is all about AI versus humans. Specifically, we'll be diving into a few areas where AI is expected to perform better than humans and a few others where humans are definitely preferred to AI.
In the study from earlier this year, we really pitted man versus machine, humans versus AI, and we found that most people prefer humans over AI, in general customer service tasks.
Very directly we asked, “If the end result was the same, would you prefer interacting with a human or an AI customer service agent?” We found that 90% prefer a human for follow-ups among each group.
We then asked why they preferred the human versus the AI. You can see here that people who preferred a human customer service agent said it was because a human understands their needs better, provides more thorough explanations and is less likely to frustrate them.
People who preferred AI better said it was because AI has better availability, addresses their issues faster and provides more accurate information.
Here's another fun one, and this might actually be my new favorite way to visualize our findings on AI versus humans, so far. You're all probably familiar with NPS, but just in case NPS stands for Net Promoter Scores. Net promoter scores are a standard way to measure customer satisfaction in pretty much any CSAT or customer satisfaction survey that you take after you stay at a hotel or try out a new restaurant or make any purchase.
In general, you'll probably get asked some version of this question, “How likely would you be to recommend this product or service to a friend, family member or colleague?”
The answer scale goes from zero to 10 and NPS categorizes respondents into three groups based on their answer choices. Anyone with a score of zero to six is called a detractor. Anyone with a score of seven to eight is passive and a nine to 10 are called promoters.
You can see here that we asked people to think about a time recently when they interacted with a human customer service agent and a time when they interacted with an AI customer service agent. The humans had an NPS score of six and the AI had an NPS score of minus 66. That's a 72-point difference right there, and this is huge.
Specifically, when it comes to product recommendations, younger generations are more bullish on AI as you probably expected. Now this isn't super surprising given their higher familiarity with getting product ads based on algorithms. They're used to it. So, they are definitely more optimistic about AI and product recommendation.
And when we ask follow-up questions on why they thought humans or AI would provide better product recommendations, we learned that humans are expected to relate to our needs better. So, we expect humans to have more authentic in-person knowledge, which allows them to understand and act on our feedback.
However, AI recommendations are trusted to provide more precise and faster suggestions for us. No surprises here, but it's also seen as having past knowledge of your preference and shopping habits. And because of that, AI is seen as providing more unbiased suggestions.
Now, if we start to expand beyond the customer service and product recommendations area, we can see that of all of the industries, there's probably the most support for AI in the customer experience space.
Despite everything I just shared on the last few slides showing that AI has a lower NPS score than humans, that people feel more negatively about AI and CX, perhaps there's still a lot of willingness to engage with AI in the customer experience space, especially for tasks like ordering food and drinks or returning an item much more than, for example, in the health care, finance or education spaces.
But health care might very well be next. Many see AI as having a huge potential to disrupt the health care industry.
So, with all of that data that we've taken a look at today, let's dig into what are some specific opportunities for marketers and market researchers specifically – What can you do with AI and where are there still some limitations?
I'll start with market researchers. As shared earlier, market researchers are actually quite optimistic about AI in their roles. Some of them are already using AI and they see the value that it provides, especially when it comes to helping them create surveys as well as data visualization and analysis.
In particular, market researchers say that using AI enabled research tools would make their insights less biased, more accurate and a lot faster.
That's why we've started to build AI capabilities right into our SurveyMonkey platform. And you'll see these across the entire journey that you'll go through as a user, right? From survey creation to data collection and analysis, we've got AI and machine learning built in, and all of this falls under our SurveyMonkey Genius umbrella.
You can really think of SurveyMonkey Genius as your own personal survey expert that uses AI and machine learning to help you create better surveys, gather higher quality data and to spot insights quicker. The idea being that these features that we've built in will help you get unstuck when it comes to survey creation while making it quick and easy to make sure you're building a high-quality survey.
In fact, our latest feature is called Build with AI, actually uses GPT technology to automate survey creation. All you have to do is write a short prompt on what your goals are from your research and it'll create a whole new survey for you.
We also want to make sure that we're improving the quality of your insights, and this starts with combing through the actual responses to your survey. And then what we do is identify and flag anything that doesn't meet quality standards. In case there's a respondent that's really speeding through your survey or clicking A B, B, B for every single question, we'll flag that user and make sure that response is set aside, and it doesn't influence your data analysis.
Finally, we also want to reduce the manual, time-consuming work of data analysis. And what we can help you do is find patterns in your data. SurveyMonkey Genius will actually sift through all of your data to identify what the top segments that you might be resonating with or how do you categorize open ends based on sentiment.
If you're interested in learning more about any of these capabilities, feel free to reach out to us directly.
And I did mention Build with AI, which is our latest capability. I'd love to give you a short little demo and show you how Build with AI can help you collect insights faster using AI powered survey generation. And again, what this does is it automates your survey creation process based on your specific feedback goals, and it'll really help deliver a custom survey to you in just minutes.
To give you an example, I'm a marketer at SurveyMonkey, a software company, and let's say I want to try and get some feedback on my website. It is one of the most important marketing channels that we have here. Let's see how Build with AI can help us create a survey to get feedback on our website.
Audio from video demonstration:
Introducing our newest AI feature called Build with AI, which allows you to generate a survey instantly using GPD technology.
To begin, click on the Build with AI button on the create a new survey page. Here you'll see we have a set of pre-draft expert prompts for the most common survey use cases. You can reference these examples of what a prompt looks like, or you can choose one of these as a starting point and modify it as needed.
Alternatively, you can type a short description of your survey goals into this text box. In this case, I work for a software company and I want to get feedback on our website and how people are interacting with it.
So, I type that in here and I click on this button to generate my survey.
And here we go. We have a full survey that has been generated based on our survey goals.
Now you can either click “Try another prompt,” which takes you back to the previous screen to make changes, or you can click on “Use survey.”
This takes you to our survey design tab where you can then make any edits or customizations to your survey. And finally, you can send it out.
Divya Shroff:
As Zoe pointed out, 60% of marketers are excited for AI's help with repetitive tasks, and 60% are also excited about AI allowing them to focus on higher level tasks. Hopefully that demo video you just saw of Build with AI got you excited about the possibilities of AI for your role as well.
Right now, AI is great at the work that humans often struggle with. Things that are rote but still require attention to detail, synthesizing multiple inputs into an easy to recognize pattern or lifting you up high enough to get unstuck from a blank page.
These are the kinds of things that AI does really well, so it can help reduce the overall cognitive load for you, but it doesn't completely take it away. Let's talk a little bit about using AI for content creation because this can get a little bit tricky sometimes.
Now, mostly marketers are using this for customer facing copy, and almost a third of them are using it for SEO copy. And then because humans are complicated creatures, a full 75% of marketers don't want AI generated content to be helping anyone with their search ranking.
So, what's the story telling us?
Now, for anyone who was working on content, web or SEO in 2011, you might recall Google's Panda Update when thin or poor content was weeded out from the results. Now, I believe that we may have a similar event occur with AI generated content as well. After all, something that's made by a machine should be detectable by another machine, right? And if not now, then eventually it probably will.
And what Google did with their algorithm update was a clear signal you must provide value, not just beat the algorithm. This is a north star that every content producer and website builder should be following, because I think one day, we will find out that content with too much of an AI quotient will actually be punished by search algorithms. I don't think it's just possible, but I do think it's somewhere between probable and inevitable in the future.
So, if we know that AI is limited in its current capacity and potentially could threaten important metrics like search results, what's the game changer that makes it more universally applicable and relevant to you as marketers?
To get there, we really need to see what's missing. And here's where I see the limitations of AI at the highest level.
AI can inspire, but it isn't inspired. What I mean by this is that AI doesn't understand the human experience. It puppets it, it recreates it. But until AI can truly, truly experience the human condition and truly outline that, we as marketers are always going to need inputs from human beings to augment, to round out, to give context to and to ultimately create the full story for what AI can only outline in the most basic ways.
We need to keep our fingers on the pulse of what people are thinking and what they're feeling. And at this point, traditional AI can't do that effectively. But those authentic inputs and actually hearing from people is how you can create winning strategies.
So here are my three takeaways for marketers and market researchers.
One, be a little careful as AI is great for straight to the point and transactional copy, but sometimes it will get the details wrong. And on, creating surveys in particular, AI is only as good and effective as the prompt and the description that you give it.
Two, stay on top of trends with AI so you make sure that the more menial and mundane tasks are getting completed by AI and not by the intern who has a gangbusters idea for your TikTok.
And three, bring together AI with authentic inputs as that is how you unlock extra power from both. You have to know what people want, what they need and what they expect from you in order to be a solid marketer and to feed into that ChatGPT scenario where you might actually ask AI for help without that authentic input, we're no better than mindless bots.
So, when you're looking to gather those authentic inputs from real humans, whether they're your prospects or customers, or even employees, SurveyMonkey Genius can help. It leverages AI and machine learning to help you create better surveys, gather higher quality responses and to spot insights quicker.
Thank you so much for joining today. We really appreciate it and hope to see you again sometime soon.
And with that, we're happy to take any questions that you may have.