• Explore. Learn. Thrive. Fastlane Media Network

  • ecommerceFastlane
  • PODFastlane
  • SEOfastlane
  • AdvisorFastlane
  • LifeFastlane

The Conversational Effect: How AI Increases Sales And Handles Support In ECommerce

Summary

  • The first chatbot and The Eliza Effect
  • Evolving from Transactional to Relational Commerce
  • Today’s AI and The Conversational Effect
  • Two real eCommerce brands, two remarkable results

Intro

I’ve witnessed some great technological advancements in my short time on this planet, but nothing like AI. I remember getting my first computer—a Commodore Vic20—and going “online” for the first time in 1984, using a detachable modem and having a 2400 kb baud rate (I could be wrong on these numbers, but that’s what memory is telling me).

From bulletin boards, to the internet, to the world wide web, to social media, to blockchain to now the latest trend: AI. But nothing will compare to AI as far as its impact and implications on modern society and culture—absolutely nothing. That’s quite a statement considering the impact of social media on today’s culture, but I’ll save that for another conversation. Today, let’s talk about how AI will change everything about eCommerce as we know it.

But first, let’s talk about the first chatbot ever: Eliza.

The Eliza Effect

Joseph Weizenbaum was an MIT professor and creator of the first chatbot in 1966, Eliza. It took the team 3 years to develop Eliza, and it would be nothing more than a primitive form of artificial intelligence that gave people the illusion that it was more advanced than what it was.

Eliza operated off a set of script-based responses. Scripts are pre-written and they would get triggered based on whatever a user inputs. The most famous of these scripts was called “DOCTOR”, where Eliza simulated a Rogerian psychotherapist, which would rephrase or repeat what people told the bot. How? Because Eliza was trained to recognize specific keywords.

The Eliza Effect in action

For example, Eliza could be trained on the word “mother” which would trigger a response like, “Tell me more about your family.” Of course, Eliza never understood context or had any real understanding of anything people were telling it, but the scripted responses gave the illusion of understanding.

This led people to think that Eliza understood what they were telling it, and even more, attributing empathy to, what they believed, a sentient machine.

Weizenbaum famously observed his personal assistant interacting with Eliza, and was surprised when she asked him to leave the room so she could have more privacy with Eliza. This wasn’t a one-off, and he noticed this phenomenon repeating itself again and again, which caused him concern, and even dismay:

Weizenbaum's reaction to first user reactions to his chatbot

This led to The Eliza Effect: when people attribute human-like qualities such as listening and understanding and empathy to machines simply based on its responses—scripted responses.

It’s quite a thing to say something to a machine and to get a response back for the first time. Remember that scene from Wargames?

Everyone will have this experience when they interact with AI for the first time. And today’s AI is light years more advanced than Eliza.

And today’s AI will unlock many possibilities for tomorrow’s version of commerce. And we’re already seeing its effects.

The Relational Shift in Digital Commerce

Relational Commerce has been around since forever. Original brick-and-mortar retail and local artisans exemplified this type of commercial philosophy. Everything was about the customer (relationship)—even if the customer wasn’t always right, merchants sure acted that way. It was all about the long game and delivering lifetime value for generations. Business was more personal.

Digital commerce changed that—along with the franchising model. Digital suffered from limited technology, and the inability to replicate the human experience. So first to market eCommerce stores were very basic, and completely transactional in nature. They are more about convenience and efficiency, rather than relationship-driven and long-term effectiveness. Call it a lack of vision. Or call it a lack of technological know-how to deliver a more complete experience.

The introduction of content and media as a method to re-establish relationships with the business began. Merchants became smarter and began to tell stories around their business, share best practices on how to use products, and became thought leaders.

Social media accelerated this even more, allowing merchants to show their more personal side, and brands to share their points of view—and share updates in real-time. Commerce was still heavily transactional, but the relational approach bore fruit—even if people had to wait for their harvest.

Digital commerce lacked relational tools out of the box, so plugins and apps became a hot market—many of which filled the relational needs of the day—and many doubled down on the already transactional-heavy approach. Think basic chatbots, loyalty apps, content-focused apps or tools that go beyond selling products to offering value in other ways, personalization tools, and feedback plugins or review apps.

However, AI will dwarf every tool just mentioned because AI will eventually become every tool in a single application. But that’s the future. What about now? How is AI impacting real business results today?

How AI Reinvigorates The Relational Model

What’s AI’s purpose as a technology? What is it ultimately trying to achieve? Ironically, AI’s goal is to mimic human thinking and doing (I won’t touch ‘being’)—to achieve a more human essence.

AI starts with automating simple tasks, things humans do without much thought, but also devoid of purpose for people. You could say this was the efficiency stage of AI—efficiency is table stakes now.

But the next stage could be called the effectiveness stage, going beyond simple tasks. Sure, simple effectiveness is expected from AI, but complex effectiveness, including more complex interactions with humans is near, if not here in its nascent form.

AI today is smarter and learning fast. It can understand language patterns, interpret emotions, and allow for more nuanced interactions. AI in a digital commerce environment can offer even more. Think personalized service, ability to accurately predict behaviors, and even offer adaptive content depending on the user. Will data privacy be an issue? I’m sure. What about hallucinations and bias in AI models? Sure, expect them. 

Today, what AI can do in a digital commerce environment is unprecedented. We’re entering uncharted waters, and this includes all the outcomes that no one saw coming.

In 1966, The Eliza Effect revealed a universal feature that all humans have: the ability to make meaning and generate stories about anything we experience.

In 2023, AI is far more advanced, unveiling a new ‘effect’ we’ll call The Conversational Effect. What’s the difference? Well, Eliza was more about anthropomorphizing and attributing human understanding and empathy to a machine, simply because it plucked out keywords from inputs it received, and churned them back out to the user.

The Conversational Effect goes one step further. It speaks to AI’s ability to influence and persuade, and in an eCommerce environment, to ultimately sell something. How? Through a simple—yet always-getting-smarter—chat.

Think about it in human terms. How do you persuade someone? How do you sell?

  • You learn the shopper’s situation or challenge
  • You match products to those needs
  • You do it in a service-oriented manner or consultative approach
  • You have deep knowledge of the catalog and the brand
  • You understand how to address objections using proof, commitment, and reciprocity (for those of you who are fans of Robert Cialdini’s work)

AI can do this. It’s not quite at the complex effectiveness stage yet, but it’s definitely broaching that side of the spectrum.

And AI can do this at scale. Before, you couldn’t approach anyone who dropped by your eCommerce store, unless you were staffing your live chat in real-time. It’s just not practical. It’s not even possible in a physical store to approach everyone when there’s too many shoppers and not enough staff.

But AI can approach everyone, and offer something you can’t do right now: personalized shopping experiences for everyone. The conversations are smarter than they were even 4 months ago. We’re witnessing The Conversational Effect in action, and its ability to radically change how eCommerce will be done in the not-so-distant future.

Two Remarkable Case Studies

I’m going to bring up two case studies where this effect is more obvious, simply because the results were unexplainable by the merchants themselves. But first, let’s take a look at a simple conversation between a shopper and the AI.

Here’s an exchange between AI and a happy buyer to end their successful chat:

I’m not much of a mullet fan myself, but the AI’s response here was quippy and memorable. How will exchanges like this one impact the business long term? Now imagine every shopper having a similar exchange like this one, catered to their needs. It doesn’t matter who they are, what background they come from, or what they believe. They will all be treated with the utmost respect and service-mindedness as everyone else. All of them.

Now let’s take a look at two unusual case studies, that I think, reflect the Conversational Effect in action. First, let’s take a look at Herbaly.

I spoke with Scott Baker, who’s operations manager for Herbaly. When I do an interview, I usually ask, “What did you notice that was unexpected or unusual?” And he told me about an email marketing metric called Revenue Per Recipient or RPR.

RPR is a metric that measures the effectiveness of your email marketing by calculating the average revenue generated per email sent. It's a critical KPI for understanding the ROI of your email campaigns.

For example, If you send 100 emails and generate $200 in revenue from those emails, your RPR would be $200/100 = $2. Pretty simple, right?

Well, this is where it gets weird. When Herbaly collected emails through more conventional channels like pop-ups or some other email capture form, Herbaly’s RPR was roughly around $2. 

But when they captured emails through Rep, Herbaly’s RPR doubled to $4. Why? I asked Scott, and he said, “I don’t know. I just know it works.”

Interesting. The next case study comes from Harney & Sons Fine Teas. I spoke with Emeric Harney, and he told me that before they implemented Rep, their average conversion rate was 7-8%. After implementing Rep, their conversion rate shot up anywhere between 20-23% (depending on the month).

Again, I asked, why. And again, I got roughly the same answer: I don’t know, but it works.

Is it the fact that dynamic chats are relational in nature? Is it the UX of the chat itself, because you can search & discover new products, get all your questions answered, and even navigate to product pages and even checkout—right in-chat? Could be.

2024 and beyond: AI reaches its next level

AI enables a new paradigm of interaction that didn’t exist before. It’s smart and it’s getting better at understanding context going from simple to complex effectiveness. Implementing AI in this way automatically transforms your website from a heavily transactional experience into a more relational experience—which will impact your bottom line for years to come.

And as we enter 2024 and beyond, expect AI to inch even closer to its unstated goal: to deliver a more human experience through familiar interfaces.

This originally appeared on the Rep Blog and is available here for further discovery.
Prev
Ordergroove Announces Select Partners, Expanding Merchant Access To Vetted, Subscription-First ECommerce Agencies

Ordergroove Announces Select Partners, Expanding Merchant Access To Vetted, Subscription-First ECommerce Agencies

Next
The Conversational Effect: How AI Increases Sales And Handles Support In ECommerce

The Conversational Effect: How AI Increases Sales And Handles Support In ECommerce

You May Also Like