5 Jobs That AI Can Never Replace in Your Ecommerce Business

It’s every business owner’s challenge: how to cut costs and make more money. Automation is often the answer — and many times, it’s a good one. Automation can save time and money, as well as increase accuracy, efficiency and profitability.

But there are times when business needs a human touch.

Fancy robots and artificial intelligence (AI) can’t do everything. There are times when you need a real, live person.

So, how do you know which are jobs that AI can never replace? Stanford professor Andrew Ng shares his POV: “If a typical person can do a mental task with less than one second of thought, we can probably automate it using AI either now or in the near future.”

Below, we take a look at five jobs that can’t be automated for your ecommerce business.

Product photography

Product photography is absolutely critical for ecommerce brands. According to one Weebly survey, product photos are important to more than three-quarters of online shoppers. And 22% of ecommerce returns happen because of a discrepancy in appearance between the photos and real life. One-third are also influenced by photo resolution, indicating a desire for high-quality shots.

According to our own analysis, the top online fashion brands have an average of eight photos per product page, and this includes a variety of plain background and contextual shots. And Weebly’s survey respondents want to see multiple angles.

This is a lot to put together, and it can be tempting to automate along the way. But when it comes to the presentation of your products, you want a careful attention to detail. This includes everything from setting up the shot to editing the photos prior to publishing.

If you’re outsourcing the photo edits, it’s a good idea to check with your service provider to ensure they don’t use any automations or AI to fine tune your shots. This can lead to Photoshop fails, unhappy customers and bad press in worst case scenarios.

Above is an example of one AI-based photo editor we tried that had trouble clipping a basic object. Imagine what would happen to your photos that need complex clipping paths for things like hair or fur.

Customer support

Automating customer support has many benefits you’d expect (saving time and money, for example), but there are instances when service needs a human touch. For simple inquiries and commonly asked questions, you can likely use an automated chatbot or even a self-serve help center. But when the questions become more nuanced and complicated, it’s time to switch to a real person.

“The whole reason behind automating parts of your customer support is to empower the support staff to handle more customer tickets when it matters most,” says Dylan Max, head of growth marketing at Netomi. “Human-powered customer support is an essential ingredient for any successful support staff because nothing can replace the human element of compassion and understanding.”

Here are a few indications from Max that can help you know when to put a person on the job:

• Your customers’ issues are unique and rarely repetitive.
• Your agents have a lot of leeway to subjectively and creatively problem-solve issues, with little to no consistency in response.
• You manage highly sensitive issues, like offering guidance to customers in critical situations (i.e., should your product cause physical harm or a substantial accidental financial charge.)

When it comes to automation in customer support, it’s also important to strike a balance.

“It can be very frustrating for customers to always connect to a bot,” says Nikola Baldikov, digital marketing manager at Brosix, a secure instant messaging software for business communication. “The balance between a human being and a bot in customer support is when you can use the quick reaction of a bot in simple repetitive situations, and the creativity of a person in more complex cases.”

Intercom recommends looking for the following when setting up your customer support processes:

• Differentiate common and easy questions from rare and hard questions.
• Quickly resolve common and easy questions with automation or self-serve content, or by allowing support staff to quickly point customers to that same content.
• When customers are helping themselves, provide a way for them to tell you if their question was answered or if your content needs improving.
• Provide a way for customers — who think they have common and easy questions, but are actually hard and rare — to easily transition from self-serve content to the live support team.

Image: Intercom

Copywriting

According to Salsify research, 87% of consumers consider product content — your product descriptions, headlines, specs, photos, reviews, etc. — to be extremely or very important when making a purchase decision.

And, the buyer journey is becoming less and less linear. This means buyers consume content and information about your products on multiple channels, in different contexts and at different points along the path to purchase.

In other words, your copy and messaging is an important component every step of the way.

It’s also not just about crafting beautiful copy that simply describes a product. It takes time to research the voice of your customer, run a competitive analysis, and draft strategic yet engaging headlines and product descriptions. That being said, copywriting is both a right and left brain activity. It’s part research-driven and logical, part creative expression.

Tom Pickersgill, founder and CEO of Broadstone — a staffing platform that leverages artificial intelligence and machine learning to match job seekers with employers — put it this way:

“Humans use their life experiences, their emotions and their creativity to bring things to life. Robotics and AI use data to learn and improve. I don’t believe data can produce genuine works of art that will truly engage an audience through shared experiences, whether that be a painting, a melody or a voice.”

There’s an interesting Medium blog which is made up of completely AI-generated content and copy. While it hasn’t been updated since 2019, it still showcases how robots can easily mix up the English language.

Take this excerpt, for example:

“Influencers can give a lot of credibility to their projects. They have relationships with the people who work in that industry and offer them access to more. Influencers can afford to work with celebrities who they know very well. And they can gain significant patronage in return. They can give small things extra publicity by working with people they know.”

Imagine the buyer experience with that copy. While technically it makes sense, it doesn’t say much — and it’s not fun or easy to read.

Ange Friesen is a brand strategist and copywriter who helps ecommerce and digital technology brands figure out who they are and how to share their ideas with the world. “It’s impossible for machines to achieve the nuance in tone that a human can create when they write,” Friesen says.

She continues: “Before you decide to automate, find yourself a great copywriter to work with. Someone who’ll take things beyond that step of writing words that ‘sound nice’ and actually works with you to create copy that’s designed to convert.”

Wholesale and retail buying

There are many things technology can't replace: strategy, critical thinking, creativity and imagination — just to name a few. These are critical traits for wholesale and retail buyers, making these retail positions less susceptible to automation. And it’s likely one reason data scientist Patrick Scott outlined in The Telegraph that wholesale and retail buyer positions (with the exception of farm products) only have a 29% likelihood of being computerized.

Now, you may be thinking that technology actually could drive these positions eventually — but to understand why automation may not affect them is to understand what these positions really do behind the scenes.

Let's talk about buyers first. Essentially, buyers are the strategic eye for brands. They plan, select and purchase based on market research, competitive analysis, product testing and evaluation, and more — all to ensure the products they purchase actually sell instead of sit on the shelves. This position especially takes skill, creativity and decision-making — all traits that automation can't replace to make a sound and strategic decision for a retail company.

According to totaljobs.com, buyers need:

• Strong analytical skills
• Creativity
• The ability to recognize future trends
• Excellent commercial awareness
• Communication skills
• Confidence and presentation skills
• The ability to make decisions under pressure
• Negotiation skills

Notice a common thread? These are skills only a human can do.

Next up: wholesale buyers. While this position comes in many forms and titles, it's all about negotiating wholesale prices based on market trends. Simply put, wholesale buyers are responsible for purchasing items at an inexpensive price to in return sell at a higher price in order to make a profit. Wholesale buyers require many, if not all, of the same traits of a retail buyer — especially the ability to effectively (and quickly) negotiate to purchase items at the right price. And, of course, this takes human interaction.

In fact, McKinsey reports that activities that apply expertise to “decision making, planning or creative work" are some of the hardest to automate with the technologies that are currently available. These activities are referred to as "knowledge work."

Returns inspections

Returns are an unfortunate reality in ecommerce. Consumers demand easy and free returns — 90% highly value free returns and 96% would purchase again if returns are easy. While you can handle returns in different ways, one truth remains constant: returns are a critical customer touchpoint and can make a big impact on your retention metrics.

When a customer returns an item, you can do a few things:

• Refund: This results in a lost sale, reverse logistics expenses and potentially a lost customer — so, you can also choose to offer merchant credit in lieu of money.
• Replace: If the merchandise has an issue, you can send an identical or similar replacement item.
• Repair: In the case of a broken product, you might be able to fix it and ship it back to the customer — you can choose to do this for free or charge the customer.
• Reject: If the item is unreturnable for any reason, you can reject the return and send it back to the customer.

See how each of these has a significantly different impact both on the customer experience and your bottom line? This is why automating return inspections can be troublesome and result in greater losses than necessary.

Amazon, for example, puts each return through a manual 20-point inspection process to assess damages, malfunctions and overall condition of the merchandise. They even take photographs to document each instance.