Amazon is the 800-pound gorilla is every brand’s c-suite.
Whether it’s their usability, recommendations, or membership perks, Amazon wins because they do customer service right – and at scale.
What many businesses don’t know, though, is that the secret to Amazon customer service success doesn’t have to mean an Amazon business size.
Even Amazon itself uses AI to maintain its’ competitive advantage in customer service and experience.
- Amazon’s uses machine learning to drive product recommendations. They use a combination of Collaborative Filtering and Next-in-Sequence models to make predictions on goods an individual consumer may need next. Amazon possesses a massive database of consumer purchase behavior to power its predictions.
- Amazon uses AI for the logistics side of the business. Artificial intelligence reroutes, changes delivery arrival times, and makes other adjustments for accuracy and efficiency. Soon, Amazon’s interest in drone delivery will start delivering to your doorstep.
- Natural Language Processing (NLP), an emerging deep learning technique, powers Amazon’s digital assistant Alexa.
Amazon made AI its competitive advantage. And don’t expect them to stop pushing the limits on the power of data and AI.
And you can expect even more innovative from Amazon, especially in the AI arena.
Because Amazon’s patent on one-click payments is set to expire this year.
Losing the one feature that led the giant to domination won’t deter them. It will only feed the fire to find new ways to be disruptive.
Amazon is filing numerous machine learning and AI-focused patents.
Soon, personalized online experiences powered by artificial intelligence will be the expectation.
The Future of AI, Retail and ROI
After all, AI enables an ecommerce website to recommend products uniquely suited to shoppers and enables people to search for products using conversational language or images, as though they were interacting with a person.
This has been one of the key missing ingredients for a larger ecommerce revenue share within the retail industry: lack of the personalization brick-and-mortars can offer.
In that same vein, other opportunities emerging include using AI to personalize the customer journey.
This alone could be a huge value-add to online retailers.
Retailers that have implemented personalization strategies see sales gains of 6-10%, a rate two to three times faster than other retailers, according to a report by Boston Consulting Group (BCG).
It could also boost profitability rates by 59% in the wholesale and retail industries by 2035, according to Accenture.
With AI, there are many opportunities, but what does it all mean? And where do you start?
That’s what we’ll cover here, including:
Embracing AI for ecommerce is no longer a feat of Amazon’s unparalleled resources.
In 2016, artificial intelligence was democratized as cloud-based microservices and made available for fractions of a penny per transaction.
The same disruptive forces powering Amazon, Google, Facebook, and Netflix is now democratized.
Right now, many businesses are getting ahead. Most, however, don’t know where or how to start.
The AI Expertise Gap and How to Solve It
Image credit KUNGFU.AI
Many ecommerce leaders are pondering how data and artificial intelligence can be their differentiator.
In retail, some patterns have emerged on where the industry is focused.
However, few companies are actually executing their AI initiatives. Most get stuck on how to start.
After all, there are common barriers to beginning with ecommerce AI.
1. Lack of vision.
A top reason companies have not piloted projects for AI is they lack a clear strategy.
It is important to have a clear vision of how AI should be used to drive ecommerce and how to execute.
Companies need to assess their data and understand where their data provides differentiation.
Once the value within your data is understood, AI can be used to better understand relationships between data sets, predict what will happen next, and/or automate processes.
It is important to have a roadmap for both near-term and far-reaching implementations.
2. Expertise gaps.
Many companies simply do not have the requisite skill set to get started.
AI projects require data, machine learning, and technology experts working together.
Data scientist expertise, for example, is often under-appreciated or over-generalized.
Companies struggle maximizing on data projects by not having the correct expertise on the right project.
Expertise is in high demand.
In the United States, for instance, there were approximately 150 million workers in 2016, but only 235,000 data scientists. OR .15% of the working population for one of the most in-demand fields (MIT Sloan Management Review, 2017)
3. Bad data.
Artificial Intelligence output is only as great as it’s input.
If you don’t have the right quality or quantity of data, the value of AI is limited.
Most companies are committed to collecting and storing data but lack resources and understanding to identify the good from the bad.
The most valuable data for AI remains hidden in unstructured or flat files.
In fact, 80% of all data is unstructured while less than 1% is being analyzed today (IBM).
4. Competing technology priorities.
Companies are often unsure how to stack rank AI spend against other technology and information spending.
Often the CIO and CTO need to align and pool budgets to execute successful projects where data is captured, stored, accessed, and processed.
5. Unclear use-cases.
Next to “lacks a clear vision” the top barrier to AI adoption is the uncertainty around finding relevant use cases.
It is important to know what is easy and what’s hard in artificial intelligence today for online retailers like you.
It is critical to move from a passive state of AI exploration to an active state of piloting projects.
To close the thinking-doing gap:
- Take a pragmatic approach.
- Identify narrow use cases well-supported with data.
- Choose open-source AI algorithms or companies offering SaaS products to quickly see success on small projects.
- Learn what an AI win looks like and the process to go about creating them.
Let’s dive in and get started.
The Economic Benefits to Ecommerce AI
Those who have evolved from thinking to doing are seeing the benefits.
Top benefits of AI in ecommerce include:
- Enhancing products.
- Making better decisions.
- Informing the creation of new products.
- Optimizing processes.
- Identifying new markets.
- Automating workflows.
These improvements touch both on automation and saving time as well as making more money faster thanks to better decisions and a clearer path to success.
Image credit KUNGFU.AI
Length of Time to AI Implementation ROI
But how long does it take to actually realize the ROI on an AI implementation?
While it’s not uncommon to see immediate success after initial pilot projects, more companies report seeing a substantial benefit after launching several projects.
The more you try the more institutional knowledge you create.
Many companies are taking a Lean Startup approach and rapidly deploying projects to test and learn.
Image credit KUNGFU.AI
Retailers who are taking action are making their intentions known.
Over the last few years, many of the enduring brands have announced their AI projects on their earnings calls.
Even in the past few years, companies that have failed to adapt to customer and technology trends are shutting down operations.
AI is both exciting and scary.
The opportunities to better serve customers, understand demands, and offer better products or experiences with less guesswork should invigorate all businesses.
However, it is scary to think about all the accelerated benefits derived from the intelligence and capability awarded to first-movers compared to the wait-and-see crowd.
The key to getting started is know which use cases to deploy, have a solid strategy, and pilot small projects for incremental ROI.
Let’s look at the top 5 ecommerce AI use cases.
5 Use Cases for Ecommerce AI Implementation
Today, AI is enabling companies to sense, predict, and automate.
In retail, for example, we see many use cases already making an impact.
Image credit McKinsey & Company
The opportunities for capitalizing on artificial intelligence in ecommerce are vast.
Many of the companies that are successfully implementing AI are using it to enhance their ecommerce engine and capability to generate revenue through sales and marketing.
Image credit KUNGFU.AI
Companies are using data and AI to enhance ecommerce features, functions, and performance.
Here are some of the most common ways.
1. Predictive Product Recommendations.
We can now collect and process petabytes of data and append mass-consumer purchase behavior to an individual’s purchase history to offer relevant and helpful product recommendations.
Online retailers are finding clever ways to collect customer data in the forms of:
- Natural language (through voice, for example)
- Product reviews
- Call center transcripts
- Purchase history
- Social media
- Web analytics.
Predictive product recommendation engines can correlate all these data sets to make hyper-personalized selections.
Amazon is a common example of how to track buying behavior by account to recommend related products.
On your own website, you can do the same manually without any additional work.
- View BigCommerce Insights “Products Frequently Purchased Together”
- Add those products to the proper product page.
With AI, you can further automate this process by using apps like Beeketing or a variety of other services.
With advances in artificial intelligence and machine learning, new deep personalization techniques have entered ecommerce.
Personalization is the ability to use mass-consumer and individual data to customize content and web interfaces to the user.
Personalization stands out from traditional marketing allowing one to one conversations with consumers.
Good personalization can increase engagement, conversions, and decrease time to transaction. For example, online retailers can track web behavior across multiple touch points (mobile, web, and email).
The data can be used to provide a seamless customer experience across all channels.
So the next time a user is on your site shopping for a new laptop, you can send a push notification on their phone with a discount code for the new line.
Tools and apps like Choice AI, Smile.io, and a variety of others.
3. Dynamic Pricing.
Dynamic pricing is a strategy based on which retailers change the price of the product based on supply and demand.
AI enabled dynamic pricing is a new disruptive force to hit the ecommerce world.
While having fluctuating prices are not new (happy hours, stock market, airline tickets), the data we can now access unlocks new potential.
We can now append customer data, competitive pricing data, and sales transaction data to predict when to discount, what to discount, and dynamically calculate the minimum amount of discount needed to ensure a transaction.
Dynamic pricing algorithms are an emerging field in artificial intelligence and ecommerce leaders may be hesitant to trust the models that inform pricing.
Early adopters are already leaving competitors behind.
Amazon is the current leader in applying dynamic pricing to their products. They are seeing massive success.
While other online retailers are experimenting with dynamic pricing, Amazon has it mastered.
It’s a key reason they are ousting competition. They’ve managed to price their commodities lower than others.
Tools like SellerActive, Price2Spy, and FeedVisor automate dynamics repricing for Amazon.
4. Artificial Agents.
Artificial agents and chatbots are a computer program designed to simulate conversations with human users, especially over the Internet.
Artificial agents are being used to interface with customers on ecommerce sites, inform customer service agents how to service inquiries, and even facilitate sales.
It is important to note that bots are not totally self-reliant. They are great tools to help facilitate simple transactions (like answer basic questions, set appointments, or triage) and provide basic decisions.
A human in the loop to act as a backup to the bot is still required to prevent user frustration.
The ecommerce giant eBay is a pioneer in the use of bots for commerce.
A screenshot of eBay’s ShopBot
The eBay shopbot functions as an AI assistant to help users easily find products of interest using natural language.
Users can communicate with the bot via text, voice or using pictures taken with their smartphone of images related to a particular product.
There is a lack of published evidence to substantiate whether the chatbot has proven to be a significant driver of revenue for the company.
However, there is evidence to suggest that machine learning is an integral component of eBay’s business strategy:
“We apply machine learning techniques to item-to-product matching, price prediction and item categorization tasks on eBay.
We also employ them for attribute extraction, generating the proper names of browse nodes, filtering product reviews and more.
Machine learning helps us optimize the relevance of shoppers’ search and navigation experiences.” – Selcuk Kopru, eBay Research Scientist (August 2016)
Brands looking to implement a similar solution on their own site can use tools like:
5. Predictive Behavior Modeling.
Ecommerce technology alone does not drive sales.
You need a successful sales and marketing strategy to support the engine.
And successful sales and marketing are predicated on a strong understanding of the customer.
Today we use our own experiences working with customers, past purchase behavior, market analysis, and personas to better understand how our customers may behave in the future.
Access to more data, sophisticated neural nets, and processing power is enabling ecommerce leaders to understand their customers and new trends in behavior better than ever.
Our ability to anticipate our customer next move is made possible by the predictive capabilities of AI.
Now, we can access a multitude of structured and unstructured data sources like social media, loyalty cards, sales, and market research to create deep psychographic profiles of our known customers to spot emerging trends, and predict unknown customers demographics.
IBM Watson Personality Insights.
Here are a few tools to get you started now:
Data Collection & Planning from AI Initiatives
Ecommerce inherently produces a lot of data and most companies understand that data is a natural resource.
Just like oil, data needs to be mined and refined to have value.
Fewer companies, however, know how to derive true value from their data.
Meaning that most of your natural resources are locked away, and not providing any value at all.
The limited data we can access through any sort of computer analysis can help us see backwards, but what about seeing forward? Can our data tell us how to act moving forward?
More companies are data rich than are producing value from their data.
It is good to collect data, sure, but it is better to have a strategy for extracting value.
Leading companies are making data-informed decisions with better intelligence than the competition.
To go from data to action, you need to know what is differentiated about your data, know how to enrich your data with other data sources and have the infrastructure to support it.
Image credit Civics Analytics
You also need the right expertise on staff to derive the value hidden in your data.
And this can be the most challenging part. After all, data science is a misunderstood field.
As mentioned previously, many business and technical leaders over-generalize the data scientist skill set and often don’t have the right expertise on the correct projects. (For a helpful guide on the definitions of roles in the data science field, see here.)
Ensuring your staff has the right data expertise is critical to making data actionable.
Having the right strategy, team, and technology to support AI allows ecommerce companies to evolve from simply collecting data to predicting behavior and taking proactive measures.
Where should you begin?
Initially, partner with firms who provide part-time, or project-based consulting and execution for projects.
Then, staff internal team-members who can work alongside hired help to learn and document the process.
AI initiatives should start small but have a snowball effect.
Data, ideas, and products will be tightly integrated and provide a perpetual loop of innovation and advanced capabilities.
Image credit AI for Business Leaders
Over time, have a plan to develop your AI center of excellence, or competency. The partnership to pervasive model reduces risks and up-front expenditures but also provides sustainable competitive advantage moving forward.
How to Build Your Brand’s Ecommerce AI Muscle
To move from data to action, you need a starting point.
Here are the steps you should follow to build out an AI center of excellence and begin to strengthen your brand’s AI muscle.
- Suppress the urge to post an opening for an AI expert to take this all on. And don’t just punt this over to your CIO or CTO.
- Start with a strategy. Winning strategies take a practical approach and start small.
Find narrow use cases that are relevant to the overall corporate strategy. The most successful AI use cases live at the intersection of business objectives, data differentiation, and readily available artificial intelligence models.
- Leverage third-party expertise on a project or part-time basis which can objectively help you build your strategy. Starting with a hire-first strategy may take 6 to 9 months to complete your AI vision. A tiger team of experts can help you build your AI roadmap in 6 weeks.
- Once you build the roadmap, ID narrow use cases that solve the biggest problems with the quickest implementations scenarios. Build or implement an MVP version of the solution. An MVP (minimum viable product) should provide the opportunity to deploy your solution quickly and also provide an interface for training employees on the AI model to confidently execute its task.
- Once confidence is achieved (both algorithmically and institutionally), build toward the full-scale solution.
Image credit KUNGFU.AI
Remember, it may take a few projects under your belt to see substantial benefit.
It’s not unusual to see moderate benefits after your first project.
The goal is to work toward full AI competency.
You can de-risk your companies adoption of AI, by working with third-party partners who can work in sprints to assess your data, build your AI strategy, and implement initial solutions.
Over time, you should up-skill your organization or hire these capabilities to see maximum benefit.
How to Develop an AI Center of Excellence
Artificial Intelligence is not a technology or just another as-a-service solution.
AI is an ecosystem of people, technologies, data, and process.
It is the logical next step in digital transformation.
Therefore AI should be approached as a corporate decision and not a technology decision.
Image credit KUNGFU.AI
AI requires some companies to break down their data, people, and technology silos and integrate them into a new center for AI.
The AI Center of Excellence is when:
- Business strategy drives solutions built from data.
- Technology and infrastructure supports data extraction, enrichment, and accessibility as break-neck speed.
- A team of experts are running AI proof-of-concept development and execute the learning process for all AI models being deployed.
- All of these components are carefully organized and managed through a carefully constructed process with DevOps.
For this reason, we are seeing the rise of the AI product manager.
The AI Product Manager takes the lead for any new initiatives requiring artificial intelligence.
They can work within individual business units to identify the needs and generate ideas on how artificial intelligence can be implemented.
The AI Product Manager then finds the data and models required to execute.
They may even build a business case for acquiring new data sources to enrich the algorithm’s output.
Over time, the AI Center of Excellence helps ecommerce driven organizations move beyond seeing project-level benefit to being truly disruptive with augmented intelligence and capabilities.
Since 2016, ecommerce companies at large have begun to focus on AI as an augmented capability for recurring, manual tasks such as dynamic pricing, virtual agents, personalized experiences, and more efficient delivery.
For larger companies, like Amazon, AI is an incredibly differentiating force.
For smaller, up-and-coming brands, AI is a change management challenge.
But the gap between the AI haves and the AI have nots is widening.
The prospect of another seismic technology shift may be daunting, especially on the heels of the digital transformation.
Don’t be a victim of analysis or change paralysis.
Use the foundation laid from prior transformation initiatives and data collection as the wind beneath your sails.
Here is a quick checklist on how to tackle the AI in ecommerce for your brand.
- Start small. Don’t worry about auditing the organization and hiring your first C-level AI leader out of the gate.
- Find a good partner who can take you from data collection through to a well-defined strategy, and then into action.
- Learn from that partner to build your own competency over time.
You’ve already accomplished the hard part. You have data and a corporate strategy.
All you need is a relevant, narrow use cases to apply to your first AI project.
AI implementation can be hard to get off the bench, but doing so is critical to survival in the future ecommerce economy where winners and losers will be defined by what they do with data, and how they scale their human capital.