Summarise with AI
Summarise with AI
In today’s fast-paced digital world, artificial intelligence (AI) has…
Almost fooled you.
Unfortunately for you, this whole blog was written by a human, and possibly the worst human you could’ve picked to entertain and enlighten you: me.
You see, forcing an AI to write a terrible blog for a terrible website cobbled together by a self-taught dev that more closely resembles the cars from the Blues Brothers rather than the technological marvel that my mum attests that it is (hi mum!), is a complete waste of potential for the Machine-Overlord-In-Training.
No, the real applications of AI you should be looking out for are the applications in the big leagues: enterprise SaaS.
With treasure-troves of data and a T-1000 style determination to maximise shareholder value, enterprise is where the ground is breaking and cash is raking (in, that got away from me, sorry Mum).
It’s a simple fact that the AI applications that are going to cause the biggest waves and change how human life is conducted in the 21st century. So, with that in mind, let’s take a look at some real ways blue chip enterprise players are embedding AI into their workflows.
A core use case that is already rolling out across enterprises is the use of AI in customer service.
Given their ability to parse large amounts of data and then spit back pertinent and succinct answers tailored to user input, LLMs are perfectly positioned to work as customer service agents, and in fact, are probably more effective than a traditional human CS agent when it comes to utilising the swathes of data that enterprises are sitting on.
Retail giants like Sephora and H&M are already investing heavily in customer service agents, for a simple reason: they have a massive amount of stock that is basically impossible for a human service agent to keep track of with legacy CS set-ups.
Now, customer service teams are able to feed knowledge bases on their vast catalogues to AI chatbots, which means they can provide up-to-date and relevant information to users, whilst Stock and instructions can be adapted and changed on a dime.
For those who repeatedly hammer in “talk to a human” whenever they are confronted with an AI customer service agent, I want to reassure you this doesn’t mean human customer service agents will become obsolete, far from it!
I’ve posted before about how I believe human-to-human connection will be a core differentiator for businesses in a lightning-fast AI-augmented world.
AI agents will essentially be a qualifier, there to alleviate manual and frequently impossible tasks for customer support reps, whilst support reps will take a more brand-building role, with a greater emphasis on activation and making customers feel heard and valued when they interact with a brand.
I could be cringey here, and say, data is the new oil, but really: data is the new oil that also cures athletes' feet and grants wishes to boot.
But collecting data alone isn’t enough.
You need a way to parse that data and produce reports and insights that allow you to move forward with confidence, especially at scale.
AI is perfect for this as it won’t ever get tired of staring at 1s and 0s until the eventual heat death of the universe, and ecommerce giant Shopify have been super forward-thinking in leveraging AI in organise their data structure and analyse that data to find new opportunities.
They’ve built a bespoke model that can analyse 70 million data points, and predict a customer’s minimum sales with a solid 90% accuracy, which they then use this data to offer loans and funding to merchants they believe will be the most commercially successful via their Shopify Capital program.
Another in-house model provides up to date tax liability information, automating possibly the most boring part of running a business seamlessly, and provides the feel of an in-house tax lawyer for their clients.
This trend of AI automation reducing boring man hours spent on analysis is one that smart enterprises are applying rapidly to beat the stagnancy and slow-movement often associated with large businesses.
Case in point, JP Morgan recently managed to reduce an insane 360,000 hours of annual work with their COiN system that can review and analyse financial documents.
Data creation has long since passed the point where mere mortals can accurately measure it alone, and big businesses have realised the necessity of using machines to properly extract value from the endless stream of information created each day.
“Leaders win through logistics. Vision, sure. Strategy, yes. But when you go to war, you need both toilet paper and bullets at the right place at the right time.”
Logistics and operations matter greatly.
Remember that time one ship getting stuck in one canal managed to grind global commerce to a screeching halt? That incident alone is A) hilarious for some reason and B) a hard reminder that “getting stuff from A to B” is wildly important.
Despite the massive emphasis on them as areas to ruthlessly optimise and tweak, Logistics and operations are also incredibly hard to execute due to how many vectors impact them.
Accidents, vehicle malfunctions, wear-and-tear, weather changes, political turmoil, giant ships stuck in comically small channels that for some reason every boat in the world has to go through, these are all things that can and have impacted global operations of businesses everywhere.
That’s why huge investments into AI are being made at the upper echelon of the business world to maximise and streamline operations to ensure timely deliveries and razor thin margins are maintained and expanded.
For example: transport giant UPS has decided to use AI to build a digital twin of its entire distribution network.
A digital twin is a digitised replica of a real-world object, and companies are quickly utilising them to model scenarios and accurately analyse and predict where efficiencies can be made, without the sometimes costly process of giving it the ol’ college try beforehand. UPS specifically uses their digital twin to model supply chain breakages, build contingency plans, identify the friction points in the network more efficiently and provide more detailed information to their customers on the progress of their deliveries.
Business operations are usually manageable at a small scale by dedicated teams, but for giant, world-impacting operations tasks, get ready to see more and more AI applications designed to organise and implement better logistics.
“By the way, if anyone here is in advertising or marketing, kill yourself” - Bill Hicks
Soon, Bill Hick’s vision of a world devoid of marketers and advertisers will become reality.
Probably not in the way he envisioned it however.
Marketing is one of the areas where AI has achieved deep market penetration due to the constant demand for more and more sweet sweet content.
Interestingly, AI is operating on a sort of self-perpetuating flywheel in the marketing space: AI makes products easier to build, putting greater emphasis on distribution, meaning more content is needed, meaning greater reliance on AI tools for content creation.
But this need for rapidly produced content isn’t just a phenomena that affects small startups looking for market share. It also affects giant businesses looking to create content on massive subject areas at a substantial magnitude whilst remaining cost effective.
Sage Publishing managed to reduce marketing spend (often the 3rd or 4th biggest expenditure for businesses) by a whopping 50% using Jasper AI to generate marketing copy for 1000s of books, enabling a greater display of their catalogue to tens of thousands of prospective buyers.
Predictive analytics (a subsect of AI) is also able to create truly special and personalised marketing experiences for users.
Take for example, the astronomically popular and successful Spotify Wrapped and in-built playlist recommendation.
By leveraging the massive amounts of data generated by music streaming, Spotify builds personalised experiences both in a massive, macro way (Wrapped) and also interwoven into the daily use of the product (Discover Weekly) . This enhances the stickiness of the product, and enables a user to constantly be finding new ways to experience music.
As we move towards fully realising the potential of the data backlog we have been building for the past 3 decades, it’s important to take a moment to reflect on the ethics of utilising this data for either marketing, operations or customer service.
I’ve written a pretty extensive framework for navigating risk when creating new tech products, but the basic premise involves consent for data mining, manageability of the risk, and whether the negative impacts of any side effects can be reversed relatively easily.
But we also can’t fear large-scale AI initiatives.
Implemented correctly, the possibilities for machine learning at a macro scale are tremendous, analogous to the widespread propagation of electricity, and could save millions of man hours, recognise opportunities that would have been lost to the mists of time or shift the paradigm of how big business functions.
Load comments
Comments
No comments yet, be the first!