johnmaconline

The fear-mongering says that AI will take your job.

Be afraid. It’s all bad. At best, our future is one where we all sit around with nothing to do because of AI. At worst, AI tries to kill us. 

In reality, AI, or at least its current and near-future state, has quite the opposite effect. Learning how to use it will make you more valuable to your company and customers. 

The current state of AI is what I call “personal automation.” 

It’s still you, but it’s you with a drill and circular saw. If you dig in, your productivity will go up. 

However, productivity isn’t the only value that AI increases. Because it’s also you with an editor, an Excel expert, a personal assistant, a graduate assistant, and an awkward teenager in his parent’s basement who’s learning to write Python code to scrape the contents of questionable websites. 

“Here is a list of highlights from last year. Create a message to send to our partners that summarizes it succinctly with an optimistic tone.”

“Create a formula that looks through my table of data and makes 3 categories: > 100, < 1, and everything else.”

“Look through this chain of emails, summarize it, and tell me if I’ve been asked to do anything, and if so, create a reply that indicates I’ll work on it next week.”

“Find the trends in this set of data.”

“Here is a Python script. Keeping this same style, add a new function that takes in a filename from the –ofile argument and directs all output to this file. The default is out.txt.”

It’s you, but more valuable. 

Each step, analysis, task, decision, operation, and assignment in this story already is or will be, theoretically possible with AI. If you broke down this story into each of its constituent parts — from entity formation and legal work, to business case analysis, marketing plans, and content creation, all the way through manufacturing and delivery — you can plausibly replace that part with AI. 

If you work in any of the areas mentioned, AI will affect you. But will you be replaced? Will you lose your job?

Autonomous automation is the dystopian nightmare of Vonnegut’s Player Piano. And you’re starting to see it in corporate road maps, venture capital pitch decks, TEDx talks, and fear-mongering journalism. 

But that’s all it is right now. It’s in the idea stage, and it’s being driven by the idea people, just like the internet was in the late 90s. It’s in the “we can apply AI to this task and get faster, cheaper, better” block diagram stage.

Some of it will happen. Some of it will start to happen in five to ten years. 

But just like with the internet, and electricity, and horseless carriages, humans and their creativity still matter. These technologies created entirely new business models and industries. 

Automation and AI will do the same. Yes, depending on your job — white-collar work is more at risk than blue-collar work — you may become a victim. 

But like most other victimhoods in this life, it’s a matter of perspective and what you do with it.

We’ve generated a set of technical drawings and specifications for your products, forwarded these to a manufacturer that meets specifications, and created the distribution pipeline. You were hoping for a same-day service from order through shipping, but the cost analysis, considering pricing and manufacturing, doesn’t support it. The manufacturer we’ve chosen to support the business case will manufacture-on-demand and has indicated that each product will require 14 hours to build. Therefore, we cannot guarantee same-day shipping for your customers, but we can guarantee next-day shipping. The manufacturer has locations in Brazil and Mexico, so 2nd-day arrival can be guaranteed for US customers but not European and Asian customers. We know this does not meet your original intentions, but given the business case, this is the best financial scenario for you.

This is where it gets really interesting.

This is supply chain management (SCM). You can get a college degree in it, just like many other fields that AI is coming for. SCM is critical to any physical product development business. Next to sales, the most important thing.

You may have yet to hear of SCM prior to Covid. It used to be a behind-the-scenes pillar of the operations department. Important but not really discussed outside of the walls of the office. But Covid shined the light on the supply chain — or lack thereof. In the post-Covid world, SCM has found its place in the sun. 

And someday, AI will both automate it and make the decisions. 

Connecting the pieces of the supply chain is pure automation, but making the decisions about this manufacturer versus that one and analyzing the delivery outcomes with regard to the business case is very interesting. Businesses pay consulting companies exorbitant fees to figure this stuff out for them.

When will AI take over?

Not very soon. Both automation and decision-making in the supply chain currently sit across a heterogeneous mix of world partners. Governments, language, geography — all deterrents in the supply chain. Even though AI might be smart enough in the near future to do it, the logistics will slow it down significantly.

But someday…

In addition to the website and social media content, we’ve generated a social media and web advertising strategy, and the first ads will start to run today at 6 pm. These ads draw the engagers to the top of the sales funnel that we’ve connected to your entire online presence. We’ve created a strategic communication campaign that resonates with your target audience and reuses much of the content they’ll see on your social media to ensure continuity and build the trust they require. In compliance with the marketing strategy and business objectives, we’ve set a pricing structure that requires two versions of your product with slightly different feature sets for each. This varies from your original idea of a single product but will result in a better overall product strategy and pave the way for follow-on products in the future. We’ve named your products the “BassBlaster Pro” and “BassBlaster+” and have adjusted your product definitions accordingly. 

This paragraph contains an interesting mix of AI automation in combination with decision-making. 

You’re already seeing AI advertising services and automated campaigns. Once the strategy is developed, using AI to automate the advertising and communications campaign itself is fairly straightforward.

But the humans are currently making the decisions and forming the strategies. In our scenario, the AI makes these decisions and forms the strategy. That’s coming in concert with the marketing strategy since they all work together.

The big leap here is that AI is performing the CEO and CCO’s (Chief Commercial Officer) role of devising the business strategy and then a product strategy to support the business. 

Although a leap for now, this is definitely coming as well. Maybe AI isn’t the CEO and CCO but more like a business consulting firm such as McKinsey, Bain, or BCG. In fact, our scenario highlights that AI might have a more meaningful impact on the business consulting world than the product development companies themselves. 

Business majors should be on notice. 

The domain “freakthefishout.com” has been purchased, and the website has been created along with a full set of social media accounts. We’ve generated a baseline of video and visual content with you as hero figure for each of those accounts. We highlight the excitement of the launch of your product and company but contrast it with the content that resonates with your audience, which tends towards introversion, outdoor pursuits, and a requirement of trustworthiness. We gathered enough video of you from our initial conversation to generate the first 30 days of content. We’ll ask for another conversation with you in two weeks so that we can generate the second 30 days’ worth of content. The social media campaign will continue for the next 60 days, as defined in your introductory AutoBizInABox plan, and can be extended indefinitely with follow-on plans. 

The domain, website, and social media account creation for a basic product business like this is, like the legal and financial pieces, straightforward and cookie-cutter. AI could even write the code and make the desirable customizations required for the marketing strategy it has devised. 

You’re now hearing about the wonderful world of AI-generated video content. Sora and deepfakes, etc. And maybe that scares the bejesus out of you.

You can relax. This part of the scenario is quite far off. 

Yes, to some extent, it’s here or very close, but fully automated, realistic content with you as the star, is much farther off. The deepfake stuff that you see today is a human-intense exercise. AI is just the skill that makes it possible. 

Sora is currently generating some interesting video, but its wholly fiction. It’s not accurately generating real people, places, and things. Certainly nothing like capturing someone’s likeness and using that to create. This is theoretically possible but maybe practically impossible for a long time.  

The more interesting use of AI here is to make decisions about the audience. Those decisions tie to an overall product and marketing strategy and guidance that AI formed itself. It’s not exactly an automation use case of AI, but it’s a decision-making use case that’s coming as well. 

“FreakTheFishOut, LLC has been created with you as the founder and sole operator. Your federal EIN number is 12-3456789, and business checking and credit accounts have been created and connected to a bookkeeping and tax service that falls within your business case’s projected financial picture. We’ve pushed a mobile application to your devices for all financial transactions. You’ll find a founder gift of $100 in your business checking account as a thank-you for using AutoBizInABox.” 

AI could rather simply handle this entity law, banking, and accounting, with an interface to the government. 

Low-cost, vanilla entity creation services already exist through sites like LegalZoom and SwyftFilings. A couple of questions and a few clicks later, you have an LLC or S-Corp. To get the EIN, you use the IRS website. It also doesn’t take long, but I suspect there might be a human at the backend of this process because you can only get service during normal business hours. 

Once you have the EIN, you can setup your financial system with checking, credit, and tax services. Services like Found already exist to combine banking with bookkeeping and basic tax filing. 

All of these services already exist, so one part of the automation with AI is just connecting it all together—a true automation task. However, the other, more important piece is when AI doesn’t just create the infrastructure you tell it to but it decides what it should be based on the conversation you had at the beginning. Automation with AI might currently be closest to handling these pieces, but it’s certainly not there yet. 

The legal and financial decision-making and tasks required for this basic type of business is cookie-cutter and broadly applicable across most businesses of this type. If you currently work in any of these areas, AI is coming, and may replace the people that do the specific tasks in the near future. However, the connective tissue required to string all of this together is further off. 

But the basic legal, banking, and accounting professions should be on notice. 

Imagine you have an idea for a business. Let’s say you’ve developed an innovative fishing rig, and now you want to share it with the world (and get paid). 

So you go to “autobizinabox.com,” and a chatbot conducts about an hour-long conversation with you, asking you a million questions. Less than an hour later, you get a notification that says.

Congratulations Mr. Fancy Pants! 

FreakTheFishOut, LLC has been created with you as the founder and sole operator. Your federal EIN number is 12-3456789, and business checking and credit accounts have been created and connected to a bookkeeping and tax service that falls within your business case’s projected financial picture. We’ve pushed a mobile application to your devices for all financial transactions. You’ll find a founder gift of $100 in your business checking account as a thank-you for using AutoBizInABox. 

The domain “freakthefishout.com” has been purchased, and the website has been created along with a full set of social media accounts. We’ve generated a baseline of video and visual content with you as the hero figure for each of those accounts. We highlight the excitement of the launch of your product and company but contrast it with the content that resonates with your audience, which tends towards introversion, outdoor pursuits, and a requirement of trustworthiness. We gathered enough video of you from our initial conversation to generate the first 30 days of content. We’ll ask for another conversation with you in two weeks so that we can generate the second 30 days’ worth of content. The social media campaign will continue for the next 60 days, as defined in your introductory AutoBizInABox plan, and can be extended indefinitely with follow-on plans. 

In addition to the website and social media content, we’ve generated a social media and web advertising strategy, and the first ads will start to run today at 6 pm. These ads draw the engagers to the top of the sales funnel that we’ve connected to your entire online presence. We’ve created a strategic communication campaign that resonates with your target audience and reuses much of the content they’ll see on your social media to ensure continuity and build the trust they require. In compliance with the marketing strategy and business objectives, we’ve set a pricing structure that requires two versions of your product with slightly different feature sets for each. This varies from your original idea of a single product but will result in a better overall product strategy and pave the way for follow-on products in the future. We’ve named your products the “BassBlaster Pro” and “BassBlaster+” and have adjusted your product definitions accordingly. 

We’ve generated a set of technical drawings and specifications for your products, forwarded these to a manufacturer that meets specifications, and created the distribution pipeline. You were hoping for a same-day service from order through shipping, but the cost analysis, considering pricing and manufacturing, doesn’t support it. The manufacturer we’ve chosen to support the business case will manufacture-on-demand and has indicated that each product will require 14 hours to build. Therefore, we cannot guarantee same-day shipping for your customers, but we can guarantee next-day shipping. The manufacturer has locations in Brazil and Mexico, so 2nd-day arrival can be guaranteed for US customers but not European and Asian customers. We know this does not meet your original intentions, but given the business case, this is the best financial scenario for you.

Congratulations again! We love helping human innovators such as yourself realize their dream of bringing human ideas to fruition. 

We suggest you hop onto freakthefishout.com and see all that you’ve brought forth into the world…

AI like ChatGPT is now creeping into the Organizational Augmentation world. 

Organizational augmentation is like outsourcing a skillset or having an excellent assistant. They’re not totally autonomous, but they do a lot by themselves without you needing to get involved. You set the rules and expectations, and they do the work. Plus, you’ve empowered them to make some decisions on your behalf, but staying with the guidelines you’ve set up. 

Organizations and teams start to feel some benefits here.

The Helpdesk is a great example. If you train an LLM on your stuff, it can effectively hold a conversation about it with humans. Even as that conversation twists and turns around an unknown number of curves. If the training data is sufficient, the chatbot can keep up.

The virtual graduate assistant is another example. You can offload a specific task, such as analyzing data to find trends, researching and summarizing prior art, or even interpreting legal briefs. 

But organizational augmentation comes with a required expertise and a financial cost. 

It requires programming your own application and incurs a dynamic cost across the interface to the LLM. To bring virtual bots into your workflow or team, you must use the LLM’s Application Programming Interfaces (APIs). Which means, “you have to write your own code to create the bot.” And you have to pay for every piece of information you send and receive to the LLM. More data == more cost.

Organizational automation is where it starts. Maybe you need less helpdesk people. Maybe you need two grad assistants instead of three. Maybe you don’t need to hire out some specific task.

But at the end of the day, the effect is minimal.

Next, we’ll look at the holy grail — Autonomous Automation.