Will AI Value Driver be Same, or Different, from Internet Impact?

Some of us believe artificial intelligence will have as big an impact as the internet did, which is to say existing industries and whole business models and processes will be disrupted, while at least a few entirely new industries could arise. 

The issue is how that impact will be primarily felt, though. If the internet largely reduced marginal costs and therefore enabled global platforms to emerge, what will AI bring?

Two broadly-different drivers of outcomes might happen: AI reshapes processes in roughly the same way the internet did, or AI reshapes processes in a new way. 

In other words, if the internet primarily recase marginal cost, AI might work the same way. If AI automates processes, it could likewise lower marginal cost of any operation. 

On the other hand, AI, by enabling massive personalization and customization of products and experiences; allowing faster innovation based on better research and development processes; might drive change primarily by allowing new products to be created and discovered. 

The primary change driver would be as a value multiplier more than a marginal-cost reduction mechanism. 

Use of generative AI to support customer contact operations is an example of reducing marginal cost by automation. Use of AI to tailor product recommendations to particular customers is another form of applied automation or lower operations cost value. 

On the other hand, it is conceivable that AI drives value primarily by enabling new product creation. AI already is used to develop new drugs and medical treatments, create new financial products, and design new software applications.

So while it is possible that AI works in both ways--to reduce costs and increase innovation--it seems likely that most of the impact could come in one of those ways, in the same way that almost all the value created by the internet revolves around lower marginal cost. 

Feature	Internet	AI
Primary Impact	Reduced marginal cost of production and distribution	Automation, personalization, and new product and service development?
Key Enabler	Connectivity, information access and global reach	Data analytics, machine learning, and predictive capabilities?
Impact on Business Models	Created new business models like ad-supported technology, sharing economy, and crowdfunding	Does AI improve existing models or enable new ones?
Impact on Business Processes	Streamlined processes, reduced costs, and increased efficiency	A mix of efficiency and quality gains
Impact on Industries	Disrupted traditional industries, created new industries	What mix of disruption and creation?

Right now, it seems as though AI will have its greatest impact as a value multiplier or insight generator or cognitive augmentor,  where the internet has the greatest impact as a marginal cost reduction mechanism. 

In other words, AI might be about analytics, insights and cognitive augmentation rather than marginal cost reduction, though it might well have that impact as well.

What is Software Product "Distribution" Cost, and are Google's Fees Paid to Apple Unreasonable?

Some observers are likely shocked that Google pays Apple 36 percent of ad revenues earned by Google search on Apple phones. 

Some of us might be nonplussed. Distribution costs for lots of products routinely amount to 25 percent or more of the total cost of getting a product from factory or source to end user customer. 

Industry	Product Type	Distribution Cost as Percentage of Retail Price
Software	Software Applications, Games	10-20%
Financial Services	Investment Advice, Financial Planning	15-25%
Legal Services	Legal Advice, Representation	15-25%
Professional Services	Consulting, Training	15-25%
Technology	Software, Web-based Services	5-15%
Agriculture	Fresh Produce	15-25%
Apparel	Fashion Apparel	10-15%
Consumer Electronics	Smartphones, Tablets, Laptops	5-10%
Food and Beverage	Packaged Food, Beverages	8-12%
Furniture	Home Furnishings, Appliances	10-15%
General Merchandise	Household Goods, Personal Care Products	12-18%
Pharmaceuticals	Prescription Drugs	20-30%
Retail	General Merchandise, Apparel	10-15%
Sporting Goods	Sports Equipment, Apparel	10-15%
Toys and Games	Children's Toys, Games	12-18%

Distribution costs for intangible products such as advice, consulting, many software products arguably are lower, but there's a difference between price (value) and cost. Actual “distribution” costs are hard to separate from the actual “performance” or “delivery” of the product. 

In some sense, for intangible services, “production” is the same as “delivery” or “distribution.” 

Unlike tangible goods, intangible services are produced, consumed and distributed simultaneously. So “distribution” as a percentage of total cost is a bit of a misnomer. Production cost is hard to separate from “distribution.” 

Advertising placements, like airline seats, are intangible and perishable products in the sense that they cannot be stored or saved for later use. Once an advertising availability is sold--or not sold--it is gone forever. 

Unlike tangible products, such as cars or books, ad avails cannot be stored in a warehouse and sold at a later date.

Additionally, both advertising inventory and airline seats are subject to demand and supply. If there is high demand for advertising placements, the price will go up. 

Also, supply can create demand. In one sense, having Apple smartphone search inventory is a form of demand creation, not simply supply expansion, as default use of Google is believed to create more demand (use of the product) by Apple customers and users. 

As a result of these similarities, advertising inventory and airline seats are often treated similarly in the marketplace. For example, both products are often sold through auctions, where the highest bidder wins the right to use the product. Additionally, both products are often subject to dynamic pricing, where the price of the product fluctuates based on demand.

The point is that “distribution cost” for an intangible product can be quite a bit higher than what we might traditionally believe, since “production cost” is bound up with “sales” and “delivery” cost. 

For a software product, “distribution can represent as much as 50 percent to 70 percent of total cost, in that sense. So a fee of 36 percent to handle “distribution” might not seem worrisome or out of line. 

Intangible Product Category	Estimated Distribution Cost (%)
Software	50-70%
Music	30-40%
Video	20-30%
E-books	15-25%
Online courses	10-20%

AI will Reshape Almost Every Industry, Including Journalism

Where Will AI Drive the Most Value, Across Industries and Functions?

Artificial intelligence is expected to affect nearly every industry, though some industries should benefit more, in terms of affect on output or productivity. Generative AI alone could add the equivalent of $2.6 trillion to $4.4 trillion annually across the 63 use cases analyzed by McKinsey consultants. 

About 75 percent of that value would  be added by GenAI application to customer operations including marketing and sales, software engineering, and research and development, McKinsey says. 

source: McKinsey

But it is not clear that the industries which benefited most from prior waves of information and communications technology will reap the biggest rewards from AI. 

According to job site Indeed, generative AI, for example, is going to supplant more human activity in software development than in driving; more replacement in information technology help desks and less for beauty or wellness jobs. 

source: Indeed

“Unlike prior advances in robotics and computing that largely impacted manual labor, roles filled by knowledge workers are potentially the most exposed to change from generative AI,” an Indeed study says. 

 

The other methodological issue is that it is difficult to impossible to measure the productivity of office or knowledge work, as “output” is hard to measure. 

Whatever we may prefer to think, the introduction of several waves of information and communications technology, from personal computing to the internet to mobile service to cloud computing has had a disparate impact on various industries, if we can extrapolate from productivity changes, and assume that information technology has played a significant role. 

Though estimates obviously vary, a reasonable assessment of information and technology impact across industries might show a pattern that shows disparate impact, in terms of productivity outcomes. 

Generally, observers would tend to cite agriculture, manufacturing and finance as among those industries which saw the most change in productivity since 1980, with government and education routinely seen as benefiting the least, if at all. Obviously, various studies and estimates vary in the precise degrees of industry impact and in some cases other changes might be more important than computing or communications changes in explaining changes. 

In agriculture, for example. Most analyses would credit genetically modified crops, precision farming techniques, and advanced machinery for the productivity gains. Obviously, precision farming is the only innovation directly provided by computing, information and communications technology. 

Improved crop breeding, better land management practices, Expanded access to inputs, improved seeds, fertilizers, and pesticides, globalization and trade, government policies, 

increased demand for food as incomes have risen, global economic growth, Infrastructure development and even education and training have played important roles in boosting agricultural productivity, many would argue.

Industry	Productivity Change (%)
Agriculture	175%
Transportation	90%
Finance	50%
Retail	40%
Manufacturing	30%
Healthcare	20%
Education	10%
Government	0%

Industries expected to benefit most from AI include those with lots of routine, repetitive tasks, large amounts of data, and complex decision-making processes. For example, AI is expected to have a major impact on the transportation, manufacturing and healthcare industries, retailing and finance. 

The industries that are expected to be least affected by AI are those that are characterized by human creativity and interaction. 

Industry	Expected Productivity Impact
Transportation	High
Manufacturing	High
Healthcare	High
Retail	High
Finance	High
Education	Medium
Government	Medium
Agriculture	Medium
Mining	Medium
Construction	Medium
Hospitality	Medium
Entertainment	Medium
Personal Services	Low
Arts and Culture	Low

The point is that some industries might benefit from AI more than they have from prior waves of information technology innovation. Healthcare is among those industries that might benefit quite a bit more than in the past. 

AI should facilitate more accurate diagnoses, personalized treatment plans, and therefore improved patient outcomes, based not only on a single patient’s history and data, but on all accumulated patient data. 

AI also should allow better proactive and preventive care, again using the accumulated data from “all” patient histories. AI can facilitate the development of personalized treatment plans tailored to individual patients' genetic makeup, medical history, and lifestyle factors.

AI is expected to play a big role in accelerating the drug discovery process by analyzing vast datasets of molecular interactions and identifying potential drug candidates. Automated procedures can streamline workflows, reduce errors, and improve efficiency.

Perhaps the important observation is that generative AI alone should have value across most industries in reshaping sales, marketing, software engineering, research and development and customer operations.