What to Expect from AI

Nobody yet knows how profound artificial intelligence will be as a tool for creating new business models, revenue streams, reducing costs or boosting productivity. But it is probably not too soon to argue that the measurable gains will be of the “process” sort that reduce costs, as that arguably has been the primary measurable benefit of virtually all new information and communications  technologies for decades. 

New types of information technology can reshape existing business models, oftentimes revenue models and nearly always profit margins or operating costs. Consider the automated teller machine and online and mobile banking. Over the last few years, the data suggests customers are conducting more transactions in total, using multiple channels. 

Decade	Percent of Transactions at Branches	Percent of Transactions at ATMs	Percent of Transactions Online
1960s	99%	0%	1%
1970s	95%	5%	0%
1980s	90%	10%	0%
1990s	80%	20%	0%
2000s	60%	40%	0%
2010s	40%	60%	10%
2020s	20%	80%	20%

Assuming that cost savings happen because of reduced labor cost and physical facilities, a shift of nearly 80 percent of transactions to ATMs or online logically reduces cost. 

Many studies purport to show both cost savings and revenue boosts because of cloud computing. 

Study	Authors	Publisher	Date of Publication	Cost Savings	Revenue Boost
The Economic Impact of Cloud Computing	McKinsey & Company	McKinsey Quarterly	2011	30-70%	10-20%
The Business Value of Cloud Computing	Gartner	Gartner Research	2012	35-50%	10-20%
Cloud Computing: The Ultimate Cost-Savings Tool	Rackspace Hosting	Rackspace White Paper	2013	30-70%	5-10%
The Cloud Revolution: How Cloud Computing is Changing Business	IDC	IDC White Paper	2014	30-60%	10-20%
The State of the Cloud 2023	RightScale	RightScale Report	2023	35-50%	15-20%

Studies of software as a service likewise tend to show cost savings compared to legacy “shrink wrap” delivery modes. 

Study	Authors	Publisher	Date	Cost savings	Revenue boosts
"How Cloud Computing Helps Cut Costs, Boost Profits"	John Engates	CIO	2013	88% of cloud users pointed to cost savings	56% of respondents agreed that cloud services have helped them boost profits
"Sales automation: The key to boosting revenue and reducing costs"	McKinsey & Company	McKinsey Quarterly	2017	Early adopters of sales automation consistently report increases in customer-facing time, higher customer satisfaction, efficiency improvements of 10 to 15 percent, and sales uplift potential of up to 10 percent	-
"Maximize SaaS Revenue: Boost Conversions with Appointment Scheduling Software"	Leadmonk	Leadmonk Blog	2019	Adopting appointment scheduling software can save you 8 hours ($400) and your clients 2 hours ($200) per month, resulting in a total monthly savings of $600	-

Likewise, many studies suggest internet of things technologies and products likewise have benefits for cost reduction and revenue lift.

Study	Authors	Publisher	Date Published	Cost Savings (USD)	Revenue Boost (USD)
The Internet of Things: Moving from Cost Savings to Revenue Generation	Michael Chui, James Manyika, Michael Osborne	McKinsey & Company	2015	100 billion - 1 trillion	1 trillion - 10 trillion
The Business Value of the Internet of Things	Gregory P. Rogers, Michael Chui, James Manyika	McKinsey & Company	2016	3.9 trillion - 11.1 trillion	6.2 trillion - 22.2 trillion
The IoT Revolution: The Internet of Things Is Transforming How We Live and Work	Richard Evans	Apress	2017	1.9 trillion - 3.9 trillion	3.8 trillion - 7.8 trillion
The Impact of the Internet of Things on Business	IDC	IDC	2018	4.5 trillion - 19.4 trillion	9 trillion - 38.8 trillion
The Global Economic Impact of the Internet of Things	Cisco	Cisco	2019	14.4 trillion - 45.4 trillion	21.2 trillion - 69.6 trillion

We always can argue about the study assumptions and the degree of outcome attribution when multiple inputs operate. We might contest the causal effects with effects that are merely correlated. 

We might contest the magnitude of “social surplus” created by applied new technologies. We can disagree about which firms and industries saw benefits and which saw losses, and to what extent. 

But it is safe to say enterprises would not keep investing in information technology if they did not believe they obtain benefits. 

Nor is it incorrect to argue that consumers have been the beneficiaries almost universally. 

Study	Authors	Publisher	Date of Publication	Cost Savings (USD)	Revenue Boost (USD)
The Economic Value of Ride-Sharing in the United States	David E. Mills and Christopher R. Knittel	Journal of Urban Economics	2018	6 billion	10 billion
The Impact of Ride-Hailing on Car Ownership and Use	Michael Sivak and Brandon Schoettle	Transportation Research Part A	2017	4 billion	5 billion
The Benefits of Ride-Hailing for Urban Transportation	Michael E. Webber and Asif Faiz	Nature Energy	2019	100 billion	150 billion
The Social Surplus of Ride-Hailing Services	Yongjie Sun, Jie Zhang, and Fei Wang	SSRN	2020	96 billion	130 billion

It is perhaps not too soon to suggest that, as has been the case in the past, most of the identifiable AI financial and business model impact will come from more-efficient operations of all sorts. 

In some cases, we should also see new business and revenue models created as well, much as ride sharing, lodging and other product substitutes have developed because of ubiquitous smartphones, internet access, software as a service, remote cloud computing and cheap processing and storage. 

Does 5G Drive Higher ARPU?

It is totally understandable that suppliers of 5G infrastructure and many service providers are anxious to show that investments in 5G networks are driving higher revenues, profit margins or subscriptions. So Ericsson argues that 5G subscriptions are correlated with higher revenue earned by mobile service providers. The implication might be that the two are causally related. 

“The launch of 5G services in the top 20 5G markets is followed by a positive revenue development of 3.5 percent CAGR compound annual growth rate over the last two years, or a total of seven percent,” Ericsson argues. 

Perhaps 5G is “responsible” for that growth, in some cases. In other cases, retail packaging policies might be the driver. In the former case, perhaps 5G subscriptions, all other things remaining equal, are priced at higher levels than 4G subscriptions. Then a reasonable case can be made that 5G actually is driving higher revenues.

source: Ericsson Mobility Report 

The same assurance is not possible when other key elements of retail pricing also are changing, even if some 5G service plans are priced identically to comparable 4G plans. 

Ericsson itself says that “only 22 percent of the 182 service providers offering 5G are showing a price difference which the consumer needs to pay in order to gain access to 5G services.” So in roughly four out of five instances, higher revenue has to be explained some other way. 

For example, 5G might be available for “no extra charge” on some higher-priced service plans, so that customers might in some sense be buying “more data, unlimited data or something else” and getting 5G as part of the package. 

To be sure, greater capacity provided by 5G does enable new thinking on the economics of “unlimited usage” packages sold for higher prices. “Service providers with 5G commonly have more unlimited packages available,” says Ericsson.

The point is that it is quite difficult to determine whether 5G itself leads to average price increases, or whether some other drivers (such as unlimited usage plans) actually are driving higher revenue per account.

Will AI Mostly Produce "Faster and Cheaper" or "Better?"

Despite all the hype, artificial intelligence is going to produce benefits and outcomes that are primarily quantitative--faster or cheaper--rather than qualitative, where “better” can include the ability to create whole new products and industries, plus revenue and business models, that didn't exist before. 

Enterprises justify information technology investments primarily because they deliver quantitative outcomes in terms of productivity, revenue, cost savings, risk reduction or higher productivity. These outcomes are important because they can be measured in some way, and are tangible outcomes. 

In that sense, IT investments might be evaluated--if with great difficulty--the same way all other investments get evaluated: using standard accounting metrics such as net present value, return on investment, payback period, internal rate of return, total cost of production or economic value added. 

Net Present Value (NPV)	NPV is one of the foremost financial key performance indicators (KPIs) used to evaluate large, capital-intensive IT projects. NPV relies on accurate cash flow projections extending over the life of the project, alongside a discount rate which is used to account for the time value of money. Project approval depends on obtaining a positive NPV. IT projects can also be compared with one another using NPV whenever firms need to ration scarce IT capital.
Return on Investment (ROI)	ROI is an accounting-based ratio that compares total project income to the level of project investment. ROI does not take account of the time value of money, meaning that projects with a longer-term return window would be treated on par with projects that generate equal returns over a shorter time period. Similar to NPV, the accuracy of ROI calculations depends on being able to identify the scale of future cash flows arising from an investment.
Payback period	The payback period is a simplistic method that calculates the time needed for a project to breakeven (recover its investment costs). In a risk averse firm, managers may gravitate towards IT projects with a shorter payback period. In practice, payback should not be used in isolation but rather alongside other metrics that take account of project risk and that consider the flow of benefits beyond the end of the payback period.
Internal Rate of Return (IRR)	Given all future cash flows and an upfront investment for an IT project, IRR is the discount rate that would return a value of zero for NPV. IRR can be considered the true rate of return in that it takes account of the time value of money and the flow of value over time. IRR can be benchmarked against desired or minimum rates of return, including the weighted cost of capital.
Economic Value Added (EVA)	EVA —also called economic profit—is a measure of residual value generated by a project after deducting the cost of invested capital. Since all capital can be allocated to different ends, EVA argues that projects should be assessed an investment cost. This allows for a more equitable comparison if managers are in a position to pick from different IT projects with unique rates of return.
Total Cost of Ownership (TCO)	TCO captures a multitude of different cost items in a single metric such as the cost of hardware, software, and services, allocated per application, user, department, etc. TCO can also be represented as a cost per period of time. TCO does not take into account the benefit or value to the organization of using the underlying resource and is, as such, a questionable metric unless accompanied by other metrics such as ROI, NPV or payback period.

source: Springer 

Whether the inputs to be measured are the application of mainframe, minicomputer or PC-based computing, use of client-server architectures, world processing, spreadsheets, business software, video streaming or digital transformation, the measured outcomes always seem to focus on quantitative improvements: faster or cheaper. 

Less tangible outcomes are cited, but are hard to quantify, such as brand enhancement or competitive advantage in core markets. Perhaps least-cited of all is the application of information technology to create entirely new products and industries. Such qualitative outcomes fall under the framework of “better.” 

source: ISSA

The use of mainframe computers in the 1960s and 1970s led to average  productivity gains in businesses of about 30 percent, according to Datamation. 

A study by the Aberdeen Group found that companies that invested in information technology had an average cost savings of 12 percent. A study by IDC found that companies that invested in IT had an average production rate increase of 15 percent.

A study by the Gartner Group found that companies that invested in IT had an average error rate reduction of 20 percent. A study by McKinsey found that companies that invested in IT had an average innovation cycle time reduction of 30 percent. 

A study by the Aberdeen Group found that businesses that invest in IT achieve an average ROI of 22 percent. A study by Gartner found that businesses that invest in IT can improve their customer satisfaction by an average of 15 percent.

A study by McKinsey found that businesses that invest in IT can reduce their costs by an average of 10 percent.

That is not to say “qualitative” outcomes are not cited. They simply are hard to quantify. Most would likely agree that the introduction of personal computers in the 1980s led to qualitative benefits for businesses that include improved decision-making, increased collaboration, and better customer service.

Smartphones, digital cameras, personal computers before them, perhaps tablets, videogame consoles, social media and the web are easy-to-understand applications of technology that created new products and industries, or, at the very least, transformed existing industries. 

Many similar claims could be made for the benefits of the internet, cloud computing and will be made for AI as well. In the end, perhaps most of the claimed benefits will be quantitative in nature. “Faster and cheaper” will be the advantages cited. 

Still, in some cases, “better” will be the outcome: whole new products, industries and revenue models will also be created. But that should not be the “main” outcome. Process improvements are likely to dominate, as they have in the past. 

The Growth Conundrum

The current effort by some “telcos” to become techcos, as with earlier language about becoming “agile” or “virtualized” or “cloud native” “solution providers” has its roots in a simple business problem: business leaders no longer see revenue or profit growth coming from the traditional “connectivity” business. 

Competition, a fundamental switch to the internet as the driver of application development, resetting of value creation and consequent commoditization of “connectivity” products all contribute to a search for new value drivers. 

All of that is somewhat perplexing. On one hand, the role of “connectivity” in the internet ecosystem is essential. On the other hand, the relative value creation seems challenged. So connectivity firms are valued by the financial markets as “slow growth” businesses, though arguably with more growth potential than most other “utility” type businesses and industries. 

Revenue Growth Rate	P/E Multiple
Slow growth (<5%)	Low P/E (<10x)
Moderate growth (5-10%)	Average P/E (10-15x)
Fast growth (>10%)	High P/E (>15x)

But that does raise a recurring question: what can “slow growing” connectivity providers do to enhance their growth profiles and escape commoditization of prices within the industry? If one believes something can be done, how much upside might there be? And can connectivity firms actually escape their roles in the value chain?

Traditionally, efforts to create additional roles have had mixed success. And even when success has been obtained, it often makes more sense to spin out those assets to reap the rewards of higher valuations than are possible when the assets are essentially trapped within a connectivity provider business, with a connectivity provider valuation.

With the caveat that one’s perspective can vary if different metrics are chosen, and with the proviso that there can be nearly as much divergence between firms within a single industry as between them, valuation differences are clear enough. 

Industry	EV/EBITDA Ratio
Social Media	25.0x
Search	20.0x
Cloud Computing as a Service	15.0x
Data Centers	12.5x
Computing Hardware	10.0x
Computing Software	7.5x
E-commerce	5.0x
Mobile Telecom	2.5x
Fixed Network Telecom	1.0x
Airports	0.5x
Natural Gas Utility	0.25x
Electrical Utility	0.125x
Commercial Real Estate	0.0625x
Commercial Air Transportation	0.03125x
Energy Producer	0.015625x
Logistics	0.0078125x

So it is hard to avoid the conclusion that connectivity providers have to create new roles in the value chain if they wish to grow revenues. But once created, there will be high incentives to spin out or otherwise separate those assets to reap the rewards of higher valuations. 

So the ever-present conundrum is the tension between needing to create higher value roles and the pressure to liquify those assets to reap the market value. 

Connectivity might always be a slow-growth business, as it has been in the past. It might face growing competition and therefore commoditization in the future, as it has in the past. Though it always will be important to look for capital and operational efficiencies, it is hard to imagine big new “connectivity” revenue sources with sustainable margin profiles.

Most of the more-lucrative opportunities seem always to lie elsewhere in the stack or value chain: devices, apps, computing services and so forth. 

So the point is that no matter what new buzzwords are applied, connectivity provider strategic options remain challenged. There is only so much business leverage to be obtained within the connectivity space. Transformative growth almost has to be found elsewhere.

But, in the end, even when successful, those assets realize their value only when parted from the connectivity provider. It’s a conundrum indeed.