The following are the main sections, subsections and topics that you’ll find in my book, The Shadow Balance: The Technology Leader’s Guide to the Hidden Costs and Risks of AI. Individual pages will apper within the sections as their podcasts are released.

Costs

The cost profile of any technology will change when AI is introduced. If the change is not done with due consideration of costs, it can result in some very surprising bills – with AI-powered technologies driving annual increases of 30% in cloud spending. Understanding the change in costs will help you develop software using AI or consume a technology that is introducing AI. This better understanding can protect you from vendors who are not currently passing full costs on to their customers and whose businesses are built on the hope that they can increase costs once they reach a certain level of adoption.

In this section, you’ll learn about the capital and operational costs that are impacted by AI by asking:

• What is the nature of this cost?
• How is AI changing the amount and relative cost?
• How can I manage and control this cost?

You’ll learn how traditional costs are changing, what new costs to expect and techniques you can use.

Structural

While we celebrate the opportunities presented by AI, many of us are ignoring or unaware of the increasing structural costs required to run AI at scale. In this section, you’ll learn about the ways that AI is changing your structural costs and the importance of right-sizing capacity and storage to meet, but not exceed, demand.

Execution

With traditional software, running a system doesn’t cost more just because your teams used it more. In the AI era, every single query, execution, and data request comes with a metered invoice, which is leading to highly variable operational expenses. In this section, you’ll learn what changes to expect to your execution-related costs when introducing AI and how to manage them.

Observability

Observing how an AI is functioning requires much more complex governance than traditional systems and, to manage costs, the complexity must be managed strategically to balance regulatory requirements and costs. In this section, you’ll learn about the way that logging and monitoring costs are changing with the introduction of AI and how to manage them.

AI Model

The lifecycle of an AI has a different cost profile than most software, with the training phase requiring massive compute investment and the inference phase having a much larger cost per transaction. In this section, you’ll learn about the different phases of an AI’s lifecycle, what types of costs to expect and how to plan for them.

Benefits

The potential benefits of using AI are incredibly exciting. With that said, realizing the hyped and promised benefits can be challenging. In this section, you’ll learn how to realize these benefits by asking:

• How does AI enable the realization of benefits for users and businesses?
• How is AI impacting the delivery of these benefits?
• What investments are needed to overcome barriers?

You’ll come away with a more pragmatic understanding of what it will take to achieve the benefits promised by AI for your products, your users and your company.

Acceleration

AI can accelerate work by improving time to completion, time to make quality decisions and overall delivery timelines.
In this section, we’ll examine some of the ways that AI is supporting organizations and software delivery teams to improve their velocity and some of the barriers preventing them from realizing the promised, increased cadence.

User Experience

AI is promising to absolutely transform the user experiences offered by software by leveraging information about its user to make decisions in real-time. In this section, you’ll learn about several ways AI enables improvements to traditional user experiences as well as some of the barriers that are preventing organizations from realizing those benefits.

Risks

While the potential of AI is enormous, the associated risks can be far-reaching and can have significant impacts on your teams, your profitability and your business’ strategy. And like any new technology, in order to conscientiously introduce AI, it’s important to understand and assess its potential risks and associated impacts. In this section, you’ll learn some of the new and increased risks that AI introduces by asking:

• What is this risk and how could it impact you?
• How is AI changing this risk and its impacts?
• What can you do to manage these risks?

You’ll gain a better understanding of some potential ways AI could impact you and your business, as well as some actions to support you in managing the potential risks while still realizing the promised benefits.

Data Breach

AI introduces an entirely new set of sophisticated, AI-specific data vulnerabilities, which sophisticated attacks can exploit to leak sensitive and proprietary information. In this section, you’ll learn about the risk of AI-specific attacks and some approaches that could help you protect your information.

Adoption

AI-enhanced technologies are constrained in the same way as any technology investment: to realize value from AI, you’ve got to ensure your users trust and use the system you’ve built. In this section, you’ll learn about the user adoption risks that exist with AI and how to manage them before they impact morale, team performance or the quality of work.

Efficiency

Just using AI isn’t going to drive efficiency. You need to manage the efficiency risks that AI introduces to a user’s workday related to the correctness and accuracy of the provided responses, and how well a user can evaluate them. In this section, you’ll learn about several risk areas that need to be managed in order to protect costs, improve performance and ensure your AI can continue to deliver value as it scales.

Reliability

AI can sometimes generate unreliable responses that cause all sorts of trouble: eroding the trust in the system itself, embarrassing employees and impacting your organization’s brand. In this section, you’ll learn about two types of common AI errors, the risks associated with them, and approaches you could use to mitigate them so that your AI remains a reliable asset.