Generative AI: Navigating AI implementation costs and calculating ROI

Implementing AI is no simple task; it has financial implications and challenges. As companies increasingly adopt AI, understanding the costs and calculating Return on Investment (ROI) becomes crucial to justifying and guiding their AI strategy.

In 2024, AI implementation costs vary widely depending on project complexity, integration requirements, and ongoing maintenance. On the other hand, the benefits can be divided into tangible and intangible, making justifying ROI even more challenging. That said, only a holistic approach justifies the initial investment and supports an ongoing commitment to AI.

Let’s dive into the typical costs of implementing AI, examine the variables involved, and explore ways to calculate and maximize ROI from AI investments. In addition, we will explore tangible and intangible returns.

Third-party software vs. Custom solutions

One of the first decisions companies must make when implementing AI is whether to build custom solutions or rely on third-party software developed by tech giants like Microsoft or OpenAI. Each approach comes with its own set of pros and cons.

Third-party AI software: Third-party solutions are faster to deploy, particularly during the proof-of-concept phase. They provide ready-made frameworks that can quickly be integrated into existing systems. Annual costs for third-party AI software can range from €0 in software fees (for example, Tensorflow as an open-source option) to €40,000 or more, depending on the provider and the scope of usage.

Custom AI solutions: For industries with strict security or compliance requirements, custom-built AI systems offer better control, technical flexibility, and heightened security. Although more expensive upfront, custom solutions enable businesses to tailor the AI precisely to their needs. Moreover, building an in-house system can provide more control over future development operational costs and potentially provide a great ROI.

Cost estimates for custom AI solutions

Costs for custom AI solutions range in price based on project complexity:

Simple Projects: €10,000 to €50,000

Medium and more complex projects can be as high as €1,000,000+

There is a wide spectrum of AI development costs, and multiple factors contribute to these expenses.

Factors affecting AI costs

Several factors influence the overall cost of AI implementation:

1. The complexity of the AI system plays a major role in determining costs. AI solutions with advanced features such as Natural Language Processing or computer vision are more expensive to develop. Additionally, the number of capabilities required and the scale of the system impact the cost.
2. AI systems rarely operate in isolation. Integration with existing IT infrastructure is necessary for a smooth workflow. Integrating AI with legacy systems or environments that use outdated technologies, however, can be challenging.
3. Increased data security and regulatory compliance have become paramount. AI projects often require customized security measures to protect sensitive data, which can increase implementation costs. Moreover, adhering to data protection laws such as GDPR in Europe or CCPA in the U.S. is critical but costly. At Smartbi, we've found that developing customized, tailored AI and ML solutions offers better security than relying solely on off-the-shelf applications.
4. Once the AI system is up and running, regular maintenance and updates are required to keep the AI model functioning at its best. These ongoing expenses, including model retraining, system fine-tuning, and regulatory compliance updates, typically add 15-20% of the original development cost per annum.

Understanding and calculating AI ROI

Calculating AI's ROI can be more complex than traditional IT projects due to AI's wide range of tangible and intangible benefits over time. Let’s break down how to approach understanding and calculating AI ROI for industrial companies, along with concrete examples.

Key aspects of AI ROI

When calculating the ROI for AI projects, it's crucial to consider both tangible (hard) and intangible (soft) returns because they provide a holistic view of the project's overall value:

Tangible ROI: the quantifiable financial benefits directly attributable to AI. These include revenue increases, cost reductions, productivity improvements, and risk mitigation.

Intangible ROI: though harder to quantify, intangible ROI is equally valuable. Intangible returns on investment include improvements in decision-making, enhanced customer experiences, and increased employee satisfaction or safety. While more difficult to measure, these intangible benefits drive long-term business success and foster a competitive edge.

ROI calculation methods

There are several methods to calculate the ROI of AI projects.

Simple ROI calculation: To calculate ROI, you divide the net profit generated by the AI system by the total cost of AI development and then multiply by 100.

ROI = (Net Profit from AI ÷ Total Cost of AI development) × 100

Payback period: The time it takes for the benefits from the AI implementation to equal the initial investment cost. For example, if a company invests €500,000 in an AI project and achieves a monthly savings of €50,000, the payback period would be ten months.

Comprehensive ROI example – GenAI for documentation and reporting in manufacturing

GenAI solutions are increasingly being used in factories to assist workers with retrieving manuals, generating reports, and producing documentation. These systems help streamline workflows, making employees more effective and efficient.

A custom GenAI assistant system that improves workplace safety and ensures regulatory compliance via real-time monitoring and automated risk assessment.

Initial investment: €200,000 on a custom GenAI assistant system. This GenAI assistant system is a custom intelligent safety and compliance assistant that assists managers by automatically generating and updating compliance documents, analyzing workplace incident investigations, and providing personalized safety and training for employees involved in incidents.

Cost savings

Time savings:

The GenAI assistant helps workers save time, resulting in €120,000 in annual savings.

Increased productivity:

Improved efficiency increases factory productivity by 5%, adding €40,000 in annual revenue.

Labor cost reduction:

The GenAI system reduces the need for manual documentation and reporting, saving €70,000 in labor costs per year.

Total Annual Benefits:

€120,000 (Time Savings) + €40,000 (Increased Productivity) + €70,000 (Labor Cost Reduction) = €230,000

Net Profit from AI:

Net Profit = Total Annual Benefits - Initial Investment

Net Profit = €230,000 - €200,000 = €30,000

Total cost of AI development:

Total Cost = Initial Investment = €200,000

Apply the ROI formula:

ROI = (Net Profit from AI ÷ Total Cost of AI Development) × 100

ROI = (€30,000 ÷ €200,000) × 100

ROI = 0.15 × 100 = 15%

In this example, our GenAI solution yields a 15% ROI in its first year of implementation.

The initial investment will be recovered within the first year, as the annual benefits (€230,000) exceed the initial cost (€200,000).

In this example, while the tangible ROI is valuable, it's crucial to consider the intangible ROI that contributes to long-term value.

Improved decision-making

-> The GenAI solution provides data-driven insights, potentially reducing errors by 20%.

Estimated value: €50,000 annually in avoided costs and improved outcomes.

Increased employee satisfaction

-> Employees report 25% higher job satisfaction due to reduced mundane tasks.

Estimated value: €40,000 annually in reduced turnover and improved productivity.

Improved safety

-> AI-assisted safety protocols lead to a 15% reduction in workplace incidents.

Estimated value: €60,000 annually in reduced insurance costs and avoided lost-time injuries.

Total estimated value of intangible benefits: €150,000 annually

In conclusion, with a solid business case backing it, a well-designed GenAI solution can provide a positive, tangible ROI from the first year and deliver ongoing significant intangible benefits. While calculating intangible benefits takes time and effort (requiring estimations and double-checking), only by considering both the tangible and intangible benefits does the true value of AI become clear.