I implement AI in my company, is it feasible?

Automating product quality decisions: Comparing organizations and Open-source vs. Commmercial AI tools.

Created by DALL-E

Everyone talks about the implementation of AI tools in our companies but no one talks about whether it is feasible, possible and/or necessary to implement it in my company, and I am not talking about using ChatGPT and/or any of the many tools created during the hype of this technology, I am talking about the effective implementation of an AI tool to automate decisions during the day to day operations of a company.
This analysis reveals the nuanced financial and operational considerations that a manufacturing company must evaluate when deciding between commercial and open source AI solutions to automate quality decisions (key decisions for any company’s survival). The choice depends on the company’s specific product mix, production volume, existing profit margins and strategic financial thresholds. It is recommended that the company perform a thorough cost-benefit analysis, taking into account both the short-term impact and long-term efficiencies of implementing AI. Balancing the upfront investment costs with the potential for significant operational improvements and personnel cost savings is key to making an informed decision tailored to the company’s unique circumstances and growth ambitions.
I hope it helps you make better decisions! Lets start!

I have chosen the following information/company as an example;
manufacturing company wants to automate decisions where the human operator has to interact with the decisions of whether a product is considered good or bad (easy). We should keep in mind that implementing an AI tool to automate product quality decisions in a manufacturing company can vary significantly in cost depending on the specific requirements, the complexity of the manufacturing process and the AI solution chosen.

Let’s get into the details:

-Manufacturing Company Overview

• 1,000 different products manufactured per month.
• 6 manufacturing process steps per product.
• 10 characteristics to control per process step.
• 10,000 units produced per year for each of the 1,000 products.
• Product sales range from $1 to $3 per unit.
• Current average product margin: 20%.

-Comparisons between AI type:

Open-Source LLM Tool

Initial Implementation Cost: $150,000 — $225,000 Annual Recurring Cost: $20,000 — $25,000

To amortize the open-source LLM tool investment over 5 years:

• Minimum required margin per unit: $0.05 — $0.07
• Percentage of current 20% margin required: 13.33% — 20%

Commercial AI Tool (e.g., Claude, ChatGPT)

Initial Implementation Cost: $200,000 — $250,000 Annual Recurring Cost: $106,250 — $160,000

To amortize the commercial AI tool investment over 5 years:

• Minimum required margin per unit: $0.16 — $0.22
• Percentage of current 20% margin required: 35% — 55%

-Key Considerations

• The open-source LLM tool has a lower initial implementation cost and lower annual recurring costs.
• The commercial AI tool has a higher initial implementation cost and significantly higher annual recurring costs due to licensing and subscription fees.
• The open-source LLM tool requires a lower percentage of the current 20% product margin to amortize the investment, making it a more attractive option if the company can maintain its current margin.

Based on the analysis, the open-source LLM tool appears to be the more cost-effective solution for the manufacturing company, as it requires a lower percentage of the current product margin to amortize the investment over 5 years. The company should carefully evaluate the long-term costs and benefits of each option based on their specific requirements, budget, and expected product margins.

Let’s see now the comparison between companies with different turnover volume and the same indicators, but now adding 2 more variables; the current operation (quality decision) is carried out by 6 people, with a gross annual salary between 30,000$-35,000$/year, and with the implementation of the tool, these 6 operators could be replaced by a person to control and maintain the IA system with a salary of 50,000$/year and another person responsible for monitoring the IA decisions, this one with an annual salary of 40,000$/year.

See where we are and what kind of company we identify with.

Annual Revenue Levels for Open-Source LLM Tool Consideration:

• $10 Million Annual Revenue: With the potential cost savings from automating the work of 6 people, the open-source LLM tool becomes a viable option even at this revenue level. The initial implementation costs ($125,000 — $250,000) would be offset by the annual personnel cost savings, making it a worthwhile investment.
• $20 Million Annual Revenue: At this revenue level, the open-source LLM tool is a very compelling option. The initial implementation costs would be easily covered by the personnel cost savings, and the ongoing annual costs ($15,000 — $30,000) would be a small fraction of the overall revenue.
• $50 Million Annual Revenue: The open-source LLM tool is an excellent choice at this revenue level. The initial and ongoing costs would be insignificant compared to the potential personnel cost savings, and the company would benefit from the automation and efficiency gains.
• $100 Million Annual Revenue: For a company with $100 million in annual revenue, the open-source LLM tool is a no-brainer. The initial and ongoing costs are negligible, and the personnel cost savings would provide a substantial return on investment.

Annual Revenue Levels for Commercial AI Tool Consideration:

• $10 Million Annual Revenue: With the potential personnel cost savings, the commercial AI tool may be worth considering even at this revenue level. The initial implementation costs ($150,000 — $300,000) and annual recurring costs ($86,250 — $180,000) could be partially offset by the savings.
• $20 Million Annual Revenue: At this revenue level, the commercial AI tool becomes a more viable option. The initial and ongoing costs would be a smaller percentage of the revenue, and the personnel cost savings would make it a worthwhile investment.
• $50 Million Annual Revenue: The commercial AI tool is a strong contender at this revenue level. The initial and ongoing costs would be a manageable portion of the revenue, and the company would benefit from the turnkey solution and support provided by the vendor.
• $100 Million Annual Revenue: For a company with $100 million in annual revenue, the commercial AI tool is an excellent choice. The initial and ongoing costs would be a small fraction of the revenue, and the potential personnel cost savings would provide a significant return on investment.

So the open-source LLM tool becomes a viable option for companies with as little as $10 million in annual revenue, thanks to the potential personnel cost savings. The commercial AI tool also becomes more feasible for companies with $20 million or more in annual revenue when factoring in the cost savings. Both options are highly attractive for companies with $50 million or more in annual revenue, with the open-source LLM tool being the more cost-effective choice, while the commercial AI tool offers a more turnkey and supported solution.

Having seen this;

Detailed indicators for decision making:
- Initial and recurring costs vs. staff savings: A critical indicator is the balance between initial and recurring costs of AI tools vs. savings from reduced staff costs.
- Margin improvement needs: The percentage increase in margin per unit needed to offset AI implementation costs provides insight into the viability of such investments.
- Revenue thresholds: Identifying the revenue thresholds that make the adoption of AI tools viable helps in strategic planning and ensures the sustainability of the investment.

Conclusion and Strategic Recommendations:

This analysis reveals the nuanced financial and operational considerations a manufacturing company must evaluate when deciding between open-source and commercial AI solutions for quality decision automation. The choice hinges on the company’s specific product mix, production volume, existing profit margins, and strategic financial thresholds. It’s recommended that the company conducts a thorough cost-benefit analysis, considering both the short-term impacts and long-term efficiencies of AI implementation. Balancing initial investment costs against the potential for significant operational improvements and personnel cost savings is key to making an informed decision tailored to the company’s unique circumstances and growth ambitions.

#AI#BA