Easing Small Businesses into AI | 16 Years of Expertise

AI Adoption in Supply Chain and Logistics

The artificial intelligence sector is currently facing scrutiny regarding potential overvaluation, yet the supply chain and logistics industries are witnessing practical and beneficial applications of the technology. Companies like Flexport and Uber Freight, alongside numerous startups, are actively developing and implementing AI solutions, attracting significant clientele.

However, the advantages of AI extend beyond large corporations. It is proving to be a valuable asset for smaller businesses seeking to optimize their operations and improve profitability.

Netstock's Opportunity Engine

Founded in 2009, Netstock, a provider of inventory management software, has recently launched a generative AI-powered tool called the “Opportunity Engine.” This tool integrates seamlessly into the existing customer dashboard, leveraging data from Enterprise Resource Planning (ERP) systems to deliver regular, real-time recommendations.

Netstock reports that its tool is generating substantial savings for its users. To date, the company has issued 1 million recommendations, with 75% of customers receiving suggestions valued at $50,000 or more.

Overcoming Initial Hesitancy

Bargreen Ellingson, a 65-year-old, family-owned restaurant supply company, initially expressed reservations about adopting an artificial intelligence product.

Jacob Moody, the chief innovation officer, explained to TechCrunch that established family businesses often approach change with caution. He stated that introducing a “black box” to manage warehouse operations would have been met with resistance.

To address these concerns, Moody presented Netstock’s AI as a supplementary tool that warehouse managers could choose to utilize or disregard, describing it as a cautious and deliberate introduction to AI capabilities.

Enhancing Efficiency and Empowering Staff

Moody noted that the AI tool is proving effective in preventing errors by analyzing the numerous reports used for inventory decision-making. While acknowledging that the AI-generated summaries are not always perfect, he emphasized their ability to “create signals from the noise,” particularly during off-peak hours.

A significant benefit observed by Moody is the increased effectiveness of Bargreen Ellingson’s less-experienced warehouse personnel.

He cited an employee with two years of experience and a high school diploma who, despite lacking extensive training in inventory management tools, can quickly assess the AI-driven insights and determine their validity. This empowers the employee and fosters a sense of ownership.

Data-Driven Success and Responsible AI

Barry Kukkuk, Netstock’s co-founder, recognizes the skepticism surrounding new technologies, particularly those that offer limited functionality beyond basic chatbot interactions.

He attributes the early success of the Opportunity Engine to the company’s extensive data collection over more than a decade of collaboration with retailers, distributors, and manufacturers. This data, protected under ISO frameworks, fuels the recommendation models, utilizing a combination of open-source and proprietary AI technologies.

The models are continuously refined through user feedback, both through explicit ratings (thumbs up/down) and implicit reinforcement based on whether suggested actions are implemented.

Kukkuk distinguishes Netstock’s approach from the attention-seeking strategies of social media platforms, prioritizing customer outcomes over mere engagement. He stated, “We care about: ‘what is the outcome for the customer?’”

Cautious Expansion and Maintaining Accuracy

Kukkuk is hesitant to expand the interactive capabilities of the AI, recognizing the potential for inaccuracies inherent in current generative AI technology. While conversational interactions could provide valuable context, they also risk introducing “hallucinations” and compromising the reliability of the recommendations.

This cautious approach is reflected in the Opportunity Engine’s integration within Netstock’s existing dashboard. Suggestions are prominently displayed but easily dismissible, avoiding the overwhelming feature saturation seen in some applications, like Google Docs.

Moody appreciates the non-intrusive nature of the AI, emphasizing that no inventory decisions are made without human review and approval. He envisions a future where AI suggestions are accepted with a 90% agreement rate before granting the system greater autonomy.

Looking Ahead: Potential and Concerns

The initial success of Netstock’s Opportunity Engine represents a promising development in the broader adoption of generative AI within enterprise settings.

However, Moody also expressed concerns about the potential long-term implications of AI, anticipating significant changes within the industry. He acknowledged the possibility of reduced demand for data science experts, but stressed the importance of retaining employees who possess a deep understanding of the underlying principles and can critically evaluate the AI’s recommendations.

Bargreen Ellingson needs personnel who can “deeply understand the theory and the philosophy” and ensure that the company doesn’t “blindly going down” an incorrect path.