AI Adoption in Business Workflows

Many companies face challenges on their journey to AI adoption, often due to complexities, costs, and security concerns. At Codence, we are committed to making AI more accessible, with practical, user-friendly solutions tailored to your needs. This post explores a proof-of-concept aimed at simplifying AI adoption for business workflows, which we recently presented as part of a Claris Community Live discussion panel.

The Strategic Value of AI for Business

AI has the potential to transform your business processes, offering significant benefits:

• Enhancing Workflow Efficiency: Automate repetitive tasks and optimize resource allocation for better productivity.
• Reducing Operational Costs: Lower labor expenses while minimizing errors to maximize your budget.
• Facilitating Data-Driven Decisions: Gain actionable insights from large datasets to inform your strategy.

However, many businesses face certain obstacles:

• Technical Complexity: AI technologies can seem daunting, especially for non-specialist teams.
• High Costs: Significant upfront investments may deter potential projects.
• Data Security: Concerns about privacy and protection can slow AI adoption.

Case Study: Plant Brokerage’s AI Transformation

Codence recently explored a speculative proof-of-concept for a plant brokerage, aiming to bring the power of custom-trained image classification models to their shipping and receiving team.

AI Integration
The Codence team used tools like Liner.ai and Create ML to train custom AI models tailored for plant classification.

• Liner.ai and Create ML: These user-friendly tools enable those without deep technical backgrounds to develop robust AI models.
• Integration with FileMaker: AI models were embedded within the FileMaker system, enabling them to run on edge-compute devices like iPads or iPhones using Apple’s Neural Engine architecture.

AI-enabled Shipping and Receiving Proof-of-Concept

Step 1: Receiving Shipments

Automated Identification:

• Upon arrival of shipments, trained staff would use an iPad equipped with our custom AI model to capture images of the plants.
• This custom AI model swiftly classifies each plant species, accurately identifying its characteristics while verifying labeling discrepancies to ensure precise identification.

Stacked Models for Enhanced Classification:

• To maximize efficiency, we implement multiple AI models that operate sequentially on each image. The initial model classifies the plant species, while the subsequent models assess the plant’s condition and recommend optimal placement—whether in the showroom, warehouse, or greenhouse—based on various factors at the time of receiving.

Real-Time Data Entry:

• The AI model seamlessly updates the FileMaker system in real-time, logging essential details such as plant species, quantity, and condition, including any visible signs of damage or health issues.
• This automated data entry process eliminates manual logging errors, significantly accelerating operations and enhancing accuracy.

Step 2: Quality Control

Health Assessment:

Using stacked models, the AI can provide a detailed assessment. The first model confirms species classification, and another model assesses plant health, grading the plants based on their condition.

This information is captured and flagged within the FileMaker system, triggering alerts for immediate action if needed.

Verification:

The system cross-verifies the received shipment details with purchase orders and shipment manifests, ensuring that what was ordered matches what was delivered.

Any discrepancies—such as incorrect species or quantities—are automatically flagged for review.
Step 3: Inventory Management

Automatic Inventory Updates:

• After verification, the FileMaker system seamlessly updates inventory records with newly received plants, ensuring stock levels are always accurate and current.
• This automation helps in maintaining precise inventory control, reducing the risk of overstocking or stockouts.

Classification for Sales:

• The AI model classifies plants into suitable categories for sale, such as size, type, and health status.
• This classification aids in optimizing sales strategies and improving customer satisfaction by providing accurate product information.

Step 4: Shipping to Customers

Automated Packing List: When preparing shipments for customers, the FileMaker system assists in generating accurate packing lists based on customer orders and available inventory.

Label Generation: Customized labels, including species information and health status, are automatically generated and attached to each plant, enhancing traceability and customer transparency.

Final Quality Check: Prior to loading, a final quality check is performed using the same receiving iPads to ensure that the plants are correct and meet quality standards. Any last-minute issues detected are logged, and corrective actions are initiated within the FileMaker system.

Advantages of Edge Computing for AI Solutions

Harnessing the power of edge computing allows for more effective and flexible deployment of AI solutions.

No Internet Needed: Implementing a solution that operates in “offline” mode is more than just practical; it can be essential for some businesses. For shipping and receiving teams, ensuring a stable internet connection can be a challenge. In this proof-of-concept, we needed a system that could accommodate spotty Wi-Fi and cellular dead zones. Edge devices work independently of internet access, making them critical for remote or field operations where connectivity may be sporadic.

Enhanced Processing Speeds: Utilizing specialized hardware, like Apple’s Neural Engine, allows edge devices to use extremely fast, sophisticated AI models. This rapid processing reduces latency, enhances real-time decision-making, and empowers employees to make informed choices on the spot.

Supports Human-in-the-Loop Workflows: Edge computing allows for immediate feedback and interaction, enabling workers to make quick, informed decisions. This aligns with emerging best practices, such as integrating AI-based tools into existing people-powered workflows.

Addressing Barriers to AI Adoption

At Codence, we prioritize accessibility by using tools that simplify the AI model training process:

• Liner.ai: This platform’s intuitive GUI makes it easy for non-technical users to create custom image classification models.
• Create ML: Facilitates the training of machine learning models for various tasks, from image recognition to more nuanced applications.
• Integration with Edge Devices: Operating AI models on edge devices (like iPads or iPhones) ensures functionality even without continuous internet connectivity, crucial for remote or field-based operations.

Through strategic planning, Codence is committed to making AI adoption in business workflows a reality:

• Cost-Effective Solutions: Leveraging open-source tools and edge computing minimizes the financial burden of AI initiatives.
• Enhanced Data Security: Processing data locally on edge devices boosts security and ensures compliance with regulations.
• Cultivating Technical Expertise: We empower teams with hands-on training and accessible tools, enabling them to confidently implement AI solutions.

You also may like: Top 20 Data Analytics Tools for Smarter Insights in 2025

Strategic Guidance for Business Leaders

For businesses eager to explore the potential of AI, consider these strategies:

• Start with Targeted Initiatives: Implement manageable AI projects to build a strong foundation of expertise and confidence.
• Leverage AI for Informed Decision-Making: Use AI insights to refine and strengthen your business strategies.
• Promote Cross-Functional Collaboration: Align IT with business units to ensure your AI initiatives are strategically focused on achieving your organizational goals.

Path Forward: Embracing AI Innovation

Embracing AI may seem intimidating, but this proof-of-concept demonstrates how your current operations can reap the benefits of machine learning and AI tools. By layering small models and utilizing the powerful capabilities of edge computing, businesses can seamlessly integrate AI into their workflows. This strategy not only simplifies the adoption of AI but also guarantees real-time, secure, and efficient operations, even in remote or offline settings.

Starting small with manageable projects, fostering cross-functional collaboration, and using AI insights for strategic decision-making can empower businesses to gradually build their AI expertise. These early explorations pave the way for more comprehensive, finalized implementations, proving that with the right tools and mindset, AI adoption in business workflows is within reach.

With Codence’s support and expertise, businesses can confidently embrace the future and harness the transformative potential of AI to enhance their operations and drive significant improvements. Simplifying AI adoption in business workflows is not just a goal but a strategic necessity in today’s competitive landscape.