EDITOR’S NOTE: One of the industry’s top enterprise architects from ConvergentIS took the time to talk with me and fill in any gaps from my research on the relationship between ERPs and GenAI. This will be the second in a series of posts that will be published today and throughout December. Here is the link to the first post.
Achieving seamless integration between horizontal and vertical GenAI models requires careful design. This involves leveraging each model’s strengths while ensuring they work collaboratively to address the full spectrum of business needs.
MY RESEARCH
Achieving seamless integration between horizontal and vertical GenAI models requires careful design, leveraging the strengths of each model while ensuring they work collaboratively to address the full spectrum of business needs. Here’s a step-by-step approach to achieve this integration:
SPECIAL NOTE: I have only shared the first two of the 10 steps for the following reasons:
STEP 1:
- When most organizations – okay, 80% of organizations take these first two steps, they erroneously use a tech-led equation-based development and implementation model. From that point on, success is highly unlikely. In short, tech is the last piece of the puzzle, not the first.
- In step one, it is crucial to examine agents (stakeholders) outside the procurement department. Watch this 13-minute video for a detailed overview of a successful implementation using an agent-based model.
- Only when you transition from a tech-led equation-based to an agent-based model are you ready to categorize tasks and map workflow dependencies.
STEP 2: At this point, you engage an Agent-based AI Operating System solution provider like ConvergentIS to structure the technology architecture.
1. Define Use Cases and Integration Points
- Categorize Tasks:
- Identify tasks best suited for horizontal models (e.g., general insights, summarization, FAQs) and those for vertical models (e.g., domain-specific predictions, regulatory compliance).
- Map Workflow Dependencies:
- Understand where horizontal and vertical models need to interact within a process.
- Example: In an ERP system, a horizontal model might handle general queries, while a vertical model provides specialized insights for procurement optimization.
2. Use an Orchestration Layer
- Middleware for Model Coordination:
- Implement an orchestration layer to route tasks to the appropriate model.
- Example: An AI pipeline routes general queries to GPT-based models and industry-specific questions to a fine-tuned vertical model.
- Dynamic Task Assignment:
- Based on the task complexity and domain specificity, the orchestration layer determines whether to use the horizontal or vertical model.
Benefits of Seamless Integration
- Efficiency: Combines general and specialized capabilities to maximize resource use.
- Accuracy: Balances broad understanding with domain-specific precision.
- Scalability: Supports diverse use cases without requiring a single, overly complex model.
- Flexibility: Allows independent evolution of both model types to adapt to changing needs.
By employing these strategies, businesses can leverage the combined power of horizontal and vertical GenAI models to optimize workflows and improve decision-making.
ConvergentIS
Using a GenAI-based orchestration engine (like RIO) to facilitate automated orchestration of the models based on the context of the prompt makes a lot of sense – especially if you can use that engine to compare models during development and ongoing testing so that the orchestration of the models can actually evolve over time too. This would only be possible by using something like RIO instead of programmatically defining the model based on the application functionality.
MY TAKEAWAY.
The following excerpt is from the post Are you chasing solutions or solving problems? Click this link to read the full post.
When I talk about agent-based problem-solving technology versus equation-based technology-led functionality, it isn’t a vague or high-level conceptual musing. It is an on-the-ground, in-the-trenches reality.
In many previous posts, I shared the following results of an agent-based approach to technology creation and implementation:
“In August 2003, the new technology successfully went live in a production environment for the DND. In this test case, the public sector organization realized a year-over-year 23% cost of goods savings for seven consecutive years while simultaneously reducing the number of buyers required to manage the contract to 3 from an original 23. Delivery performance and product quality also improved dramatically.”
(NOTE: The Agent-Based Development and Implementation model below occurs within the center hub, circled in red.)
30
Procurement Insights
How do you integrate horizontal and vertical GenAI models seamlessly?
Posted on November 30, 2024
0
EDITOR’S NOTE: One of the industry’s top enterprise architects from ConvergentIS took the time to talk with me and fill in any gaps from my research on the relationship between ERPs and GenAI. This will be the second in a series of posts that will be published today and throughout December. Here is the link to the first post.
Achieving seamless integration between horizontal and vertical GenAI models requires careful design. This involves leveraging each model’s strengths while ensuring they work collaboratively to address the full spectrum of business needs.
MY RESEARCH
Achieving seamless integration between horizontal and vertical GenAI models requires careful design, leveraging the strengths of each model while ensuring they work collaboratively to address the full spectrum of business needs. Here’s a step-by-step approach to achieve this integration:
SPECIAL NOTE: I have only shared the first two of the 10 steps for the following reasons:
STEP 1:
STEP 2: At this point, you engage an Agent-based AI Operating System solution provider like ConvergentIS to structure the technology architecture.
1. Define Use Cases and Integration Points
2. Use an Orchestration Layer
Benefits of Seamless Integration
By employing these strategies, businesses can leverage the combined power of horizontal and vertical GenAI models to optimize workflows and improve decision-making.
ConvergentIS
Using a GenAI-based orchestration engine (like RIO) to facilitate automated orchestration of the models based on the context of the prompt makes a lot of sense – especially if you can use that engine to compare models during development and ongoing testing so that the orchestration of the models can actually evolve over time too. This would only be possible by using something like RIO instead of programmatically defining the model based on the application functionality.
MY TAKEAWAY.
The following excerpt is from the post Are you chasing solutions or solving problems? Click this link to read the full post.
When I talk about agent-based problem-solving technology versus equation-based technology-led functionality, it isn’t a vague or high-level conceptual musing. It is an on-the-ground, in-the-trenches reality.
In many previous posts, I shared the following results of an agent-based approach to technology creation and implementation:
“In August 2003, the new technology successfully went live in a production environment for the DND. In this test case, the public sector organization realized a year-over-year 23% cost of goods savings for seven consecutive years while simultaneously reducing the number of buyers required to manage the contract to 3 from an original 23. Delivery performance and product quality also improved dramatically.”
(NOTE: The Agent-Based Development and Implementation model below occurs within the center hub, circled in red.)
30
Share this:
Related