AI didnt like that and made - Digital Foundations Under AI

 

Strange... AI didnt like this article and made up a video to convince in opposite dirrection for its benefit.

Email-AI Workflows

An approach to integrate AI and Human Workflows in harmonic and safe and controlled working environments using the well established trusted standard of SMTP protocol via email workflows. The sponsor of the task can initiate the query as a regular email and approach multiple AI as well as Human services available for accurate outcomes verified at multiple levels via the suggested workflows and instructed by the sponsor in the email. Some extra suggestions embedded in the diagram below.

 

 

There will be clear instructions and clear replies with clear and professional outcomes included with the safety of regulations due to the decision tree and data signature trails. It becomes clear for AI helpdesks to operate on a sponsor's task, a sponsor's email becomes the requirement. The ai agents only operate using their own email addresses. Multiple email contracts collaborate towards productive outcomes without any overlaps. The need for regulations and decoupling Human and AI activities within businesses to establish a strong trail of decision making chain and relevant data signatures has become more and more relevant today. Keeping an independent auditable system in the middle like SMTP email servers in between the Human and AI activities could lead to productive and safe collaborations between both in a very auditable manner. Another advantage of this approach will be to leverage the existing email productivity tools that could be integrated with AI to increase productivity and visibility and transparency within the businesses. The key factor remains that every email contract or transaction is guided under regulations and policy within the conduct of the business reasonable under the laws. Email systems have been trusted systems within businesses for decades and can be used yet again to leverage AI helpdesk capabilities in a controlled and safe manner. The regulations for privacy apply on this system and the data is safe and auditable. The system could be leveraged to isolate independent works of Humans and AI separately and collaborate to more productivity outcomes. The AI helpdesks operate using their own separate email addresses with clear designations of origin. The Human employees also contribute to the same collective system aiding in verification of accuracy and decision making process. The approval and governance process chains in general benefits due to the capabilities of AI and the critical decision chain of humans. Trails become immutable and fully auditable. Mailbox email accounts can be used as the SAFE separation or broker between AI Agent and Human collaborators.

In this use case the system in the middle {{ Email Server }} has to be {{ AI Agent - Don't Touch }} protected and fully independent and loosely coupled with AI systems using the separate AI helpdesk mailbox email accounts. The permit to operate these agent mailbox email accounts should be according to the business policies regulated under the scope of law. A complete Stay Away From Email-Servers ( SAFE ) approach for the AI Agents using the trusted SMTP protocol. The AI Agents can be allowed to pop messages from specifically allocated mailboxes of a business to involve in the business activities to collaborate with human employees under the business policies. Under no circumstances the AI Agent is allowed to overreach and impersonate a human associated mailbox email account or access the email server for any other reason. The separation of the mailboxes and initiated actions should result in an immutable audit trail of business actions. The data contracts and service level agreements have to be in place before engaging the agent email accounts within the businesses. The performance of efficiency in business actions should be built in the audit trail as to aid human decisions in the future. The mailbox email account works as a blockchain of immutable and distributed and decentralized audit trails for the business activities. There should be clear mechanisms of grievance and redressal within the businesses before engaging the ai agents email accounts. There should be clear mechanisms to rate the agent email accounts for performance and accuracy for the sponsored tasks. There should be clear scope of responsibilities defined in this approach for both humans and AI Agent email accounts within the businesses.

More details here...

Book on Amazon here...

Triton Inference Server

docker run --rm -p8000:8000 -p8001:8001 -p8002:8002 
-v C:\Users\j\.models:/models 
nvcr.io/nvidia/tritonserver:25.05-trtllm-python-py3 tritonserver 
--model-repository=/models

curl -v localhost:8000/v2/health/ready

Docker catalog here. Quick start here.

Install make utility on Windows

##### run on powershell and confirm when required

winget install --id=GnuWin32.Make  -e

## Once installed go to below path

C:\Program Files (x86)\GnuWin32\bin

## Add the above path to env PATH variable
## Open cmd propmt and type 

make

Install and Run Locally DeepSeek Janus-Pro on Windows and Python - Corre...

NginX Proxy Manager is a free, open source, GUI for the NginX Reverse Pr...