Learn how to control your Smith settings, manage your chat history, and understand the enterprise-grade security that protects your data. We'll cover conversation settings, chat history management, data security, compliance, and best practices for keeping your account secure.
What Settings Can You Control?
Smith provides straightforward account management and enterprise-grade security features designed to keep your work organized and your data protected. This article covers the essential settings you can control, how to manage your chat history, and the security measures that protect your information.
Conversation Settings
Control Memory, Creativity, and Multi-Model options
Chat History
Search, resume, and manage your conversation history
Security & Privacy
Enterprise-grade protection for your data
Accessing Settings
Smith's settings are easily accessible from the main interface. The Settings gear icon appears in the bottom-right corner of the prompt input area, next to the model and persona selectors.
Quick Access Tip: Settings are context-aware and apply immediately to your current chat session. Changes you make will affect how Smith responds in ongoing and future conversations.
Conversation Settings
Smith offers three key settings that control how the AI interacts with you. These settings help you customize the conversation experience based on your needs.
Memory
The Memory toggle determines whether Smith retains context from earlier messages within the same chat session. When enabled, Smith can reference previous exchanges and maintain conversational continuity across multiple questions.
Smith remembers context from earlier in the conversation and can reference past exchanges. Best for multi-turn discussions, iterative analysis, building on previous responses, and complex problem-solving.
Each prompt is treated independently without reference to previous messages. Best for single questions, unrelated queries, fresh perspectives, and avoiding context contamination.
Creativity (Temperature Control)
The Creativity setting controls the AI model's temperature, which affects how deterministic or creative the responses are. You can choose from three levels:
Most deterministic and focused. Responses are consistent and predictable. Ideal for factual analysis, data extraction, technical documentation, and situations requiring precision.
Balanced approach with moderate variation. Good for general business writing, summarization, email drafting, and everyday tasks requiring both accuracy and natural language.
Most creative and varied. Produces diverse, imaginative responses. Best for brainstorming, creative writing, exploring multiple perspectives, and generating unique ideas.
Multi-Model
The Multi-Model toggle enables side-by-side comparison of responses from different AI models. When turned on, Smith will display responses from multiple selected models simultaneously, allowing you to compare perspectives and choose the best answer for your needs.
This feature is particularly valuable when you want to validate information across different AI systems or explore how different models approach the same prompt.
Recommended Settings by Task Type:
Managing Chat History
Smith automatically saves your chat sessions so you can return to previous conversations, review past analyses, or continue where you left off. The Chat history button appears in the top-right corner of the interface.
Open Chat History
Click the "Chat history" button in the top-right corner to open the history drawer.
Search or Browse
Use the search bar to filter by name or browse through your sorted history (most recent first).
Resume or Delete
Click any chat to resume the conversation, or use the delete option to remove individual chats or clear all history.
What You'll See in Chat History
Each saved chat displays detailed metadata to help you identify and locate specific conversations:
Deleting Chats
Smith provides two options for removing chat history:
Deletion is Permanent:
Both individual chat deletion and clearing all history are permanent actions. Smith will display a validation message before proceeding to ensure you want to delete the selected content. Once confirmed, deleted chats cannot be recovered.
Data Security & Privacy
Smith is built with enterprise-grade security at its core. Your data protection is not an afterthought—it's a fundamental design principle that guides every aspect of how Smith handles your information.
Persistent Storage
Files remain in your workspace until you delete them
Instance Isolation
Your data is logically separated from other organizations
User-Level Access
Only authorized users can access workspace content
Data Ownership & Control
You retain complete ownership of all content, documents, and data you upload to Smith. EBM Software acts solely as a processor to help you interact with your data through AI capabilities. Your intellectual property rights remain entirely with you.
Data Storage & Isolation
Your uploaded files are securely stored in your dedicated workspace environment managed through EBM's Ontario platform. Each workspace is isolated within your specific instance, ensuring that your document collection remains private and accessible only to authorized users within your organization.
Smith's security architecture is designed to eliminate the primary risk hindering enterprise AI adoption: data security. Unlike consumer-grade AI tools that may use your data for model training, Smith routes all queries through a single, governed gateway to top-tier enterprise models via end-to-end encrypted APIs. This architecture formally and technically ensures your proprietary prompts and corporate data are never used for model training, safeguarding your intellectual property.
Enterprise-Grade Infrastructure
Smith is built on Microsoft Azure's enterprise infrastructure, providing multiple layers of security and compliance:
Azure Cloud Platform
Hosted on Microsoft Azure (East US and East US 2 regions), keeping all data within the United States with enterprise-grade reliability and availability
Zero Training Policy
Your data is never used to train underlying AI models from OpenAI, Anthropic, or Google
Private Instances
For highest security requirements, physically segregated instances provide complete data isolation
Certified Compliance
ISO 27001:2022 and SOC 2 Type II certified for information security management
Encryption Standards
All data is protected using industry-standard encryption protocols:
Encryption in Transit
All data transmitted between your browser and Smith's servers uses TLS/SSL protocols to prevent interception
Encryption at Rest
Files stored in the workspace database are encrypted using AES-256 encryption, ensuring they remain secure in storage
Data Usage & Training
Smith uses your data strictly for retrieval-augmented generation (RAG) to provide context for answering your specific queries. Your data is never used to train the underlying AI models (GPT, Claude, Gemini, DeepSeek) or shared with model providers to improve their base systems.
Your Data Stays Yours: The AI models process your files temporarily to generate responses, but they do not retain, learn from, or incorporate your proprietary information into their training data. Your business intelligence remains confidential and contained within your instance.
Access Controls
EBM employees do not have access to your private workspace data or document contents. Access is strictly controlled through role-based permissions:
👤 User Access
Full access to their own workspace files and folders based on organizational permissions
🔧 System Maintenance
EBM technical staff can access system metadata for maintenance but cannot view document contents
🛟 Support Access
Support personnel can only access your data if you explicitly grant permission for troubleshooting purposes
Advanced Security Features
Smith incorporates sophisticated security measures designed specifically for enterprise AI deployment:
Intelligent Query Routing
Sophisticated semantic models understand the precise intent and context of your queries, routing them to the most appropriate AI model while maintaining security boundaries.
Knowledge Graph Grounding
Queries are grounded in your organization's internal knowledge graphs and verified data sources, dramatically reducing the risk of AI "hallucinations" and ensuring responses are based on factual information.
Multi-Model Validation
Responses from multiple leading AI models (OpenAI, Anthropic, Google) are compared against verified facts from your knowledge graph, providing cross-validation and reducing model-specific biases.
End-to-End Encryption
All communications with enterprise AI models use encrypted APIs, ensuring your queries and responses remain confidential throughout the entire processing pipeline.
Session & Usage Monitoring
Smith includes session management capabilities that allow administrators and users to track interaction history. You can review the specific queries and outputs generated during sessions, providing transparency and oversight of how the tool is utilized within your organization.
Why Smith is More Secure Than Consumer AI Tools
The distinction between consumer-grade AI tools and enterprise platforms like Smith is fundamental. Consumer services are designed for the general public and often treat user data as a resource for model improvement. Smith, by contrast, is architected around the principle of data protection:
| Security Feature | Consumer AI Tools | Smith Enterprise |
|---|---|---|
| Data Used for Training | Yes, by default (opt-out required) | No, never (contractual guarantee) |
| Data Encryption | Basic transit encryption | AES-256 at rest, TLS/SSL in transit |
| Access Control | Basic user account | Role-based permissions, enterprise SSO |
| Network Isolation | Public internet only | Private Azure infrastructure |
| Compliance Certifications | Generally not applicable | ISO 27001:2022, SOC 2 Type II |
| Data Retention | Provider-controlled (often years) | Customer-controlled with audit trails |
| Cross-Customer Isolation | Shared infrastructure | Isolated instances, optional physical segregation |
| Data Residency | Variable / unspecified | US-only (Azure East US regions) |
The Shadow AI Risk: When employees use consumer AI tools like public ChatGPT for work tasks, they inadvertently expose sensitive company data to systems designed to learn from that data. Smith eliminates this "Shadow AI" threat by providing a governed, enterprise-grade alternative that routes queries securely to leading AI models without sacrificing your data privacy.
Workspace Permissions
While individual user settings are personal, workspace access operates at the organizational level. Understanding how workspace permissions work helps ensure the right people have access to the right information.
"My Workspace" is unique to each user and is not shared between team members. Each person has their own isolated workspace for private files and folders. Shared organizational access—allowing multiple team members to query a common knowledge base—is on the product roadmap and will be introduced in a future release.
Shared Workspace Access — Coming Soon
Team-level shared folder access is on the roadmap. When available, it will be governed by the same instance isolation, role-based permissions, and AES-256 encryption that protect individual workspaces today—so your organization's collective knowledge base will be just as secure as your personal files. Contact your EBM account representative for availability timelines.
Workspace Collaboration Today: While workspaces are currently personal, users can share access to the same underlying data source through the Ontario platform. This allows multiple team members to query and generate insights from a collective knowledge base while maintaining individual workspace organization. For more information on organizing your workspace, see the Organizing Your Work with Workspaces article.
Compliance & Governance
Smith is designed with enterprise security and data sovereignty in mind, addressing the needs of IT, Legal, and Compliance teams. Our platform provides the contractual guarantees, technical controls, and compliance assurances necessary to protect corporate assets in regulated industries.
Current Certifications
Smith maintains industry-leading security certifications that demonstrate our commitment to information security:
ISO 27001:2022
Information Security Management System certification demonstrating a systematic, audited approach to managing sensitive company information. Recertified December 2025, valid through January 2029.
SOC 2 Type II
Audited controls for Security, Availability, and Confidentiality of customer data over time
Compliance Framework
Smith's architecture supports secure enterprise environments where access control and data sovereignty are priorities:
Data Isolation
Your data remains in your Ontario-connected workspace, separate from other organizations
Source Grounding
Responses can be restricted to verified internal documents through workspace file selection
Usage Tracking
Session history and usage monitoring provide visibility into how the tool is utilized
Audit Capability
Chat history and file metadata support audit requirements for regulated industries
Decision Support, Not Decision Making
Smith is designed as an assistive tool to augment human intelligence, not replace it. The platform provides information retrieval, synthesis, and drafting assistance based on the data you provide. Final decision-making authority and liability rest with you as the user.
Professional Review Required:
All critical outputs, especially those regarding legal, financial, or compliance matters, should be reviewed by a qualified professional before implementation. Smith provides intelligent assistance, but human expertise remains essential for high-stakes decisions.
Security Best Practices
While Smith provides robust security infrastructure, following these best practices helps ensure your data remains protected:
Tips for Keeping Your Account Secure
|
🔑Credential Management
|
Never share your login credentials. Each user should have their own account for accountability and audit purposes. |
|
🔐Sensitive Data
|
Verify you have proper authorization before uploading confidential or regulated information to your workspace. |
|
📁Folder Organization
|
Use private folders for sensitive personal files and shared folders only for content appropriate for team access. |
|
🧹Regular Cleanup
|
Periodically review and delete outdated files and chat history to minimize your data footprint. |
|
💾Export Important Work
|
Download critical analysis or responses to your local system for backup and version control. |
|
✓Verify Permissions
|
Before uploading to shared folders, confirm the appropriate team members should have access to the content. |
Common Questions
Can I recover deleted chat history?
No, chat deletions are permanent and cannot be recovered. Always verify before confirming deletion actions.
How long are my files stored in the workspace?
Files remain in your workspace indefinitely until you manually delete them. Smith does not automatically remove uploaded content.
Can I see who has accessed my shared folders?
File metadata includes upload and modification information. For detailed access logs and permissions management, contact your Smith administrator.
What happens if I reach the file size limit?
Files larger than 25MB cannot be uploaded. Consider splitting large files into smaller sections or compressing them before upload.
How do I request support access to my workspace?
Contact EBM support through the standard support channels. Support personnel will request explicit permission before accessing your data for troubleshooting.
Can I export my entire chat history?
Individual chat conversations can be exported to Word or PDF format. To export multiple chats, you'll need to download them one at a time. See Starting and Managing Conversations for export instructions.
Does Smith support Single Sign-On (SSO)?
Yes. Smith supports SSO through Microsoft Entra B2B guest accounts, allowing your organization to authenticate users through your existing identity provider without separate credentials. This simplifies access management and keeps authentication within your existing IT governance framework. Contact your EBM account representative to configure SSO for your organization.
Does my data stay in the United States?
Yes. All data is stored and processed exclusively within Microsoft Azure's East US and East US 2 regions. No data is transferred outside the United States as part of normal platform operations, making Smith well-suited for organizations with US data residency requirements.
Does EBM conduct regular penetration testing?
Yes. EBM conducts penetration testing on a regular cadence to proactively identify and remediate vulnerabilities. The most recent test was completed in August 2025. Results are used to continuously harden the platform's security posture.
Does Smith have a disaster recovery plan?
Yes. Disaster recovery planning is a formal component of EBM's ISO 27001:2022-certified Information Security Management System. Our policies cover business continuity, data backup, and recovery procedures to minimize service disruption and protect data integrity in the event of an incident. Specific RTO/RPO details are available to enterprise customers upon request.
Related Articles
Related Articles
Organizing Your Work with Workspaces
Learn about workspace structure, file organization, and folder permissions to keep your knowledge base secure and organized.
Starting and Managing Conversations
Explore conversation features, export options, and how to leverage chat history for productive AI interactions.
Welcome to Smith
Get started with Smith's core features and understand how to make the most of the platform.
Next Steps: Configure your conversation settings to match your workflow preferences, organize your chat history by deleting outdated conversations, and review the security features to understand how Smith protects your data. A well-managed account enables more productive and secure AI interactions.
Comments
0 comments
Article is closed for comments.