Plan AI support around your organization
Bring your use case, knowledge sources and operational requirements. We will help you assess fit and define an initial rollout with clear scope and agreed terms.
Start with the core workflow
Define what the Agent may answer, what information it can use and when your team should take over.
Workspace permissions
Agents, knowledge and conversations are organized by Workspace. Review membership and role permissions with the people who will operate the service.
Knowledge and configuration
Connect approved sources, wait for processing and test representative questions. Review draft Agent instructions and workflows before publishing.
Human support
Define who monitors the inbox, when they are available and how unresolved conversations are handled. A handover request does not guarantee an immediate reply.
Discuss deployment requirements
The standard product is a hosted service. Additional infrastructure requirements need a feasibility review and written confirmation.
Hosted application
- Configure Agents in the hosted dashboard
- Install the website widget where your site permits scripts
- Review the privacy policy and product limits
- Test the published configuration before customer rollout
Infrastructure needs
- Describe any regional hosting requirements
- Ask about dedicated infrastructure or private networking
- Confirm whether your requirements can be supported
- Agree scope and delivery in writing before relying on an option
A bounded first rollout
- Choose one supported task and a small knowledge source
- Use representative questions and expected answers
- Test integration failures and human escalation
- Expand only after reviewing the pilot results
Review your data requirements
Describe your data and obligations before rollout. Confirm the required contractual and technical controls with your reviewers and the LoopReply team.
Data handling review
Read the published privacy policy and describe the data you intend to process. Ask the team to confirm contractual requirements before sending sensitive records.
Model and provider settings
Identify the AI provider and model used by your Agent. Confirm the relevant processing and retention terms rather than assuming every provider has identical policies.
Retention requirements
Explain your retention and deletion requirements during evaluation. Do not assume that an automatic per-Workspace purge schedule is available.
Conversation records
The inbox uses stored conversation history for support. If you require processing without stored conversations, confirm feasibility before using the service.
Location requirements
Confirm hosting and subprocessors for your proposed setup. A region-specific or fully local deployment requires an explicit agreement.
Compliance evidence
Request the evidence your organization needs. This page does not certify compliance or promise a particular certification, BAA or data-residency arrangement.
Prepare an access and security review
Use these review steps when preparing your configuration. They are deployment checks, not a security certification.
- Limit Workspace access to the people who need it and review their roles.
- Use dedicated, least-privilege credentials for connected services.
- Review which files, pages and database rows an Agent may access.
- Use read-only accounts for supported database knowledge connections.
- Test that private account data cannot be obtained through an unauthenticated chat.
- Test prompt-injection attempts and missing-source behavior with your own configuration.
- Agree escalation, deletion and incident contacts before rollout.
Plan capacity and readiness
Test the workload you expect and agree operating limits before expanding beyond the pilot.
Usage and plan limits
Estimate active users, messages, Agents and team members. Review current allowances and request a quote when your needs exceed the public plans.
Knowledge readiness
Source processing is asynchronous. Monitor status and test answers after a sync; an accepted upload does not mean indexing is complete.
Service expectations
Discuss traffic patterns, response expectations and required service commitments. Specific uptime, recovery and support guarantees apply only if agreed in writing.
Define each integration explicitly
Specify the data or action needed from each service, then verify authentication, supported actions and failure handling.
Business-system APIs
Check endpoint access, authentication, available workflow actions and error handling. A generic API connection does not establish a native integration with every ERP or CRM.
Database knowledge
Supported PostgreSQL and MySQL sources sync bounded, read-only query results into saved knowledge. Chat does not run unrestricted SQL against your systems for every answer.
Website and booking links
Embed the chat widget and direct visitors to existing account or booking services. Returning a link is different from performing an action in that service.
Agree the help your team needs
Discuss configuration, integration work and support expectations before selecting the final plan.
Configuration help
Describe the task, source material and workflow you want to build. The team can help you assess how to configure an initial Agent.
Integration review
Share the intended integration and its API documentation so feasibility, permissions and implementation effort can be discussed.
Commercial terms
Request a written quote covering the agreed scope, allowances and support terms. Dedicated staffing, custom development and SLAs are not assumed inclusions.
Discuss your first Agent workflow
Send us your intended workflow and requirements. Review the product facts and current plans while we assess the proposed setup.