Frequently Asked Questions

REDA AI is a 100% Salesforce-native AI agent platform delivered as a managed package from the AppExchange. It lets you build and deploy intelligent AI agents directly inside your Salesforce org — with no code, no middleware, and no external infrastructure.

Agents built on REDA AI understand your custom schema, inherit your complete Salesforce security model automatically, and can be deployed across four channels — Lightning Pages, Experience Cloud, an External Widget, and Salesforce Flows — in hours rather than months.

Yes. REDA AI is delivered as a fully managed package via the Salesforce AppExchange. Installation takes under five minutes — one click from the AppExchange listing, no external infrastructure to provision, no middleware to configure.

The Salesforce AppExchange Security Review is currently underway. REDA AI is ISO 27001:2022 certified and SOC 2 Type II certified, and these certifications were a prerequisite for the Security Review submission.

REDA AI is designed for three personas who each get real value from day one:

• Salesforce Admins — build, configure, and deploy agents using point-and-click tools. No AI expertise or prompt engineering background required.
• CTOs and Technical Leaders — evaluate REDA AI against enterprise AI architecture requirements: data residency, security model, extensibility, TCO, and scalability.
• Business Process Owners — deploy AI agents that understand your specific Salesforce implementation and deliver measurable process outcomes.

Developers can also extend agents further via the Apex API and Tool Builder, but they are not required for core deployment.

No. Core agent setup — persona definition, prompt configuration, tool assignment, and channel deployment — requires zero code and no AI background.

The AI-assisted agent builder reads your org schema and automatically generates the optimised persona, prompt configuration, context fields, and suggestion chips. Admins with Salesforce experience can build a production-ready agent in under 10 minutes.

For custom tool creation, Tool Builder guides you through a structured form and generates the underlying Apex code automatically. Developers can review and extend the generated code if needed, but it is not a requirement.

Under one hour for most teams. The four steps break down as:

• Install the package — under 5 minutes from AppExchange
• Activate your licence — 30 seconds in the REDA AI Admin tab
• Build your first agent with AI — under 10 minutes; REDA AI reads your schema and generates the configuration automatically
• Deploy to a channel — minutes per channel

For comparison: external AI integrations typically take 3–6 months; custom development takes 4–8 months. REDA AI compresses that to the same afternoon.

Yes. The Trial plan is free, requires no credit card, and activates instantly. It includes:

• 500 AI credits (enough for meaningful testing)
• One agent
• One month of access
• Internal Salesforce Lightning channel

Trial credits do not roll over, and External Widget, Experience Cloud, and Flow deployment are not available on the Trial plan. At the end of the trial, your agent configuration is preserved when you upgrade to any paid plan.

No. REDA AI is entirely self-contained as a managed package. It does not require Data Cloud, MuleSoft, Heroku, Einstein, or any other Salesforce add-on beyond the base platform itself.

This is a deliberate architectural decision. Additional dependencies mean additional licensing costs, additional complexity, and additional points of failure. REDA AI eliminates all of those by running natively inside your existing Salesforce org with zero external dependencies.

REDA AI reads your org's metadata — custom objects, fields, relationships, picklist values, and validation rules — through the Salesforce Metadata API at agent build time.

When you use the AI-assisted agent builder, it analyses this schema and auto-generates Schema Guidance prompts that accurately describe your custom objects and fields to the AI model. The result: an agent that understands your specific implementation, not a generic Salesforce template. Context Fields then pull live record data automatically when a user opens a record page, so the agent always starts informed.

REDA AI is compatible with Salesforce Enterprise, Unlimited, and Developer editions. It requires a Salesforce org with managed package installation rights.

If you have a non-standard configuration — sandbox environments, Shield-enabled orgs, or government cloud instances — contact the REDA AI team to confirm compatibility before proceeding.

Yes, with two independent layers of control. First, REDA AI enforces your existing Salesforce Field-Level Security, object permissions, sharing rules, and profile restrictions on every interaction — the agent can only see what the logged-in user can see. No special configuration required.

Second, you define Context Fields per agent to specify precisely which fields are surfaced as agent context. Sensitive fields — SSNs, salary data, health records, financial information — can be excluded entirely from any agent's context, even if the user technically has FLS access to them. Both layers apply simultaneously.

Yes to both. The REDA AI Apex API allows developers to invoke any agent programmatically from any Apex context — triggers, batch jobs, Queueable classes, or custom controllers. This enables deep integration with existing org logic without requiring a UI.

REDA AI also integrates natively with Salesforce Flows. You can invoke agents from record-triggered Flows, scheduled Flows, or platform event-triggered Flows, enabling fully unattended AI interactions as part of your existing workflow automations.

REDA AI is positioned as the extensible alternative — designed for the majority of Salesforce orgs that need production AI without the overhead and dependencies that come with Agentforce.

• No Data Cloud required — REDA AI works with your existing org data without a separate data layer
• Works on all supported editions — not limited to orgs with additional Salesforce products
• 3-level extensibility model — granular control over persona, knowledge, and tools through 5 structured prompt types
• Org-based pricing — no per-seat or per-user licensing that scales with headcount
• Hours to deploy — not months of professional services engagement

REDA AI is complementary to the Salesforce ecosystem, not competing with the platform itself.

Never. 100% of data stays inside your Salesforce org at all times. This is not a policy — it is a structural property of the architecture.

REDA AI stores everything as standard Salesforce objects inside your org: conversation transcripts, audit logs, agent configurations, incident records, credit usage data. There are no external databases, no cloud storage, no third-party data warehouses, and no message brokers. Data egress is architecturally impossible.

REDA AI inherits and enforces your entire Salesforce security model automatically, with zero configuration required. Field-Level Security, object permissions, sharing rules, profile restrictions, and record visibility rules all apply to every AI interaction.

The agent sees exactly what the logged-in user sees — no more, no less. No separate permission model is created, no privilege escalation is possible, and no security gaps are introduced. The security model is the same one you already manage in Salesforce; REDA AI simply operates within it.

REDA AI currently holds the following certifications:

• ISO 27001:2022 — Information security management system
• ISO 9001:2015 — Quality management system
• SOC 2 Type II — Security, availability, and confidentiality trust service criteria
• GDPR compliant — All data remains inside your Salesforce org; right-of-access and right-to-erasure requests are handled through standard Salesforce data management tools

Salesforce AppExchange Security Review is currently underway. Contact the REDA AI team for audit documentation or specific regulatory enquiries.

Every interaction is fully audited and stored as standard Salesforce objects. The Conversation Audit Log captures per interaction:

• User identity and timestamp
• Record context and channel
• Agent and tools used
• Credit usage (per message, per session, per user)
• Full conversation transcript

The Incident Log tracks errors by severity (Critical, High, Medium, Low) with root cause data, user attribution, and resolution status. All audit data is available through standard Salesforce reports and dashboards — no special tools required, and nothing leaves your org.

REDA AI provides multiple independent guardrail layers — each configurable per agent:

• Guardrail Prompts — restrict agent behavior in plain English at the knowledge level
• Action Confirmation Cards — require explicit human approval before any write operation (record update, email send, task creation)
• Tool-level controls — enable only the specific tools each agent needs; all others are disabled by default
• Per-user credit limits — cap usage per user per month to prevent runaway consumption
• Per-agent credit caps — set maximum spend per agent across all users
• Channel-specific restrictions — External Widget deployments have daily credit caps and origin whitelisting

REDA AI's architecture supports HIPAA-compliant deployments. Key properties that satisfy HIPAA requirements:

• All data — including any PHI — remains within your Salesforce org at all times with zero egress
• Full audit trail of every AI interaction satisfies the accountability requirements under HIPAA
• Field-level access controls allow explicit exclusion of sensitive health fields from agent context

For enterprise deployments requiring an explicit HIPAA Business Associate Agreement (BAA), contact the REDA AI team to discuss your specific compliance requirements and documentation needs.

Credits are the unit of AI work in REDA AI. Each plan includes a monthly credit allocation that refreshes at the start of each billing period. Credits are consumed per conversation turn — the exact amount depends on interaction complexity.

Your Admin dashboard shows real-time credit usage broken down by user, agent, and individual interaction. You can set per-user monthly credit limits and per-agent caps to manage consumption proactively, with efficiency ratings (Optimal / Good / Needs Review / High Cost) to identify optimisation opportunities.

Credit consumption scales with interaction complexity. Typical ranges per conversation turn:

• Simple interaction (FAQ, field lookup, basic query) — 5–10 credits
• Standard interaction (record intelligence, multi-message) — 15–25 credits
• Workflow interaction (multi-step reasoning, Flow invocation) — 25–40 credits
• Complex interaction (deep reasoning, multi-object analysis) — up to ~50 credits

Most real-world Salesforce use cases — record summaries, guided actions, Q&A on record data — fall comfortably in the 15–25 credit range per turn.

Yes, on all paid plans. Unused credits roll over automatically up to a maximum of 2× your monthly allocation. For example, if your plan includes 60,000 credits per month and you use 40,000, the remaining 20,000 carry forward — giving you up to 80,000 available the following month.

Trial plan credits do not roll over. Top-up credit packs purchased separately also do not roll over at period end.

You can purchase top-up credit packs at any time, available immediately upon purchase. Top-up pricing is $0.030 per credit with a minimum purchase of 25,000 credits ($750). Top-up credits are used before your monthly allocation and are available across all agents and channels.

To avoid running out unexpectedly, use per-user monthly credit limits and per-agent caps in your Admin dashboard. You'll also see real-time credit usage dashboards so you can monitor consumption trends before they become a problem.

Per org, per month — billed annually. There are no per-seat fees, no per-user licensing, and no headcount-based charges. Your entire user base can interact with REDA AI agents without changing your plan cost.

This is intentional. AI value should scale with usage and outcomes — not with the number of people you allow to benefit from it. Plans start at $1,500/month (Growth) and scale by credit allocation and agent count, not by user count.

The Trial plan is free with no credit card required. It includes 500 credits, one agent, one month of access, and deployment on the internal Salesforce Lightning channel only. Build Agent with AI, all 5 prompt types, and the full admin console are all available during the trial.

External Widget, Experience Cloud, and Flow deployment are not available on the Trial. Trial credits do not roll over. At the end of the trial period your agent configuration is fully preserved — upgrading to a paid plan continues from where you left off.

REDA AI's extensibility model has three independent levels — each adds capability without requiring the next:

• Level 1 — Agent Persona & Context: Define who the agent is, its tone, role, and purpose in plain English (up to 5,000 characters). Context Fields pull live Salesforce record data automatically when a user opens a page, so the agent always starts informed.
• Level 2 — Agent Knowledge: Five structured prompt types assembled hierarchically at inference time — Persona Override, Custom Instructions, Schema Guidance, Action Directives, and Guardrail Prompts. The AI-assisted Prompt Builder helps write prompts without prompt engineering expertise.
• Level 3 — Agent Ability: 12 pre-built tools (CRUD operations, email, tasks, events), Salesforce Flow integration for any custom business action, and Tool Builder for admin-created custom tools — REDA AI generates the Apex code automatically.

No. Tool Builder guides you through defining a custom tool's purpose, inputs, and outputs in a structured form. REDA AI generates the underlying Apex code automatically from that definition. Admins with no Apex knowledge can create fully functional custom tools.

For organisations with developers, the generated Apex code can be reviewed and further customised. The REDA AI Apex API also allows programmatic agent invocation from any Apex context for deeper integration with existing org logic.

When you click "Build Agent with AI" and describe your agent's purpose, REDA AI runs a four-step automated process that completes in under two minutes:

• Schema Analysis — reads your org metadata (objects, fields, relationships, picklists)
• Prompt Generation — creates an optimised persona, schema guidance prompts for your custom objects, and guardrail prompts appropriate for the target channel
• Context Field Recommendations — identifies the highest-signal fields for the target object type
• Suggestion Chip Generation — creates ready-to-use interaction chips based on the agent's role and common user actions

You review the generated configuration, make any adjustments, and deploy. The AI does the heavy lifting; you retain full control.

Yes. Through Tool Builder, you can create custom tools that make Apex callouts to external APIs, REST services, or any system accessible from your Salesforce org. Combined with Salesforce Flow integration, your agents can interact with virtually any external system — ERP, marketing automation, support platforms, financial systems.

All orchestration logic, audit data, and agent configuration remain inside Salesforce. The agent calls out; it doesn't export data.

Yes. Each plan supports multiple independently configured agents — 5 on Growth, 20 on Business, unlimited on Enterprise. Each agent has its own persona, prompt configuration, tool set, channel deployment, and credit tracking.

A common deployment pattern: a sales intelligence agent on Lightning Pages for the sales team, a support triage agent on Experience Cloud for customers, and an operational reporting agent triggered by Flow for back-office automation — all running simultaneously, each monitored independently.

REDA AI supports four deployment channels, each with its own security model:

• Salesforce Lightning Pages — embedded directly on record pages; Salesforce session authentication; full FLS, sharing rules, and profile permissions enforced; Action Confirmation Cards for write operations
• Salesforce Experience Cloud — inside customer or partner portals; inherits authenticated user's Experience Cloud permissions
• External Widget — JavaScript snippet embeddable on any website; Guest User authentication; allowed-origins whitelist; daily credit caps; read-appropriate tools recommended
• Salesforce Flows — invoke agents programmatically from any Flow type; enables fully unattended automation without a user present

Yes. Once an agent is configured, you can deploy it to any combination of channels. The core persona, knowledge, and tool configuration is shared, while you can customise guardrail prompts and tool assignments per channel to match the security requirements of each context.

For example: the Lightning version of an agent might have full read-write access to records, while the External Widget version of the same agent is restricted to read-only tools and informational responses. Same agent, different capability profiles per channel.

The External Widget is a lightweight JavaScript snippet you embed on any website. Visitors interact with your agent without a Salesforce login — authentication runs through Guest User credentials in your org.

The agent's data access is restricted to what your Guest User profile permits. Security configuration for this channel includes:

• Allowed-origins whitelist — specify which domains can embed the widget
• API key encryption for the embed token
• Daily credit caps to prevent abuse or unexpected overages
• Conservative tool assignment recommended (read-only operations)

You add the REDA AI component to any Experience Cloud page through the standard Experience Builder — drag, drop, configure. The agent inherits the authenticated user's Experience Cloud permissions automatically, seeing only the records and fields that user is permitted to access.

This makes REDA AI on Experience Cloud ideal for customer self-service portals (where customers can query their own account data), partner portals (where partners can access relevant pipeline and case information), and community sites — all without any compromise to your sharing model.

Yes, through the Salesforce Flows channel. You can invoke REDA AI agents from any Flow type — record-triggered Flows (when a record is created or updated), scheduled Flows (run nightly or weekly), or platform event-triggered Flows (on system events).

This enables fully unattended automation: an agent can analyse a newly created lead record and enrich it, generate a post-close summary when an opportunity is won, draft a renewal notice 60 days before a contract expires — all without a user opening a page, running as part of your existing workflow automation infrastructure.

Onboarding scales with your plan:

• Trial & Growth — guided setup walkthrough in the Admin console, access to the full REDA AI knowledge base and documentation, video tutorials, and a Dedicated Account Executive to answer questions
• Business — everything in Growth, plus structured onboarding sessions and priority access to the support team
• Enterprise — Dedicated Implementation Team with hands-on configuration support, a structured onboarding programme, and Consulting & Managed Services access
• Enterprise+ — white-glove onboarding with an executive sponsor, custom onboarding timeline, and quarterly business reviews

All plans include access to the REDA AI knowledge base, documentation, and standard support channels. Plan-specific inclusions:

• Growth ($1,500/mo) — Dedicated Account Executive
• Business ($2,500/mo) — Dedicated Account Executive, priority support queue
• Enterprise ($5,000/mo) — Dedicated Implementation Team, Consulting & Managed Services
• Enterprise+ (Custom) — Custom SLA, executive sponsor, quarterly strategic reviews

You can reach the REDA AI team at sales@reda.one or through your account executive for plan-related enquiries.

Yes. REDA AI offers Consulting & Managed Services on Enterprise and Enterprise+ plans. Engagements cover custom agent design and configuration, implementation support for complex use cases, ongoing optimisation and performance tuning, and strategic AI roadmap advisory.

If you need to design agents for a highly specific workflow, migrate from another AI platform, or run a structured proof-of-concept before committing to a full rollout, the REDA AI team can engage directly. Contact sales@reda.one with your scope and timeline.

REDA AI's Incident Log inside your Salesforce org automatically captures and classifies errors — Critical, High, Medium, Low — with root cause data, user and record attribution, and resolution status tracking. AI-assisted root cause analysis helps identify the issue without exposing raw user data. This is your first port of call for most issues.

For platform issues that require REDA AI team involvement, contact sales@reda.one or reach out through your Dedicated Account Executive. Enterprise and Enterprise+ customers have priority response SLAs. All error data stays inside your org — nothing is sent externally for diagnostics without your explicit consent.