Help people find answers, make informed decisions and get work done. Configure agents to work as copilots inside Salesforce, provide record recommendations, run through automations and APIs, and support interactions in Claude, customer and partner portals, and your website.
Shape their instructions, context and enabled actions around your processes, with visibility into usage and controls over access and spending.
Your foundation for assistance & automation
Bring your own AI accounts or use REDA credits. The funding model applies to your Salesforce org; talk to REDA to switch.
Pay your provider directly. REDA supplies the orchestration platform, with no markup on provider usage.
Anthropic · OpenAI · Gemini · DeepSeek
Use the shared REDA credit allowance in your plan. Choose processing tiers and track consumption inside Salesforce.
REDA supplies and operates the AI provider accounts. Additional credits are available as top-ups.
Your provider supplies the AI. Your REDA subscription supports the platform that puts it to work: Salesforce orchestration, configurable governance, admin tools, support and continued product improvement as AI evolves.
Customer-account usage is still metered for reporting and your own usage limits. It is not deducted from a REDA allowance. The provider’s console and invoice show the actual provider charges.
Review the data and credential boundaries →Compare both pricing models →Your admin knows the business. REDA AI helps turn that knowledge into a configuration they can review, test and improve.
Create an agent with a name and a role, then ask REDA AI to review its setup. It proposes improvements across instructions, settings, record context and suggested questions.
You decide which suggestions to apply. Test the configuration against real scenarios before making it available to your team.
See how Build Agent with AI works (opens in a new tab)
7:11Describe the work in everyday language, review the proposed configuration and place Alex on an Opportunity page.
Start with a business need and test a working service assistant.
Watch demoVideos play on YouTube. Privacy policy.
Set the role and purpose. Add guidance or a representative sample file.
See proposed changes and their rationale alongside your business requirements.
Keep what fits, refine what needs work and test the result with your team.
Use AI to draft or refine instructions around the agent’s purpose. Compare the suggested rewrite with your original before deciding what to keep.
Built-in quality checks flag repetition, length and structure. Scores describe these checks—not answer accuracy—and token figures are estimates.
Explore the Prompt Optimizer (opens in a new tab)
4:20Extend Alex’s service workflow with warranty guidance using Prompt Builder, then verify the case update and customer email.
Adapt an existing agent as your business requirements evolve.
Watch demoVideos play on YouTube. Privacy policy.
Provide the procedure, purpose and any specific changes you want.
Review the wording, the explanation and estimated token counts.
Confirm that important rules remain and test representative scenarios.
Choose which record fields help explain the situation. REDA AI can suggest context fields and descriptions using your Salesforce object schema.
Your administrator reviews the proposed configuration. Relevant context starts with understanding your data model and who should have access to it.
See how context is configured (opens in a new tab)
3:01Extend Alex from Opportunities to Cases, review the added context and test a service follow-up from the Case page.
Help the same agent work across another part of your process.
Watch demoVideos play on YouTube. Privacy policy.
Choose the object and review its existing context configuration.
Get suggestions for fields and descriptions that support the agent’s purpose.
Check relevance and access, then test the information available in the experience.
Start with support. We build the first scoped pilot with you. These tools support your administrator’s work; business knowledge and testing remain essential. AI-assisted analysis uses REDA credits with AI included, or your configured provider account with Bring your own AI.
Read the setup guide (opens in a new tab)Start with built-in capabilities. Add business-specific tools when your process needs more.
Choose what each configuration needs, subject to channel availability and permissions.
Explore the full catalogue (opens in a new tab)Query records, retrieve details and work with your Salesforce object structure.
Records · queries · schemaCreate or update records, prepare follow-up work and send communications through permitted actions.
Records · tasks · emailSchedule an action, continue a configured workflow or consult another configured agent.
Scheduling · workflows · delegationExpose a parameterised Salesforce query for a specific business question.
SOQL and configuration knowledgeMap CSV columns to Salesforce fields for a defined import process.
Mapping and validation setupUse an Apex-backed tool for specialised logic or an integration.
Development, deployment and testingBuilt-in count verified for REDA AI 1.16000. Tools and actions are building blocks; your configuration determines how they are used.
Make configured assistance and automation available to your team, your processes, your customers and your partners. Explore each experience below.
Copilot for your team
Give your people a copilot in Salesforce, configured around their roles, your business processes and policies. Your people start the conversation: ask questions in plain language, review the findings and decide which actions to take—all within the permissions and guardrails you set.
Less time piecing things together. More time for your people to move work forward.
more issues resolved per hour on average in a study of 5,172 customer-support professionals using an AI assistant.
Independent research · Brynjolfsson, Li & Raymond · QJE, 2025. Read the study ↗
Study findings, not a measured REDA AI result.
See Maya review a loan application and Clara help a financial controller explore business records. Each copilot is configured for a different role and process.
4:24Review application documents, update Lead fields and follow up with a branded email—all from one Salesforce workspace.
See the findings become reviewed actions your team can track.
Watch demoVideos play on YouTube. Privacy policy.
4:52Ask about overdue invoices, investigate customer balances and compare revenue movements with Clara in REDA Financials AI.
Move from a finance question to supporting records and the next follow-up.
Watch demoVideos play on YouTube. Privacy policy.
Recommendations for your team
Give your people guidance right on the Salesforce record, without having to start a chat. REDA AI uses the record’s information and the procedures and policies you configure to highlight what needs attention, explain why and suggest the next step—helping new joiners and experienced colleagues follow a consistent approach.
Less guesswork. Clearer priorities. More consistent ways of working.
months of tenure. In a customer-support study, staff with two months’ tenure using AI assistance performed comparably to staff with more than six months’ tenure without it.
Independent research · Brynjolfsson, Li & Raymond · QJE, 2025. Read the study ↗
Customer-support AI assistance; not a test of record-based recommendations or REDA AI.
This sample B2B sales configuration highlights follow-ups, potential risks and information that needs checking—all within the opportunity. Your recommendations are tailored to your own records, roles and procedures.
A concise recommendation brings the opportunity’s context into focus while your people work in Salesforce.
Enlarge screenshot ⤢
Open the detail to review suggested actions, potential risks and gaps in the available information before deciding how to proceed.
Enlarge screenshot ⤢Automation & APIs for your processes
Let a business event start the work. Configure REDA AI to respond, work through your procedures and carry out the actions you allow, bringing in your people when judgement or approval is needed.
Connect it through Salesforce Flows and APIs, and extend its tools and actions as your business needs evolve.
Less routine work to coordinate. More time for the decisions that need your people.
AI-assisted email work
2 hours
less time on email per week among workers who used the AI tool, in the second half of a six-month randomized study across 66 firms.
Dillon et al. · NBER, 2025.
Read the study ↗
Study of staff using AI assistance; does not measure autonomous workflows or REDA AI.
AI in accounting workflows
7.5 days
shorter month-end reporting lag associated with AI adoption in field research covering 79 small and mid-sized firms.
Choi & Xie · Journal of Accounting Research, 2026.
Read the study ↗
Association with AI adoption; does not isolate autonomous workflows or measure REDA AI.
Follow a WhatsApp enquiry through REDA Connect and REDA One, with Mira using Salesforce records and configured procedures to respond and update the inquiry.
6:26Follow a WhatsApp enquiry from REDA Connect into Salesforce, as Mira checks property records, gathers requirements and updates the inquiry.
Connect your data and procedures to work arriving through other channels.
Watch demoVideos play on YouTube. Privacy policy.
Claude & MCP for your team
Let your people work with Salesforce from Claude through REDA AI. Requests run through agents configured with your business knowledge, procedures and permitted tools and actions, using the signed-in user's Salesforce access.
Bring the results into Claude to analyse, chart and turn into reports—without rebuilding your business instructions in every conversation.
Less exporting and re-explaining. More time to turn answers into useful work.
nonprofit applications evaluated in under an hour. Smartsheet reports that its corporate social responsibility team used its MCP connector with Claude to evaluate the applications.
External customer case study · Smartsheet / Anthropic. Read the case study ↗
Company-reported example; not a measured REDA AI result.
Follow Clara as a financial controller explores Salesforce records in Claude and prepares a monthly briefing. The same connection can support other roles, records and procedures.
5:02Use Clara through MCP to explore Salesforce finance records, compare months and prepare a management briefing with charts and a PDF.
Keep Salesforce access scoped while working with returned findings in Claude.
Watch demoVideos play on YouTube. Privacy policy.
Customer & partner portals
Give customers and partners a clearer way to get help and move their requests forward. REDA AI brings your business knowledge and procedures into your Salesforce portal, helping people understand what they need to do and take the actions you make available, using the information they are permitted to access.
When your team needs to step in, they can continue with the context already captured.
Less back-and-forth. A smoother experience for the people you serve.
A 29-percentage-point increase in self-service rate in Nubank's card-delivery support workflow, compared with earlier AI assistant versions.
Industry research · Gupta et al. · KDD, 2026. Read the study ↗
Company-authored study of an external deployment; not measured REDA AI performance.
This sample use case shows how REDA AI can be configured for customer support in a portal. Your setup can reflect other customer and partner journeys.
3:16See Gungun help a signed-in customer check a case and raise a new support request with the details the service team needs.
Carry customer context into support, under the portal user’s permissions.
Watch demoVideos play on YouTube. Privacy policy.
Built in your Salesforce. Available on your website.
Help visitors explore your products and services, understand what fits their needs and take the next step towards buying. REDA AI brings your business knowledge and sales processes into the conversation, configured around the information and actions you allow.
Connected directly to your Salesforce, it can capture requirements, create an enquiry or arrange a conversation with your team—with the context already there.
Less distance between a visitor's interest and your team's next opportunity.
Higher purchase conversion in a randomized field experiment with an AI-powered pre-sale chatbot on an online retail platform.
External research · Fang et al. · Reported by Columbia Business School, 2025. Read the research ↗
External online-retail research; not a measured REDA AI result.
This sample use case shows how REDA AI can be configured for product enquiries. Your setup can reflect different products, services and sales processes.
3:29Watch Penny answer product questions using current catalogue records, understand a visitor’s requirements and capture a Salesforce Lead.
Give your sales team the product discussion and enquiry context together.
Watch demoVideos play on YouTube. Privacy policy.
Go beyond setup. Give your admin visibility into how AI is being used and practical tools to investigate when something needs attention.
Set usage limits and review conversation activity, tokens and credits to understand consumption and focus your optimisation work.
Explore conversation analytics (opens in a new tab)Use the conversation debugger to examine an interaction and get suggestions to review against the agent’s configuration.
See the conversation debugger (opens in a new tab)Use AI-assisted incident analysis to investigate recorded issues and review potential next steps.
Explore incident analysis (opens in a new tab)Refinement is a review process. Diagnostic suggestions help your administrator investigate; they do not automatically fix configuration or prove a business outcome.
Understand prompt quality checks (opens in a new tab)REDA AI connects the model’s reasoning with your configured business context and the tools that perform the work.
A person asks for help, or a configured process invokes REDA AI.
REDA AI prepares the configured instructions, relevant data and available tools.
The configured external model processes the supplied context and may request tool calls.
REDA AI executes allowed tools under the applicable channel controls and returns results, with review where configured.
Be clear about the data flow. Relevant business content can be sent to the AI provider. Salesforce execution, external processing and each channel’s access controls are explained in our trust guide.
Governance & trustWe work with you to configure a first agent around an agreed problem and prove it in your Salesforce sandbox. The free pilot is a bounded proof of concept; a full production rollout is scoped separately.