SleekFlow Product Requests

Today, AgentFlow treats all knowledge sources equally during retrieval. When multiple sources contain overlapping information (e.g. Custom Answers, indexed website URLs, PDFs, Knowledge Base articles), the model determines which source to retrieve from and there is no way for administrators to control source priority. This creates challenges for enterprise deployments where customers have curated and approved answers that should always take precedence over website content. Expected behavior: If a customer asks a question that exists in the approved Q&A set, the AI should prioritize and return the approved answer. Actual behavior: The AI frequently ignores the Custom Answer and retrieves information from indexed website pages instead, generating a different response. This results in: Inconsistent answers. Hallucinations when website content is interpreted incorrectly. Difficulty validating AI behavior during UAT. Loss of customer trust because approved answers are not consistently used. Significant CS and SE effort spent troubleshooting behavior that cannot currently be configured. Proposed Solution **Introduce Knowledge Source Prioritization settings within AgentFlow. Example: Priority Order: Custom Answers PDFs Indexed Website URLs Live Web Search Alternative approach: Allow administrators to assign a priority score to each source type. Benefits, Customer Benefits: More predictable AI behavior Greater trust in approved responses Easier compliance and governance Better enterprise adoption Faster onboarding and implementation Easier UAT validation Reduced troubleshooting effort Clearer expectation setting with customers Improved answer quality Reduced hallucination risk Stronger enterprise readiness for regulated industries such as education, finance, healthcare, and government. Additional Consideration: A future enhancement could include displaying the source used for each answer and allowing admins to configure fallback behavior when higher-priority sources do not contain a relevant answer.

The current Broadcast module is too limited for customers running large-scale, automated engagement programs. Today, broadcasts primarily follow a simple workflow: Select contact list Select template Send or schedule For more advanced use cases, customers are forced to move into FlowBuilder because Broadcasts do not support dynamic audience management, trigger-based sending, or recurring automated campaigns. This creates several challenges: Increased Flow Builder complexity. Higher maintenance overhead. Excessive use of flows for use cases that are fundamentally campaign/broadcast activities. Additional MAC consumption as contacts must enter flows to receive automated outbound communications. Requested Enhancement: Expand the Broadcast module to support advanced automation capabilities, including: Dynamic Audiences Automatically updated recipient lists based on contact properties, labels, lifecycle stages, custom objects, or segmentation rules. Dynamic inclusion/exclusion criteria. Recurring Broadcasts Daily, weekly, monthly, and custom recurring schedules Automated re-send logic based on audience conditions. Trigger-Based Broadcasts Send broadcasts when a contact meets specific criteria. Send broadcasts based on lifecycle stage changes. Send broadcasts based on labels, custom object updates, or contact property changes. Send broadcasts after a defined period of inactivity. Campaign Journey Automation Ability to create lightweight campaign journeys for engagement and re-engagement use cases. Business Impact: Many outbound marketing and customer engagement scenarios do not require a full FlowBuilder journey. A more powerful Broadcast module would: Reduce operational complexity. Improve scalability. Reduce reliance on Flow Builder for campaign automation. Provide a more intuitive experience for marketing and lifecycle teams. Better support enterprise-scale outbound programs. Customer Example: MVF currently uses FlowBuilder for many automated outbound campaigns because Broadcasts cannot support dynamic audiences, recurring sends, or trigger-based execution. These use cases would be more naturally managed within an enhanced Broadcast .module.

Customers deploying AI agents at scale need visibility into both individual AI conversations and aggregate AI performance in order to evaluate quality, identify improvement opportunities, and scale AI responsibly. Today, labels can be used for segmentation and operational tracking, but there is limited native reporting specifically focused on AI conversation outcomes, customer intent, sentiment themes, and AI performance trends. Requested Enhancement: Provide dedicated AI conversation analytics and reporting capabilities, including: Conversation-Level Analysis Ability to review AI conversation outcomes on a per-customer basis. Visibility into whether an AI conversation was successful. Visibility into handoff reasons. AI resolution/completion tracking. AI vs human escalation tracking. Aggregate AI Reporting AI engagement rates. AI completion rates. Human handoff rates. Conversion outcomes from AI conversations. Complaint/escalation trends. Performance by AI use case or chatbot. Intent Analytics Automatic identification and reporting of common customer intents. Intent distribution trends over time. Ability to filter/report by intent category. Sentiment Analytics Common themes surfaced in customer sentiment. Positive/neutral/negative sentiment trends. Escalation risk indicators. Business Impact: As customers expand AI usage, they need both conversation-level visibility and aggregate reporting to: evaluate AI effectiveness. identify areas for optimization. understand customer behaviour. measure AI ROI. confidently scale AI across additional use cases.

Allow SleekFlow's AgentFlow AI Agent to accept and read PDF files shared by customers during conversations. The agent should be able to extract text content from the PDF and respond intelligently based on its contents. For example use case is Lab Report Processing: When a customer sends a lab report PDF, the AI Agent should be able to extract key fields such as: Name: Patient / customer name Report Date: Date the report was issued Report Name: i.e. Full Blood Count, Lipid Panel Report Type: Blood Test, Urine Test, Radiology Once extracted, the agent can use this data to personalise the response, route to the right team, update a CRM record, or trigger a follow-up flow all without human intervention. This would unlock high-value verticals like healthcare, insurance, and wellness where document intake is a core part of the customer journey. Thank you!

This feature provides a simplified setup process for the AI assistant, enabling users to build their own internal co-pilot agent. By streamlining the configuration, users can quickly create a customized digital assistant tailored to their specific workflows and needs.

This feature provides an internal co-pilot agent designed to assist users by answering questions based on uploaded knowledge base materials. Users benefit from immediate, automated support and easy access to relevant information, helping them find answers and complete tasks more efficiently.

The AI is currently limited to plain text responses. While it can present options to end-users within a guided flow, users are required to type out their answers manually, which creates unnecessary friction in the conversation. We'd like to request the ability for the AI to send messages with List options or Quick Reply buttons, allowing end-users to simply tap their choice rather than typing it out. This would significantly improve the overall customer journey on WhatsApp and make guided flows feel a lot more seamless and intuitive.

For the data source from AI Assistant reply, I would like to have both the URL link and File Name to be shown and the URL link should be clickable for user to easily cross check with the data source

This feature allows the AI to automatically transcribe incoming voice messages into text. Users benefit by receiving these transcriptions as internal notes, making it easier to quickly understand and manage customer communications without needing to listen to the audio.

A built-in evaluation system that lets you set quality standards for your AI agents and continuously monitor performance against those benchmarks. Detect regressions, track improvements over time, and replace manual spot-checking with structured, ongoing quality measurement.