Kenyt

API triggers let the agent retrieve and update information in connected CRMs, databases and other business tools during a conversation. The integration catalog spans Salesforce, Zoho, Shopify, Google Sheets, Stripe, Razorpay and telephony providers such as Exotel. Workflows can combine booking, verification, payment processing and follow-up actions. Custom connectors extend the same action paths to customer systems, while a unified dashboard carries the resulting customer data and campaign context.

Sourcekenyt.ai/agentic-airead 2026-10-20

Sequential workflows connect conversation decisions, messaging, orders and payment actions. A customer response or existing business context selects a decision-tree branch, and slot filling avoids asking again for known values. Builders define intents, actions and dialogues, then connect the workflow to customer systems. Examples include confirming an order, processing its payment, sending the receipt and starting later delivery or renewal communication, with departments able to set their own handling rules.

Sourcekenyt.ai/business-automation/multistep-workflow-automationread 2026-10-20

A self service assistant builds its knowledge from customer web pages and technical documentation, then uses that material for answers about the business. Teams can supply their own policies for employee questions and maintain searchable knowledge articles for support. The email product connects this content with response suggestions and links to relevant articles. Uploaded or extracted business knowledge stays associated with the configured assistant, while customer systems supply additional live context.

Sourcekenyt.ai/chatgpt-powered-ai-assistantread 2026-10-20

Email handling supports draft approval, tone controls and escalation rules before a team releases an assisted response. Campaign templates also pass through an administrator approval step inside Kenyt after the messaging provider approves them. The messaging API rejects inactive or unapproved templates. For live support, a teammate accepts an incoming conversation from the service desk, and human override preserves the interaction context while the person takes control.

Sourcekenyt.ai/docs/campaigns/creating-message-templatesread 2026-10-20

Organization administrators invite teammates and set separate access levels for Conversations, Campaigns and organization details through Manage Access. The data protection program uses encryption, least privilege, API access controls, backups and incident response, with a designated data protection officer, internal risk reviews and records of processing. Subprocessors sign processing agreements and international transfers use contractual safeguards. Kenyt names no SOC 2 report, ISO certification or other independent attestation of its own.

Sourcekenyt.ai/docs/dashboard/how-to-invite-a-new-team-member-to-the-kenyt-dashboardread 2026-10-20

The Conversations dashboard keeps history with date filtered exports, and large reports arrive by email when a direct download is too big. Email operations add a send activity tracker for delivery, open and read status, alongside resolution, response time and sentiment metrics, and customer and campaign dashboards tie those outcomes to user journeys. Review stays at the level of the conversation and the message: the intents matched, actions called and API responses inside an automated exchange are not laid out step by step.

Sourcekenyt.ai/docs/dashboard/how-to-export-conversation-reports-from-the-dashboardread 2026-10-20

Contextual Memory retains user preferences, earlier interactions and behavioral patterns for subsequent responses. Known values fill conversation slots, while the shared interaction history supports continuity across email, chat and voice. Email replies can also draw on previous threads held in the CRM. The persistent context follows the customer and business workflow; it has no separate memory lifetime or per-person memory controls independent of the underlying customer information.

Sourcekenyt.ai/personalization/user-contextread 2026-10-20

Customers configure assistants in Kenyt's hosted dashboard and deploy them to websites and messaging channels through integrations. The terms place the website service in India, with contractual safeguards covering international data transfers, and no setting lets a customer choose another hosting region. The self managed plan means the customer configures the account and agents itself; the software still runs only on Kenyt's hosted service, with no customer cloud or on premises installation.

Sourcekenyt.ai/terms-and-conditionsread 2026-10-20

Prebuilt conversational blueprints include more than 500 intents, 20,000 entities and 100 ready-to-use flows. Customers adapt those starting points to their business instead of writing every conversation from scratch. Named assistants address support, sales, marketing, HR, recruitment and IT service management, with industry configurations for areas such as real estate, retail and hospitality. The builder also generates additional training examples and entity variations for the selected workflow.

Sourcekenyt.ai/agentic-airead 2026-10-20

Customer actions such as booking an appointment, downloading content or sending a query can trigger a message sequence. Campaign stages support behavior-based follow-ups across email, WhatsApp and text messaging, while scheduled sends control timing. In a live conversation, inactivity can trigger contextual prompts after a configurable delay, with a five-second default. An intervening user response suppresses that prompt, and individual intents can disable the behavior.

Sourcekenyt.ai/developers/promptsread 2026-10-20

The assistant pairs structured natural language understanding for intents and entities with generative responses, and the self service assistant uses ChatGPT for answers grounded in the customer's web pages and documentation. Customers configure intents, examples, knowledge and response behavior inside that hybrid engine, and generative tools write extra training utterances and entity variants. Kenyt sets the model; administrators get no choice of model maker, no routing control and no way to bring their own model.

Sourcekenyt.ai/chatgpt-powered-ai-assistantread 2026-10-20

External applications can create or update a lead through the native Create Lead Webhook, using an authorization token, organization identifier and bot identifier. The request supports ownership, pipeline, stage and custom fields. A separate Instant Messaging API queues WhatsApp, text or email messages from approved templates, with arguments and attachments. Both interfaces return structured success or failure responses, including authorization errors, invalid inputs and rejected template states.

Sourcekenyt.ai/developers/create-lead-webhook-documentationread 2026-10-20

Built in testing simulates real world scenarios before launch, running test cases and edge case queries against the configured bot. Visual analysis then points out low performing intents and missing entities, and response benchmarks compare the bot's output with the customer's performance goals. After deployment, conversation feedback refines intent mapping as question patterns change, and feedback on email replies supplies a further improvement signal for the email agent.

Sourcekenyt.ai/generative-airead 2026-10-20

Website scraping builds the assistant's knowledge from customer pages before it answers questions. At runtime the work runs through conversational flows, API triggers and connected business tools, and the agent never opens a browser or desktop session to click through another application. Prompt cards show options the user picks, and the chosen text feeds back into intent matching inside the chat. Video calls and screen sharing on the service desk are tools for human staff, not for the agent.

Sourcekenyt.ai/developers/promptsread 2026-10-20