We are excited to announce the release of Kanvas Ecosystem API v1.99.5! This release significantly expands Kanvas’ core architecture as the operational nervous system for modern businesses. In this update, we focus on three major pillars: ingesting company knowledge from active workspace conversations, AI-powered commerce intelligence, and highly contextual, human-in-the-loop workflows.
Here’s a deep dive into what's new in v1.99.5:
The biggest highlight of v1.99.5 is the introduction of the Slack Company Brain ingestion layer.
Historically, AI agents have been restricted to static document repositories, CRM databases, or manually indexed knowledge. With the new Slack ingestion layer, Kanvas can continuously capture real operational conversations across your Slack workspace and turn them into dynamic context.
Slack Conversations ➔ Kanvas Brain ➔ Search / RAG ➔ AI Agents
By preserving the original channel and user context, agents don't just read words—they understand where information originated, allowing the Company Brain to act as a living, breathing memory layer for the entire organization.
We have rolled out a major upgrade to our commerce intelligence layer, featuring product recommendation capabilities backed by Retrieval-Augmented Generation (RAG) and Typesense.
Instead of navigating classic category trees or entering exact keyword matches, customers can now interact with conversational commerce agents to describe exactly what they need in natural language (e.g., "I'm looking for a luxury SUV for a big family" or "Algo elegante para mi mamá").
Kanvas indexes product and variant data dynamically so agents can understand customer intent, match it with the catalog, and serve perfect contextual product recommendations.
Following the Slack-driven approvals introduced in v1.99.4, we have taken back-office automation a step further.
Now, when automated finance agents like Apex (AP) or Arc (AR) prepare a pending bill or invoice, the actual source invoice PDF is attached directly to the Slack approval request.
Invoice ➔ Agent Processing ➔ Kanvas Pending ➔ Slack + PDF ➔ Human Review ➔ Approval ➔ Execution
Reviewers no longer have to log into secondary tools to audit transactions. They can view the source document directly inside their Slack thread, ensuring a seamless, high-trust human-in-the-loop workflow.
Our Voice Agent runtime is now smarter. When a call comes in, the voice agent automatically resolves the incoming phone number to its corresponding Kanvas lead or contact record.
Incoming Call ➔ Phone Number ➔ Lead / Contact ➔ Context ➔ Voice Agent
By retrieving lead context instantly, voice agents can engage in personalized conversations right from the start, opening up unified omnichannel communication that flows seamlessly from voice to SMS, email, and Slack.
Kanvas continues to bridge the gap between AI reasoning and business execution. By giving your agents access to real-time communication channels, business files, and human-in-the-loop reviews, we are building a more connected and autonomous workforce.
Explore the full changelog on GitHub: v1.99.4...v1.99.5
We are excited to announce the release of Kanvas Ecosystem API v1.99.8! This release marks an important milestone in bridging autonomous agent intell...
Expanding Company Knowledge, AI-powered Product Discovery, and Human-in-the-Loop Agent Workflows. We are thrilled to announce the release of Kanvas Ec...
There’s a new term going around: agentic commerce . Sounds complicated. It really isn’t. Today, you go to a website, search for what you want, compare...
Let's talk about how MCTEKK can help you scale your commerce operations with a backend that just works.
