Relay.app Features 2026: Automate Business Workflows Fast

What Is Relay.app and Why It Matters for Business Automation in 2026

Relay.app is an AI-powered workflow automation platform that lets business and operations teams connect their everyday tools — Gmail, Slack, HubSpot, Notion, and dozens more — without writing a single line of code. What separates it from older automation tools is its native support for human-in-the-loop (HITL) approvals, branching logic via Paths, and a built-in AI Assistant that can summarize, classify, and extract data mid-workflow.

As of early 2026, Relay.app has rapidly expanded its integration library (adding 101 new integrations in January 2026 alone), introduced native Tables for in-platform data storage, and rolled out MCP (Model Context Protocol) support — positioning itself as a serious competitor to tools like Zapier and Make for teams that want AI built into their automation logic rather than bolted on.

This guide covers every major Relay.app feature, who each capability is best suited for, where the platform falls short, and how it stacks up against alternatives.

Core Relay.app Features Explained

Workflow Automation with Triggers and Paths

The foundation of Relay.app is its visual workflow builder. You set a trigger — a new lead in HubSpot, a form submission, a scheduled time, or an inbound webhook — and Relay executes a sequence of actions across your connected apps. Crucially, each workflow can branch using Paths, which create conditional logic based on data values or outcomes.

A real-world example: a SaaS company uses Relay to handle customer onboarding. When a new deal closes in HubSpot, Relay automatically creates a Google Workspace account, posts a Slack welcome message to the internal team, assigns onboarding tasks in Notion, and then holds — waiting for an HR manager's approval before sending the client their login credentials. This combines app automation with a controlled, human-reviewed checkpoint.

The main limitation of Paths is visibility: deeply nested branching flows can become hard to audit or debug, especially as workflows grow. Teams running 10+ condition branches should expect to invest time in documentation or risk creating unmaintainable automations.

Human-in-the-Loop (HITL) Approvals

One of Relay.app's most distinctive features is its native HITL support, updated significantly in early 2026. Workflows can pause mid-execution and route a decision to a human reviewer before continuing. This is invaluable for compliance-sensitive processes: contract approvals, large purchase orders, client-facing communications that need sign-off.

The current limitation worth knowing: approvers must be registered Relay.app users. This means you can't route an approval to an external stakeholder or a client via email without them having a Relay account — a friction point that teams handling external approvals will feel immediately. Compare this to tools like Workato, which supports external approval via email or Slack without requiring platform accounts.

Built-in AI Actions

Relay.app embeds AI directly into workflow steps rather than treating it as a separate integration. You can insert an AI action to summarize a long email thread, classify a support ticket by priority, extract structured data from unstructured text, or generate a draft response — all within the same workflow that also creates a CRM record or posts a Slack notification.

As of late 2025 and into 2026, Relay supports multiple AI models within workflows:

• GPT-5.1 (OpenAI)
• Claude 4.5 Opus (Anthropic)
• Gemini 3 Pro (Google)
• Video generation via Sora 2 and Veo 3.1

This multi-model flexibility means you can pick the right model for the task — use Claude for nuanced text classification, GPT-5.1 for code-related steps, and Gemini for tasks involving Google Workspace data.

Native Tables

Launched quietly in November 2025 and enhanced through early 2026, Tables is one of Relay.app's most significant recent additions. Teams no longer need to maintain an external Google Sheet or Airtable database for lookups, logging, or data storage — they can create native Tables inside Relay that plug directly into workflows and agents.

Tables support multi-select fields, default values, multiple data types, and can be triggered on row creation or updates. For teams that previously relied on a Google Sheets integration as a lightweight database layer, this eliminates a fragile dependency and keeps automation logic fully self-contained.

Integrations and Connectivity

Relay.app connects to major business tools across CRM, communication, project management, finance, and developer infrastructure. The integration library expanded significantly through 2025 and into 2026, with notable recent additions including:

• Microsoft Planner — full task and plan management (create, update, trigger on task creation)
• Microsoft SharePoint — file management, list operations, drive triggers
• Snowflake — execute SQL, add/find rows, navigate databases
• Telegram — trigger on new messages, send messages
• Notion — 12+ new actions including page copy, trash, comment management
• Shopify — article and blog management

The honest caveat: Relay still has fewer niche connectors than Zapier, which covers thousands of apps. If your stack includes a specialized industry tool — a niche CRM, a vertical SaaS product, or legacy software — there's a real chance Relay doesn't support it natively, and you'll need to use webhooks or the API to bridge the gap.

Relay.app Feature Comparison Table

Relay.app Pricing Overview

Relay.app offers a free plan to get started — suitable for individuals testing workflows or small teams automating a handful of processes. Paid plans begin at approximately $9 per user per month on the Pro tier (billed annually), which unlocks higher run volumes, more active workflows, and access to premium AI model steps.

Business-tier plans run approximately $18–$29 per user per month and add priority support, advanced collaboration features, and higher limits on Tables and AI usage. Enterprise plans for larger teams with SSO, audit logs, and custom data retention typically start around $500+/month depending on team size and AI usage volume.

For comparison: Make starts at $9/month for 10,000 operations — a task-based model rather than user-based — which can be more cost-effective for low-user, high-volume automation. Teams evaluating both should map their actual workflow run frequency to decide which pricing structure favors their use pattern.

Common Mistakes Teams Make with Relay.app

Mistake 1: Over-nesting Paths Without Documentation

Teams often start with a simple branching workflow and keep adding conditions over time. Six months later, a workflow with 8 nested Paths becomes impossible to debug when something breaks. The fix: treat each major branch as a potential candidate for a separate sub-workflow, and add a description to every Path condition at build time — not retroactively.

Mistake 2: Assuming All Approvers Have Relay Access

HITL approval workflows fail silently when an approver isn't a Relay user. A common scenario: a finance team sets up an expense approval flow and routes it to a department head who has never logged into Relay. The workflow hangs indefinitely. Before launching any approval-based automation, verify every approver has an active account and knows how to respond.

Mistake 3: Using Tables as a Full CRM or Database Replacement

Native Tables are excellent for lightweight lookups and logging — tracking processed records, storing reference data, or maintaining a rolling log of workflow outputs. They are not a replacement for a relational database. Teams that try to run complex queries, manage relationships between multiple tables, or store thousands of records often hit performance limits. Keep Tables for simple, flat data structures.

Mistake 4: Skipping Error Handling on External API Steps

Relay's retry policies for failed steps are not comprehensively documented. Teams that build workflows relying on third-party API calls — especially to rate-limited services — often find runs silently fail or get stuck. Build explicit error Paths for every external API step: one branch for success, one branch that pings a Slack channel when the step fails, rather than assuming errors will surface naturally.

How Relay.app Compares to Key Alternatives

Relay.app sits in a competitive market. Here's how it compares honestly against the tools you're likely considering:

• vs. Zapier: Zapier has a vastly larger integration library (6,000+ apps vs. Relay's hundreds). Relay wins on native AI integration and HITL — Zapier's AI features feel bolted on by comparison. For teams whose stack is fully covered by Relay's integrations, Relay offers a more cohesive experience.
• vs. Make: Make is more powerful for data transformation and complex multi-step flows, with a lower price per operation. Relay is more beginner-friendly and has better collaboration features built in. Make users who need human approvals typically resort to workarounds; Relay handles it natively.
• vs. n8n: n8n is self-hostable and open-source, making it the default choice for teams with strict data residency requirements. Relay is fully managed and cloud-hosted. If self-hosting is a requirement, n8n wins outright; otherwise Relay's AI features are more polished.
• vs. Microsoft Power Automate: Power Automate is deeply integrated into the Microsoft 365 ecosystem. Teams running on SharePoint, Teams, and Outlook get deep out-of-the-box value. Relay now supports SharePoint and Microsoft Planner, but for Microsoft-first organizations, Power Automate remains the default choice.

Who Should Use Relay.app in 2026

Relay.app delivers the most value for operations and business teams at SMBs and mid-market companies (roughly 10–500 employees) that:

• Use a modern SaaS stack (HubSpot, Slack, Notion, Google Workspace, Shopify) that Relay covers natively
• Need workflows where humans must review or approve outputs — not just fully automated pipelines
• Want AI-powered steps embedded in automation without managing separate AI infrastructure
• Are looking for a Zapier alternative with a better collaboration model and more modern AI capabilities

It's a poor fit for teams with niche or legacy integrations not yet supported, organizations requiring self-hosted infrastructure, or enterprises needing complex role-based access controls beyond the current plan limits.

The platform's trajectory through 2025 into 2026 — MCP support, multi-model AI, native Tables, 101+ integrations added in a single month — signals an aggressive roadmap. Teams evaluating workflow automation tools today are looking at a platform that will be meaningfully more capable in six months than it is now.