Video summary

Marketplace & open platform 2026 roadmap session

Main summary

Key takeaways

Business

Executive direction (2025 lessons → 2026 AI-first execution)

  • 2025 was “disruptive,” forcing partners to change how they build products and run businesses—where to invest and what to build next.
  • 2026 is positioned as the first year to “go all-in with AI”, shifting from AI that assists users to AI that executes work.
  • Monday’s strategic platform shift (Monday → 2026):
    • From: “AI helping humans work on Monday”
    • To (2026): “best agentic work platform” where humans + agents (inside and outside Monday) execute work.

Platform strategy: the “4 pillars” playbook for marketplace/AI app success

These are the core focus areas repeated throughout the platform strategy:

  • Trust (foundation for adoption)

    • “No trust, no adoption” (especially for enterprise and regulated environments).
    • Emphasis on:
      • Keeping data within customer-defined boundaries
      • Providing admin controls over AI
  • Richness & Depth (differentiation + stickiness)

    • Compete against “lighter/vibe apps” by embedding deeply into workflows.
    • Value comes from:
      • Deep domain expertise
      • High-impact use cases
      • Frequent reliance
  • Adoption (discoverability → monetization → retention)

    • Solutions must be:
      • Found
      • Used
      • Paid for
    • Focus on:
      • Net retention (reduce churn)
      • Activation (increase usage of installed apps)
  • Developer Experience (speed + low friction)

    • AI-era tooling to iterate faster and ship with fewer blockers.
    • Reduce integration and operational overhead for builders.

AI solution types partners can build (with examples)

The platform categorizes marketplace offerings into three build types:

  • AI-Powered Apps (LLM value-add)

    • Example: A risk analyzer on a board that highlights issues, summarizes them, and helps users understand what needs addressing.
  • Agents (decision + action + ongoing/background execution)

    • Example: The same risk app can include an agent that:
      • Pings assignees for updates
      • Escalates to stakeholders with improved visibility
  • Skills (triggered tasks, often via Sidekick / in-workflow)

    • Example: After escalation/summarization, trigger a Slack/email update to notify the team without leaving context.

Trust: 2025 capabilities recap + 2026 security roadmap

Key 2025 trust features (concrete mechanisms)

  • Data residency / regional storage

    • Monday Code provides region-scoped storage aligned with account regions.
    • New: added IL region for multi-region apps.
    • Storage services mentioned:
      • Key-value storage
      • Document storage (structured + semi-structured)
      • Object/blob storage (large files like PDFs/media)
      • Secret storage (“vault”) for sensitive tokens/secrets
  • Network/IP controls

    • IP allow list (apps can send data only to allowed domains/IPs)
    • Static IP exposure for firewall whitelisting
  • Code security scanning

    • Backend code security scan to detect vulnerabilities and hard-coded secrets/tokens
  • Marketplace trust signals

    • A partner page with reliability indicators (e.g., when joined marketplace, total installs)
    • Ratings/reputation visibility
    • Feedback loop via the developer center

2026 trust/security enhancements (forward-looking)

  • Service token (long-awaited)
    • Allows apps/agents to operate independently (not tied to a specific user), preventing breakage when users leave.
  • Refresh token support
    • Shorter-lived access tokens via refresh flow
  • Frontend code security scanning
    • For files uploaded to the CDN
  • AI governance emphasis
    • Permissions controls at admin/workspace/board levels
    • Activity logging
      • Board activity log
      • Admin audit log
    • Agent Center to manage agents (including kill/suspend/pause/resume if they go rogue)

KPI / business metric targets mentioned (in trust/adoption context)

  • No explicit CAC/LTV/CAC payback targets were provided.
  • Trust is treated as a prerequisite; adoption and churn are quantified elsewhere.

Richness & Depth: embedding + building blocks

Developer platform expansion for deeper workflow integration

  • Enhanced developer account

    • Access to all Monday products to test apps across surfaces
    • Default 1k items test capacity (partners can request more for scale testing)
  • Extension points / UI programmability

    • Ability to customize columns/fields such as:
      • Formula
      • Dropdown
      • Status
      • People column
      • File column (customization mentioned “very soon”)
  • API expansions

    • APIs for managing board views, dashboards, columns, and other “building blocks”
    • Aggregation API for widget calculations (max/sum/average) at scale
    • Access to higher-layer objects (e.g., project portfolios)
    • Bulk API to import large item sets from files into a board
  • Workflow & automation integration

    • “Long awaited” ability to manage/control automations from apps via:
      • Webhooks to register/subscribe when users interact with app features (columns/views) and then dispatch logic
    • Mention of a set of events; guidance to request missing events if needed
    • Multi-level boards
      • Access rollups and aggregated stats on hierarchical structures

Integration and execution capabilities (operations enablers)

  • Simplified credential handling
    • Token + secret management support to reduce partner implementation complexity
  • Automation improvements
    • Synchronous automation handling (wait for success/failure)
    • External trigger services with polling services as triggers (roadmap/mentioned; some subtitle confusion noted)

Agent/Sidekick exposure strategy

  • “Meet the user wherever they do work”:

    • Add capabilities inside Sidekick (conversational AI)
    • Encourage building skills and agents surfaced inside Monday
  • Agent building block (forthcoming)

    • Agents behave like assignable “users,” including:
      • Mentioning agents in updates
      • Adding agents to automations to run in the background
    • Exposed via an Agent Store (“released very soon”)
    • Partners encouraged to share agent use cases for collaboration

Adoption: monetization + retention mechanics (with concrete churn metrics)

Discovery, activation, and monetization improvements

  • Mobile support expansion

    • For:
      • Board views
      • Item views
      • Objects
    • Enables adoption scenarios beyond desktop (frontline, executives, teams)
  • Uninstall/unsubscribe friction UX

    • Adds friction during unsubscribe/uninstall decisions
    • Immediate reported impact:
      • -12% unsubscribes
      • -10% uninstalls
  • Churn reason exposure

    • In AMP (Application Management Product), churn reasons are exposed as an optional field to help diagnose why accounts leave.
  • Retention insight / targeting heuristic

    • Accounts with at least 3 active users have ~20–30% lower likelihood to churn
    • Strategy: boost adoption of installed apps by suggesting them to users not yet using them
  • Revenue protection when payment fails

    • Notify in product (and already started via email)
    • Offer a discount when users try to unsubscribe
    • Enforce payment based on accurate Monday seat count
    • Discount/payment enforcement described as opt-in
  • Listings/search optimization

    • Improve app listings so users better understand capabilities and use cases before install
    • Emphasize search engine discoverability and clearer expectations
  • Admin approval and purchasing flow

    • Expanded admin app management covering approvals for:
      • installed apps
      • purchases
    • Goal: shorten approvals and reduce admin friction

Explicit timeline/targets mentioned

  • “7 16 months” appears in subtitles without clear context.
  • Migration deadline: end of April for workflow blocks migration (see below).

Developer Experience: “AI-first” tooling and faster time-to-value

AI-driven developer workflow shift

  • Introduced concept: MCP as a conversational interface for agents accessing APIs.
    • Builders can get apps MCP to ship faster by delegating heavy lifting to AI coding assistants
    • Platform MCP is also mentioned for broader natural-language interaction with the platform

Concrete developer features and operations

  • Partners can gain programmatic control via:

    • CLI
    • API
    • Apps MCP
  • New platform/developer items:

    • Account slug: a user-friendly unique identifier for API access
    • Workflow blocks migration
      • Move from old sentence builder blocks to workflow blocks
      • Deadline: end of April
      • Enables automations across “all Monday surfaces,” including workflow builder and custom builder
    • Community hub
      • Product announcements, feature requests, conversations in one place
    • Issue tracker planned for this year
      • Track common issues and help partners catch up

Commercial enablement features for partners

  • Ability to grant discounts to existing subscriptions via dev center UI + API
  • Gradual ability to force customers to upgrade to the latest pricing version (for partners with older plan tiers)

Event/task guidance

  • Attend Alex’s talk at 11:30 for MCP details (encouragement noted).

Actionable recommendations implied by the roadmap

  • Prioritize trust-by-design
    • Ensure your app/agent follows governance and data-handling expectations; avoid patterns that trigger enterprise security concerns.
  • Build for embedded workflow value (Depth)
    • Use Monday extension points + APIs to integrate into views/columns/dashboards and workflow automation.
  • Design for adoption loops
    • Expect increased focus on:
      • discovery (listings/search)
      • activation (mobile + multi-surface)
      • retention (churn reason + friction UX)
  • Use AI developer tools
    • Adopt apps MCP / AI-first dev workflows to reduce iteration cycles and speed validation.

Presenters / sources mentioned

  • Noriel (leading the open platform group)
  • Rotel (leading apps framework and Monday Code product management)
  • Maria (referenced as “thank you, Maria” during the trust discussion)
  • Alex (talk at 11:30 about MCP)

Original video