Skip to main content
Intempt
Coming soon

Next best action orchestration

A rules-based journey that routes each person to a next step from a candidate set using attributes and segment membership at branch points.

Authored by Intempt Technologies

Recipe curators

Agentic workflow

Coming soon

7 steps Blu runs end to end. It drafts a plan and nothing runs until you approve it.

  1. 1

    Decide the next step per user

    Coming soon

    Blu builds a step

  2. 2

    Only route confident calls

    Coming soon

    Blu builds a step

  3. 3

    Write an email per action

    Coming soon

    Blu builds a step

  4. 4

    Write the in app versions

    Coming soon

    Blu builds a step

  5. 5

    Build the recommendation set

    Coming soon

    Blu builds a step

  6. 6

    Re-read the signal at each gate

    Coming soon

    Blu builds a step

  7. 7

    Check the model against a control

    Coming soon

    Blu builds a step

About this recipe

When to use it

A fixed drip sequence sends the same next step to everyone regardless of what they actually need. This picks the next action per-user from a live signal, so the path adapts instead of marching on a schedule.

What it delivers

Selects the next step per-user from a candidate set - content, offer, feature nudge, human touch, or recommendation surface - based on a live AI attribute, replacing fixed cadences with adaptive paths.

How you prompt this

Describe the outcome in your own words and Blu Agent maps it to this recipe, or run it by name. It drafts a plan first and nothing runs until you approve it.

Tell Blu Agent what you want

“A rules-based journey that routes each person to a next step from a candidate set using attributes and segment membership at branch points.”

recipe.md preview

Built from the published catalog. Blu Agent's full step prompts load in your workspace when you run it.

Sign up to read the full recipe.md and run it with Blu Agent.

Sign up to see the full recipe

Inputs & outputs

What the recipe needs before it runs, and what it leaves in your workspace.

Inputs

  1. Nothing declared beyond an Intempt workspace. Blu asks for anything else it needs as it goes.

Outputs

    How the recipe runs

    Blu follows the steps in order and shows you each one before anything goes live.

    1. 01Decide the next step per userCreate an AI-derived attribute 'next_best_action' on the User object, refreshed daily and on every significant behavioral event. Inputs: current lifecycle stage, engagement velocity (sessions trend), feature usage breadth, recent intent signals (pricing visit, feature first-click, support touch), and prior message engagement. Outputs: a structured object with (a) recommended_action one of [educate / nurture / offer / surface-feature / surface-recommendations / handoff-to-agent / wait]; (b) recommended_channel [email / sms / in_app / push / slack-internal]; (c) recommended_content_theme; (d) confidence (0-100); (e) expected_uplift_signal. The attribute is the decisioning brain: the journey routes from it at each gate.
    2. 02Only route confident callsBuild a segment 'NBA-orchestrated users' capturing active users where next_best_action.confidence >= 50 AND user is not currently in another high-priority journey (no double-orchestration). Refreshed continuously, as users' NBA changes, they move between sub-cohorts of this parent segment. Use the result of "Decide the next step per user".
    3. 03Write an email per actionGenerate email content variants per recommended_action: (a) educate variant (short value tip matched to the user's stage; (b) nurture variant) case study/customer story relevant to user's segment; (c) offer variant (personalized incentive (discount / trial extension / credit) matched to plan and tenure; (d) surface-feature variant) feature spotlight with deep-link to the in-app destination. Each variant pulls dynamic content blocks based on the user's full profile, not just first name. Use the result of "Decide the next step per user", "Only route confident calls".
    4. 04Write the in app versionsThis step builds a page. Generate in-app message variants for the surface-feature and surface-recommendations branches. Surface-feature: contextual tooltip pointing to the feature, with a 1-line value prop and 'try it' CTA, rendered in the relevant in-app location. Surface-recommendations: a personalized recommendation card pulling from the user's product affinity. In-app fires only during active sessions, never outside the app (avoids interruption fatigue). Use the result of "Decide the next step per user", "Only route confident calls", "Write an email per action".
    5. 05Build the recommendation setConfigure a recommendation surface 'NBA-driven product/content recs' that activates when next_best_action.recommended_action = surface-recommendations. Sources: the user's prior interaction history + similar-user purchase/usage patterns + currently-trending items in their segment. Render in: in-app sidebar + email content block + page personalization slot. Refresh weekly. Use the result of "Decide the next step per user", "Only route confident calls".
    6. 06Re-read the signal at each gateBuild an adaptive journey wired to the NBA segment. At each gate (signup+7d, +14d, +30d, +60d, ongoing weekly), the journey reads next_best_action attribute and routes the user down the matching branch: educate to email variant a; nurture to email variant b; offer to email variant c with discount; surface-feature to in-app tooltip + email variant d; surface-recommendations to activate recommendation surface + email digest with recs; handoff-to-agent to trigger agent conversation; wait to skip touch, recompute next gate. Each branch's outcome (clicked / engaged / converted / ignored) feeds back into the next NBA computation so the decisioning learns. Exit on: conversion event (Deal created / Subscription started / activation_milestone: depending on lifecycle), unsubscribe, or sustained no-engagement (NBA returns wait 3 gates in a row to suppress). Use the result of "Decide the next step per user", "Only route confident calls", "Write an email per action", "Write the in app versions", "Build the recommendation set".
    7. 07Check the model against a controlCompose an NBA orchestration dashboard: distribution of recommended_action across users (which actions does the model favor: sanity check on model balance), per-branch engagement rates (which actions actually convert), confidence-vs-conversion correlation (does higher-confidence routing actually predict higher conversion (model-quality signal), uplift vs control (a 5-10% holdout that gets fixed cadence) the proof-of-value chart), and per-segment NBA quality (model may serve some segments better than others: informs retraining priorities). Use the result of "Decide the next step per user", "Only route confident calls", "Write an email per action", "Write the in app versions", "Build the recommendation set", "Re-read the signal at each gate".

    Install Intempt and run it

    1. 1

      Sign up

      Create a free Intempt workspace.

    2. 2

      Connect your data

      Add the sources this recipe reads from.

    3. 3

      Ask Blu

      Type /next-best-action-orchestration in Blu once this recipe is live there.

    4. 4

      Review and approve

      Blu builds each step and waits for your OK before anything goes live.

    Related recipes

    Intempt TechnologiesDiscovery call qualification fieldsGenerates an AI summary for Discovery calls from meeting data using the qualification framework in the summary recipe.View recipe Intempt TechnologiesPost demo nurtureKeeps a demo warm for 90 days with a playbook, a case study, an ROI calculator and a customer story, and pulls the AE in early if they bite.View recipe Intempt TechnologiesReturn rate by categoryShows which product categories get returned most, and which ones have got worse since last quarter.View recipe

    More recipes by Intempt Technologies

    Discovery call qualification fieldsGenerates an AI summary for Discovery calls from meeting data using the qualification framework in the summary recipe.View recipe Post demo nurtureKeeps a demo warm for 90 days with a playbook, a case study, an ROI calculator and a customer story, and pulls the AE in early if they bite.View recipe Product swap in a sceneReplaces the product in a shot you already have with a different one, keeping the scene, lighting and shadow the same.View recipe

    Got a play worth sharing?

    Publish it to the Intempt Collective Marketplace and earn when other teams run it.

    Submit a recipe