Power user pattern detection
Detects users with daily or batch patterns of repeated manual work: 5+ same task in 7 days, one-at-a-time batch operations, or repeat exports.
Authored by Intempt Technologies
Recipe curators
Agentic workflow
Coming soon7 steps Blu runs end to end. It drafts a plan and nothing runs until you approve it.
- 1
Spot work done the hard way
Coming soonBlu builds a step
- 2
Find who could automate it
Coming soonBlu builds a step
- 3
Nudge them mid task
Coming soonBlu builds a step
- 4
List their power features
Coming soonBlu builds a step
- 5
Ask adopters for a review
Coming soonBlu builds a step
- 6
Nudge, remind, then let it go
Coming soonBlu builds a step
- 7
See which features stay hidden
Coming soonBlu builds a step
About this recipe
When to use it
Someone doing the same task five times a week, or exporting the same report repeatedly by hand, is telling you exactly which feature they need - they just don't know it exists. This detects the pattern and surfaces the fix.
What it delivers
Detects users repeatedly performing manual workflows the product can automate - the same task 5+ times in 7 days, batch operations done one at a time, repeat exports - surfaces the relevant power feature contextually, then invites them to an advocacy program if adopted.
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
“Detects users with daily or batch patterns of repeated manual work: 5+ same task in 7 days, one-at-a-time batch operations, or repeat exports.”
recipe.md preview
Built from the published catalog. Blu Agent's full step prompts load in your workspace when you run it.
Inputs & outputs
What the recipe needs before it runs, and what it leaves in your workspace.
Inputs
- 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.
- 01Spot work done the hard wayCreate an AI-derived attribute 'detected_manual_patterns' on the User object, refreshed daily. Detects repeated manual sequences that have automated counterparts in the product. Example patterns: (a) user runs same multi-step report 5+ times in 7 days (has unused 'scheduled reports' feature), (b) user exports data manually 3+ times in 7 days (has unused API/webhook feature), (c) user assigns same task type repeatedly (has unused task templates feature), (d) user filters dashboard same way 10+ times in 14 days (has unused saved-view feature). Output: list of detected patterns with the automate-it feature name and adoption-likelihood score (based on user's plan, skill level, prior automation adoption).
- 02Find who could automate itBuild a segment 'Manual-pattern detected' capturing paying users where detected_manual_patterns is non-empty AND the user hasn't yet used the recommended automation feature. Partitioned by feature-to-introduce. Excludes users who have dismissed feature-recommendations 3+ times (respect the no) and users with plan limits that exclude the suggested feature. Use the result of "Spot work done the hard way".
- 03Nudge them mid taskThis step builds a page. Generate in-app nudge content per detected pattern. Format: tooltip or floating card that appears WHEN the user is mid-pattern (e.g. on their 6th manual export). Content: 'You've done this 6 times this week: did you know [Product] can automate this?' + 60-second 'how it works' GIF + 'enable now' CTA + dismiss option. Renders at the exact moment of friction, not in a generic feature-discovery surface. The contextual timing is the magic: this is the 4x feature-adoption uplift pattern from the research. Use the result of "Spot work done the hard way", "Find who could automate it".
- 04List their power featuresConfigure a recommendation surface 'Power features for your workflow' on the user's dashboard. Pulls: the top 3 detected_manual_patterns with their automate-it counterparts, ranked by adoption-likelihood. Each recommendation: 1-line description + 'try it' deep link. Updates when user adopts a feature (rotates in the next-best). Renders persistently for 30 days after detection. Use the result of "Spot work done the hard way", "Find who could automate it".
- 05Ask adopters for a reviewGenerate follow-up email content sent 14 days after a user adopts a recommended automation feature. Content: 'Congrats on automating [feature]: you've saved an estimated [time] per week. Mind sharing your experience?' Two CTAs: write a review on G2/Capterra (with deep-link to the right product page), refer a peer (with referral program info). This is where power-user-detection becomes an advocacy pipeline: the user just had a positive experience, the moment is warm. Use the result of "Spot work done the hard way", "Find who could automate it".
- 06Nudge, remind, then let it goBuild a journey wired to manual-pattern-detected segment. Touch 1 (real-time, on the 5th+ pattern repetition mid-session): in-app contextual nudge fires. Recommendation surface activates persistently. Touch 2 (Day 3, if user didn't dismiss): email reinforcement with the same feature pitch + a customer story of someone who automated it. Touch 3 (Day 14, IF user has adopted the feature): advocacy follow-up email asking for review/referral. Touch 4 (Day 14, IF user has NOT adopted): exit gracefully: pattern stays detected, surface stays active, but stop pushing. Exit on: feature adoption + advocacy ask sent, explicit dismiss, or 30-day timeout. Use the result of "Spot work done the hard way", "Find who could automate it", "Nudge them mid task", "List their power features", "Ask adopters for a review".
- 07See which features stay hiddenCompose a power-user detection dashboard: top 10 detected manual patterns by frequency (which features have the biggest discoverability gap (informs in-app UI redesign priorities), in-app-nudge-to-adoption rate (the headline metric) target: 15%+, baseline generic feature emails get 4-6%), 30-day retention of adopters vs. non-adopters (proves the journey's value beyond direct adoption), and advocacy-ask conversion (% of adopters who write a review or refer: this is the unexpected revenue side-effect of feature discovery). Use the result of "Spot work done the hard way", "Find who could automate it", "Nudge them mid task", "List their power features", "Ask adopters for a review", "Nudge, remind, then let it go".
Install Intempt and run it
- 1
Sign up
Create a free Intempt workspace.
- 2
Connect your data
Add the sources this recipe reads from.
- 3
Ask Blu
Type /power-user-pattern-detection in Blu once this recipe is live there.
- 4
Review and approve
Blu builds each step and waits for your OK before anything goes live.
