Deployment
Guide
Your audit identified the revenue leaks. Now deploy the agent.
Four categories. One 90-day program each.
Getting started
Decisions before you start
You'll need to make several key decisions that will shape your deployment:
- Which category? AI Pipegen, Lifecycle Revenue, Revenue Attribution, or AI Creative Ops
- How many initial users? (White-glove setup covers up to 5)
- What's your timeline and baseline metrics?
- What systems need to be connected? (inbox, calendar, pipeline, webstore, payments, ad platforms)
This guide walks through each category's 90-day deployment. $5,000/month. Platform access, white-glove setup, agent configuration, and measurement included.
AI Pipegen
Sales tools record. Agents execute.
Clay enriches. Outreach sequences. Gong records. Apollo prospects. HubSpot logs. Five tools, zero connection from intent to close. One agent runs the full loop on first-party data.
01. Meet your deal execution agent
Reps juggle 5 disconnected tools. Each sees one signal. The rep is the integration layer. This program installs an agent that works wherever the rep works:
- Triages inbox by deal intent and value - acts on signals the moment they arrive
- Drafts replies in your voice from the full account context (saved to Drafts)
- Schedules inside threads - no calendar ping-pong, pre-call briefs auto-delivered
- Captures meetings into 60-second summaries, follow-up drafts, and next-step tasks
- Syncs deal state from call + email + product signals - pipeline reflects reality without rep data entry
This isn't a trial. It's a 90-day deployment.
02. What we deliver
Deal execution on desktop, mobile, and web
A/ Integration + setup (white-glove for 5 reps)
White-glove setup for up to 5 reps: inbox, calendar, pipeline, qualification rules, and tone calibration.
B/ Inbox → qualification → scheduling
Triage by intent, draft contextual replies, schedule inside threads. No manual back-and-forth.
C/ Meetings → follow-through
Auto-generated summaries, follow-up drafts, next-step tasks, all synced automatically.
D/ Deal velocity + pipeline accuracy
Stalled deal detection, auto-logged activity, and pipeline stages that reflect reality. Reps sell; Blu keeps the system honest.
03. How it works
| Step | Action | What happens |
|---|---|---|
| 1 | Connect inbox + calendar + pipeline | Threads map to deals + contacts across devices |
| 2 | Lock qualification rules | 'Ready for meeting' criteria set |
| 3 | Run daily execution | Triage, draft, schedule inside threads |
| 4 | Capture meetings | Summary + follow-up + tasks created |
| 5 | Measure deal velocity | Time-to-close and meeting-to-opp conversion vs baseline |
Lifecycle Revenue
Point tools automated channels. The Lifecycle Marketer automates the lifecycle.
Optimizely tests pages but can't email. Klaviyo emails but can't test pages. Nobody closes the loop. One agent experiments the page, orchestrates the journey, and attributes the dollar.
01. Meet your lifecycle agent
Marketing stacks split lifecycle across 4 tools. Each bolted on AI, but the AI only sees what that tool sees. This program installs an agent that orchestrates the full lifecycle:
- Updates lifecycle stages automatically (New → Activated → Engaged → At-Risk → High-Value) from behavioral + payment signals
- Triggers journeys by lifecycle stage, not channel - onboarding for New, repeat-purchase for Engaged, winback for At-Risk
- Runs lifecycle-targeted experiments - variants that know who they're testing, not blind A/B splits. CUPED variance reduction gets reliable winners in days
- Promotes winners to personalization rules automatically - experiment feeds the journey, journey data informs the next experiment
- Measures impact on activation, repeat purchase, retention, and LTV - not just opens and clicks
This isn't a trial. It's a 90-day deployment.
02. What we deliver
Lifecycle + journeys + experiments + personalization included
A/ Behavioral data foundation + lifecycle design
Webstore + payments connected, identity resolved, lifecycle stages defined (New → Activated → Engaged → At-Risk → High-Value).
B/ Stage-triggered journey orchestration (4-6 core flows)
Onboarding, repeat purchase, winback, at-risk intervention - email/SMS/push + on-site, triggered by lifecycle stage.
C/ Lifecycle-targeted experiments (3-5 tests)
Experiments that know who they're testing. CUPED-powered, AI-generated creative. Winners promoted to personalization rules.
D/ Lifecycle measurement
Cohort-based reporting on activation, retention, repeat purchase, and LTV direction. Before-and-after by lifecycle stage.
03. How it works
| Step | Action | What happens |
|---|---|---|
| 1 | Connect webstore + payments + tracking | All behavioral + revenue signals unified in CDP |
| 2 | Define lifecycle stages + segments | The Lifecycle Marketer auto-assigns customers to stages from live signals |
| 3 | Build stage-triggered journeys + on-site experiences | Flows, templates, personalization rules shipped |
| 4 | Run lifecycle-targeted experiments with CUPED | Reliable winners in days, promoted to personalization rules |
| 5 | Measure lifecycle progression | Cohorts compared vs baseline across stages |
Revenue Attribution
Analytics tools show charts. The Data Analyst gives answers.
Not a dashboard you stare at. Not a BI tool you query. An agent that answers from web, revenue, product, and attribution data on one identity graph.
01. Meet your analytics agent
Teams run 5 analytics tools that can't see each other. Each added NL querying, but an AI that only sees one slice gives confident wrong answers. This program installs an agent that sees the full picture:
- Unifies web, revenue, product, and attribution analytics on one identity graph - same numbers everywhere
- Connects ad spend to actual MRR (not form fills) with multi-touch attribution across the full journey
- Tracks funnel conversion with clickable drop-offs - see the users behind every number
- Scores every account for churn risk and LTV from behavioral + payment + product signals combined
- Surfaces anomalies and flags at-risk accounts before they churn - proactive, not reactive
This isn't a trial. It's a 90-day deployment.
02. What we deliver
Replaces GA + Mixpanel + ChartMogul + Triple Whale + Gainsight
A/ Data foundation + unified identity graph
SDK installed, Stripe + webstore + CRM + ad platforms connected. One identity graph resolving across devices, sessions, and channels.
B/ Web + revenue + attribution dashboards
Traffic, revenue, subscription, and multi-touch attribution from one source of truth. Same numbers everywhere.
C/ Funnel + retention views (user-level, clickable)
Core conversion funnels built, retention heatmaps configured. Every drop-off step clickable to the user list behind it.
D/ Predictive scoring + Data Analyst alerts
Churn probability and LTV models from combined signals. At-risk accounts auto-flagged. Anomalies surfaced proactively.
03. How it works
| Step | Action | What happens |
|---|---|---|
| 1 | Connect Stripe + webstore + CRM + ad platforms | One identity graph across all signals |
| 2 | Configure unified dashboards | Web, revenue, attribution, subscription - same numbers everywhere |
| 3 | Build conversion funnels | Every drop-off clickable to the users who left |
| 4 | Enable predictive models | Churn + LTV scored from combined behavioral + revenue signals |
| 5 | Set up Data Analyst alerts | Anomalies surfaced, at-risk accounts flagged before they churn |
AI Creative Ops
Creative tools make one thing. The Creative Designer produces everything.
Pletor knows your brand. Marpipe knows your catalog. Neither knows your customer. Three brains - brand kit, customer data, product catalog - every asset personalized to who's seeing it.
01. Meet your creative production agent
Creative production is split across 5 tools for 5 surfaces. Each handles one. Nobody handles all five from one product feed. This program installs an agent that produces everything:
- Extracts your brand from your URL and enforces it across every surface - catalog, paid, owned, site, channels
- Generates pack shots, lifestyle, on-model, and video from your product feed at catalog scale
- Produces catalog ads (Meta, TikTok, Google, Pinterest), email/SMS/push templates, site modules, and landing pages
- Formats for marketplace compliance (Amazon, Etsy) and refreshes automatically as products change
This isn't a trial. It's a 90-day deployment.
02. What we deliver
Brand Kit + all five surfaces included
A/ Brand Kit + product feed foundation
Brand extracted from URL, product catalog connected, generation rules configured, channel specs locked.
B/ Catalog creative (product photography + video)
Pack shots, lifestyle, on-model, and product video for up to 200 SKUs - print-resolution, multi-angle, scene-controlled.
C/ Paid + owned + site (ads, email, SMS, push, landing pages)
Catalog ads for every platform, message templates, site modules, and matched landing pages - all from one brief.
D/ Marketplace + ongoing refresh
Amazon/Etsy format-compliant listings. Auto-refresh as products change. Performance variants biased toward winners.
03. How it works
| Step | Action | What happens |
|---|---|---|
| 1 | Brand Kit + product feed connected | Brand governs every surface, SKU data ready |
| 2 | Generate catalog creative | Photos + video for up to 200 SKUs |
| 3 | Build paid + owned + site assets | Ads, email, SMS, push, landing pages from one brief |
| 4 | Format for marketplaces | Amazon, Etsy listings compliant and published |
| 5 | Measure production velocity | Cost-per-asset and time-to-market vs baseline |
After 90 days
Choose your model
At the end of the 90-day deployment, you have two options. Both are available for every category.
Tech-only
Keep the platform and run it internally. Lower cost. Your team owns execution, and you retain all the configuration, journeys, experiments, and integrations built during the deployment.
Continue month-to-month
We keep optimizing and expanding workflows. No long-term contract. Useful when you want to deploy additional categories, increase user count, or maintain the managed execution model.
Not quite ready for deployment?
Learn how to hone your Intempt pitch or run a pilot to strengthen your case.