# Track engagement during enterprise POCs

> Daily engagement per account for active POCs. Flag any account that has gone quiet for 3+ days so AEs can act.

## Agent adaptation contract

- Canonical human page: https://agentry.sh/workflows/demo-poc-tracking
- Execution mode: on_demand
- Immutable automation template: none
- Applies to: b2b-saas, enterprise-sales
- Required example events: feature_used, implementation_milestone, demo_completed
- Required Agentry resources: none declared
- Do not use when:
  - Do not use until the example events are mapped to observed project signals, the current onboarding state is verified, and live event/property reads prove the required data is present.
  - Do not recommend this when the product has no stable account or workspace identity.
  - Do not present illustrative custom analysis as a built-in query blueprint or runnable Agentry product surface.
- Ask before using:
  - Which observed events map to feature_used, implementation_milestone, demo_completed? Is the current onboarding state verified, and do live event/property reads show non-synthetic traffic for them?
  - Which live properties provide feature_used.account_id, feature_used.is_poc, implementation_milestone.account_id, implementation_milestone.milestone, demo_completed.account_id, and which stable user or account identifier joins the signals?

This is an adaptable workflow example, not an API recipe. Map event and property names to the project's saved signal map, require status: "verified" from GET /v1/projects/:project_id/onboarding, and confirm the required signals through live event/property metadata and rows. Fetch current OpenAPI or query-blueprint details before making calls. Do not infer unattended authority from this page.

## Why this matters

Sales-led POCs are won by paying attention. The AE running the POC is on 6 other deals at the same time and can't manually check whether each prospect is logging in. A POC that goes quiet on day 4 of a 14-day trial is almost certainly lost — but only if someone notices on day 4, not day 14.

Salesforce has POC tracking; nobody fills in the daily fields. Hubspot has "engagement score" but it counts emails, not product usage. The hot signal — "is the prospect actually opening the tool you sold them on?" — lives in your event database.

Agentry's one prompt: per-POC daily activity, ranked by trend, with red-flag callouts on accounts that stopped. AE walks into Monday standup with the call list pre-prioritized.

## What you get

- A per-POC daily activity chart for the active trial window
- A "red flag" list: POCs with no activity in 3+ days
- A "green flag" list: POCs accelerating week-over-week (close-ready)

## Walk through it

**You:**
> I'm running 4 enterprise POCs. For each account, show daily event volume during the trial period. Flag any that have gone quiet for 3+ days.

**Agent:**
> Which accounts? Should I look for `is_poc=true` on your events, or do you have a specific list?

**You:**
> List: acct_megacorp, acct_innovate, acct_scaleco, acct_bigfish. Trial started 12 days ago for all of them.

**Agent:**
> Pulling daily event volume per account for the last 12 days.

```text
Endpoint: POST /v1/projects/:project_id/analytics/query
Exact shape: /v1/openapi.json?path=%2Fv1%2Fprojects%2F%7Bproject_id%7D%2Fanalytics%2Fquery&method=post
Custom HogQL goes in the OpenAPI-defined `query` field.
Concept fields:
  query: "SELECT properties.account_id AS account, toDate(timestamp) AS day, count() AS events, uniq(distinct_id) AS users FROM events WHERE properties.account_id IN ('acct_megacorp','acct_innovate','acct_scaleco','acct_bigfish') AND event != '$exception' AND timestamp >= now() - INTERVAL 12 DAY GROUP BY account, day ORDER BY account, day"
```
```text
{
  "rows": [
    { "account": "acct_megacorp",  "day": "2026-05-04", "events": 412, "users": 14 },
    { "account": "acct_megacorp",  "day": "2026-05-05", "events": 380, "users": 12 },
    { "account": "acct_megacorp",  "day": "2026-05-06", "events": 88,  "users": 4  },
    { "account": "acct_megacorp",  "day": "2026-05-07", "events": 12,  "users": 2  },
    { "account": "acct_megacorp",  "day": "2026-05-08", "events": 0,   "users": 0  },
    { "account": "acct_megacorp",  "day": "2026-05-09", "events": 0,   "users": 0  },
    { "account": "acct_megacorp",  "day": "2026-05-10", "events": 0,   "users": 0  },
    { "account": "acct_innovate", "day": "2026-05-04", "events": 88,  "users": 4 },
    { "account": "acct_innovate", "day": "2026-05-12", "events": 612, "users": 9 },
    { "account": "acct_innovate", "day": "2026-05-13", "events": 880, "users": 11 }
  ]
}
```

**Agent:**
> Mixed picture. Status of all four POCs:
>
> - **megacorp** — Day 1-3 strong (~400 events), Day 4-5 tapering, **Day 6-12 = silence**. RED FLAG. Almost certainly lost; AE should call today.
> - **innovate** — Slow start, **accelerating: 88 → 880 events/day**. GREEN FLAG. Schedule the close call.
> - **scaleco** — (no rows shown but the agent has them) — flat, low-volume throughout. Probably the champion left.
> - **bigfish** — Steady 200-300/day, ~6 users active. Healthy but not exciting.
>
> This is a custom account-scoped engagement query, not a built-in public-query blueprint. I can put it in a protected internal view for the sales team. If you want a browser-safe public URL, I should first add a real query blueprint for active POC engagement and then publish that blueprint.

## The output

```
Active POCs — Day 12 of 14

RED FLAGS (3+ days silent)
  megacorp     Day 6-12: 0 events / 0 users
               Started strong (400/day), dropped Day 4, dead since Day 5
               AE action: call champion TODAY, find out what stopped

GREEN FLAGS (accelerating)
  innovate     Day 1-7:  ~80-200 events/day (exploring)
               Day 8-12: 600-900 events/day (operationalizing)
               AE action: schedule close call this week

STEADY
  bigfish      ~250 events/day, 6 active users
               AE action: continue scheduled check-ins

NEEDS DIAGNOSIS
  scaleco      <50 events/day throughout, 1 user
               AE action: champion likely disengaged — re-onboard or pivot

Protected internal view:
  /admin/agentry/poc-health
  ↳ Server-side Agentry read for the sales team unless/until a real active-POC query blueprint is added.
```

## Setting it up

Same instrumentation as `account-health-score`: every event needs `account_id`. The only POC-specific addition is marking the trial start so you know which window matters.

```ts
// When you provision a POC, fire a marker event:
await fetch(`https://api.agentry.sh/v1/analytics/`, {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.AGENTRY_SERVER_API_KEY}`,
    "Content-Type": "application/json",
    "User-Agent": "myapp-sales/1.0",  // REQUIRED — Cloudflare 403s default UAs
  },
  body: JSON.stringify({
    event: "poc_started",
    distinct_id: account.id,
    properties: {
      account_id: account.id,
      account_name: account.name,
      poc_owner_ae: "alex@us",
      poc_duration_days: 14,
    },
  }),
});

// And on every per-user event from that account, tag is_poc:
await fetch(`https://api.agentry.sh/v1/analytics/`, {
  method: "POST",
  headers: { /* same */ "User-Agent": "myapp/1.0" },
  body: JSON.stringify({
    event: "feature_used",
    distinct_id: user.email,
    properties: {
      account_id: user.workspace_id,
      is_poc: account.is_poc ?? false,  // ← lets the agent filter cleanly
    },
  }),
});
```

If you don't want to instrument `is_poc`, you can also just pass the account_id list at query time (as in the conversation above).

## Variations

- *"For each POC, show which FEATURES they used most. Tells me what the demo should focus on at close."*
- *"Compare POC engagement curves of accounts that closed-won vs closed-lost last quarter. What patterns do winners show?"*
- *"On demand, list any POC with no activity yesterday and draft a #sales-pocs summary for review."*
- *"Prepare an AE-specific view with their accounts highlighted; let a human choose the recipients and sharing channel."*