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
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.
Which accounts? Should I look for is_poc=true on your events, or do you have a specific list?
List: acct_megacorp, acct_innovate, acct_scaleco, acct_bigfish. Trial started 12 days ago for all of them.
Pulling daily event volume per account for the last 12 days.
agentry_analytics_query {
project_id: "default",
hogql: "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"
}{
"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 }
]
}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.
Want me to publish this as a dashboard you can share with the AEs and check daily?
agentry_publish_query {
recipe_id: "poc_engagement_tracker",
description: "Active POC Engagement (daily)",
params: { account_ids: ["acct_megacorp", "acct_innovate", "acct_scaleco", "acct_bigfish"], window_days: 14 }
}{ "publication_id": "pub_poc_1", "public_url": "https://agentry.sh/q/pub_poc_1?key=agp_…" }Dashboard live. Pin it in your sales channel.
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
Public dashboard:
https://agentry.sh/q/pub_poc_1?key=agp_xxx
↳ Auto-refreshes; share with the sales teamSetting 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.
// When you provision a POC, fire a marker event:
await fetch(`https://api.agentry.sh/v1/analytics/${PROJECT_ID}/`, {
method: "POST",
headers: {
"Authorization": `Bearer ${process.env.AGENTRY_DSN}`,
"Content-Type": "application/json",
"User-Agent": "myapp-sales/1.0", // REQUIRED — Cloudflare 403s default UAs
},
body: JSON.stringify({
event: "poc_started",
distinct_id: account.id,
properties: {
account_id: account.id,
account_name: account.name,
poc_owner_ae: "alex@us",
poc_duration_days: 14,
},
}),
});
// And on every per-user event from that account, tag is_poc:
await fetch(`https://api.agentry.sh/v1/analytics/${PROJECT_ID}/`, {
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?”
- “Daily cron: at 9am, post any POC with no activity yesterday to #sales-pocs.” (uses a Routine)
- “Send the dashboard URL via email to each AE every Monday at 8am with their accounts highlighted.”