Customer Scenarios
How operators actually run Stratam
Four detailed walkthroughs from beta operators —
the persona, the workflow, the actual monitor setup, the
measurable outcome.
The pain
"I'd wake up to 60 emails. Half were 'thanks!' replies, 10
were genuine support questions, 5 were urgent. I was spending
my first hour every morning sorting before I could do any
actual work."
What they wired up
One Gmail OAuth connection. One recurring monitor that fires
every weekday morning. Vault preloaded with their pricing,
common support answers, and a "do not autoreply to" allowlist
of investors + key customers.
Monitor: Morning inbox triage
Schedule: 6:30 AM ET weekdays (cron 30 11 * * 1-5)
Notify: Discord DM
Instruction:
Pull unread emails from the last 12 hours. For each:
- If from do_not_autoreply list: leave alone, flag in summary.
- If a clear support question matching the vault FAQ: draft a
reply, save to drafts (do not send).
- If a "thanks!" / acknowledgment: archive.
- If urgent (mentions outage, refund, legal): flag in summary
with full preview.
Then DM me a 5-line summary: urgent count, drafts saved, archived
count, anything that needed a judgment call.
The outcome
By 7 AM they open Discord, read 5 lines, and know what
needs them. They review the saved drafts at lunch and send
80% of them as-written. The "anything that needed a
judgment call" line surfaces ~2 emails a day — the only
ones that actually want their attention.
Self-reported: "Got my mornings back. The
operator paid for itself in the first week of saved time."
The pain
"Our top three competitors change pricing without
announcing. We'd find out three weeks in when a prospect
casually mentioned the new number. We'd lose deals because
our pitch was stale."
What they wired up
Three monitors, one per competitor pricing page.
browser_visit renders the page, agent extracts
the price table, compares to last week's snapshot from the
vault, only pings on a material change.
Monitor: CompetitorX pricing watch
Schedule: daily 9:30 AM ET (cron 30 14 * * *)
Notify: #sales Slack channel (via webhook)
Instruction:
Visit competitor-x.com/pricing. Extract the public tier table
(name, monthly price, key features). Recall the last snapshot
saved under tag pricing:competitor-x. Compare.
If anything changed materially (price, tier name, feature
added/removed): ping with a diff. Otherwise stay silent.
Save the new snapshot via remember, tag
pricing:competitor-x,snapshot, content
is the structured table.
The outcome
Three months in they've caught two competitor price hikes
and one new tier launch within hours of going live. Sales
team updates the battlecard same day. One deal saved
explicitly attributed to "we knew about the new tier and
had a counter ready."
Self-reported: "Cheaper than any
competitive intel SaaS we evaluated. Faster too."
The pain
"Every morning: open CoinGecko, open Yahoo Finance, open
Bloomberg, open my brokerage, open Twitter. 25 minutes
before I'd even seen what mattered. Half of that was
scrolling past stuff that didn't move."
What they wired up
One pre-market monitor + a vault preloaded with their
watchlist tickers and price targets. Uses
web_search + http_fetch for live
prices, recall against their notes.
Monitor: Pre-market brief
Schedule: 6:00 AM ET weekdays
Notify: Discord DM
Instruction:
Watchlist tickers (recall tag watchlist):
pull current price + overnight change for each.
Then web_search for last 12h news on each ticker. Summarize
any catalyst-level news (earnings, M&A, regulatory, major
filings) in 1 line per ticker.
Crypto watchlist (recall crypto-watchlist):
same drill — price + overnight + any major news.
Macro: 1 line on overnight S&P futures, dollar, gold, oil.
Format: Discord-readable, under 25 lines total. Bold any
ticker that moved >3% overnight or has catalyst-level news.
The outcome
One Discord message at 6 AM, read in 90 seconds, sized
for their thumb. The ticker bolding does the prioritization
— they go deep on the 2-3 bolded names, skim the rest.
Self-reported: "I'd pay $25/mo just for
the morning brief. Everything else Stratam does is bonus."
The pain
"We had PagerDuty for incidents, Slack webhooks for
deploys, and a separate uptime tool for the API. None of
them talked to each other. When something broke, we'd
spend 10 minutes figuring out which tool was screaming."
What they wired up
One org workspace with all 5 engineers. Two inbound
webhooks (from CI + uptime monitor) routed to Stratam,
plus a recurring monitor checking the API health endpoint.
Stratam correlates events: a deploy followed by an uptime
alert gets flagged as "likely deploy-related."
Webhook: ci-deploy-finished
Source: GitHub Actions success hook
Behavior: remember(tag=deploy, content=commit_sha + msg)
Webhook: uptime-alert
Source: uptime monitor (PagerDuty-style)
Behavior:
recall(tag=deploy, since=10min)
if a deploy happened in the last 10min:
post to #eng "uptime alert correlates with deploy
{sha} ({minutes_ago}min ago) — likely cause"
else:
post to #eng "uptime alert — no recent deploy,
investigating infra"
Monitor: daily incident summary
Schedule: 9:00 AM ET weekdays
Pull every "uptime alert" + "deploy" event in last 24h,
summarize for standup.
The outcome
The "likely deploy-related" correlation is the win — the
on-call engineer knows within 30 seconds whether to roll
back or page infra. They retired the PagerDuty seat and
the standalone uptime SaaS.
Self-reported: "Net savings of $180/mo on
retired tools. Stratam org tier costs $75/mo for 5 of us.
That math just keeps getting better."