How does WarnCast detect when a client relationship is at risk?
▾
WarnCast monitors five client signals from Gmail every 6 hours: how fast your client responds to emails, how often scope-change keywords appear, whether their message sentiment is shifting, whether senior stakeholders have gone quiet, and how frequently budget topics come up. Each signal gets a score from 0–100. The aggregate becomes your Health Score, updated automatically — no manual check-in required.
What's a Health Score and what do the numbers mean?
▾
The Health Score runs from 0 to 100. Green (70+) means the client relationship is stable. Yellow (40–69) means signals are drifting and you should pay attention. Red (below 40) means something is actively wrong and an alert fires immediately. On first connect, WarnCast scans the last 30 days and shows a provisional score labeled "Estimated — updating in 6h" so you see value the moment you log in.
What does WarnCast actually do when it fires an alert?
▾
When a Health Score drops below your threshold, WarnCast fires an alert that includes: the exact signal that triggered it, what changed and by how much, an AI-drafted follow-up message personalized to your client's communication style, and a recovery playbook based on what worked in similar situations. The free trial shows the score. The paid plan gives you the full alert intelligence — what triggered it, the draft, and the playbook.
Can I control how sensitive the alerts are?
▾
Yes. Every project has a sensitivity slider: Low, Medium (default), or High. You can also snooze individual alerts for 7 days, mark them as resolved, and review the full alert history per project. If you're on vacation or stepping away, the global alert pause feature lets you silence all alerts for up to 14 days — no noise, no FOMO.
What data does WarnCast read from our emails?
▾
We'll be straight with you: to score a relationship, WarnCast reads the actual emails — the words, the tone, the timing — because that's the only way to detect cooling sentiment, scope creep, or budget tension. What we don't do is keep them. The content is read in memory, the signal is computed, and the text is discarded in the same pass. We persist only the resulting numbers (0–100 per signal). There is no copy of your inbox sitting in our database. Access is read-only, and you can revoke it any time from your Google account.
Is client email content stored anywhere?
▾
No — and this is the promise we hold ourselves to. Your emails are pulled through Google's official read-only API, analyzed the moment they're read, and the text is thrown away as soon as the signal is computed. The only thing written to our database is the score itself — numbers between 0 and 100, plus timestamps. We never archive, index, or build a searchable copy of your inbox. You can revoke access at any time from your Google account settings, and the content was never ours to keep in the first place.
Is my data used to train AI models?
▾
No. Your project data and signal history are never used to train any AI model. Two of WarnCast's features — sentiment scoring and drafting a follow-up — do pass text through Anthropic's Claude API at the moment they run, because reading tone and writing a reply require the words. That processing is transient: nothing is stored by us afterward, and Anthropic does not train its models on data sent through the API. Benchmarks and recovery playbooks are built only from anonymized, aggregated signal numbers — never your client names or email content.
Is data shared with third parties?
▾
We don't sell or share your data, and we have no data-broker relationships — ever. We use two processors to run the product: Anthropic's Claude API scores sentiment and drafts follow-ups (text is processed transiently and never stored, and Anthropic doesn't train on API data), and Stripe handles payments. Aggregated benchmarks shared across the network are stripped of client names and email content — only the anonymized signal numbers remain.
How is my data encrypted and protected?
▾
All data is encrypted at rest and in transit. OAuth tokens are stored in encrypted database rows with row-level security — only your account can access your data. We use Google's official Gmail API with read-only scopes. You can revoke access at any time from your Google account settings without losing your signal history.
What integrations does WarnCast support right now?
▾
V0.1 is Gmail-first. Every agency has Gmail, Gmail alone covers the 5 most critical client signals, and setup takes under 5 minutes. Slack, Asana, and Google Calendar are V0.2 power-ups — they unlock the full 14-signal Health Score. You'll be prompted to connect them after your first Gmail score is live.