What are Signals?
Signals are AI-powered monitors that automatically analyze data from your connected tools and the web, then deliver structured insights on a schedule you define. Each Signal focuses on a specific question or area you want to stay on top of.
Think of a Signal as a persistent, intelligent monitor — you define what to watch, and Samepage handles the analysis and delivery.
How Signals Work
Each Signal has five components:
Title — A descriptive name for what the Signal monitors (e.g., "Customer Feedback & Sentiment")
Data Sources — Which connected tools and/or web sources to analyze (e.g., Slack + Jira, or Gong + HubSpot)
Instructions — Natural language description of what to look for and how to present the findings (e.g., "Summarize recurring feature requests from customer calls this week, grouped by theme")
Schedule — How often the Signal runs (e.g., daily, weekdays only, every Monday)
Time Range — What window of data to analyze (e.g., last 7 days, last 30 days, or auto)
When a Signal runs, Samepage pulls relevant data from the selected sources for the specified time range, combines it with context about you (your name, role, timezone, connected integrations, and prior Signal history), and runs your instructions through AI. The result is a structured output with insights, summaries, tables, and deep links back to the source data.
What Goes Into a Signal
Behind the scenes, Samepage assembles several inputs before generating a Signal:
Your data — The actual content from your selected integrations within the time range (messages, tickets, issues, meeting transcripts, CRM records, etc.)
Your profile — Your name, email, department, and timezone so the Signal can personalize its output
Your prompt — The instructions you wrote, which drive what the AI looks for and how it formats the result
Content preferences — Your chosen output length (concise, moderate, or detailed), style (bullets, mixed, or paragraphs), and tone (friendly, balanced, or direct)
Prior context — History from previous Signal runs, so the AI can identify what's changed
Magic - Samepage uses best practices, proven prompt strategies, feedback loops and other techniques to deliver useful content and avoid noise.
Single-Source vs. Multi-Source Signals
If your Signal uses a single data source, Samepage uses that integration's built-in analysis template — optimized prompts and structured output formats designed for that specific tool's data. For example, a Slack-only Signal produces a structured topic-and-highlights format, while a Google Calendar-only Signal produces an event summary.
If your Signal uses multiple data sources or a custom prompt, Samepage concatenates context from all sources and runs your prompt across the combined data. This is where cross-referencing gets powerful — you can ask Samepage to compare what customers are saying in Gong calls with what's in the Jira backlog, for instance.
Schedule Options
Samepage offers flexible scheduling so Signals run at the right cadence for the information they track:
Daily — Runs every day
Weekdays — Runs Monday through Friday
Weekly — Runs on specific days you choose (e.g., every Monday and Thursday)
Monthly — Runs on a specific day of the month (1–28)
Custom interval — Runs every N days, weeks, or months (e.g., every 2 weeks, every 3 days)
Signals are generated automatically on schedule. You can also manually trigger a Signal at any time using the Preview feature.
Time Range Options
The time range controls how far back the Signal looks when it runs. Options include:
Auto (default) — Samepage determines the best lookback based on the data source and schedule
Today or Yesterday — For daily operational Signals
This week, Last week, or Last 7 days — For weekly summaries
This month or Last 30 days — For monthly reviews
This quarter or Last quarter — For broader trend analysis
Custom — Specify an exact number of days, weeks, or months
Each data source in a Signal can have its own time range if needed.
What Signals Can Monitor
Signals can analyze data from any connected integration or from the web. Common categories include:
Product & Engineering:
Weekly product and engineering progress — what shipped, what's in progress, what's next
Risk radar — work items at risk, in progress too long without movement, blocked items
Bug tracking and triage — recurring bugs, unresolved issues, QA trends
Customer Intelligence:
Customer feedback synthesis — feature requests and pain points from Gong calls, support tickets, Slack
Customer sentiment trends — emerging themes from Intercom, Zendesk, or Slack conversations
Feature requests linked to revenue — attach contract value from CRM to each request
Competitive & Market:
Competitive tracking — monitor competitor product releases, pricing changes, news via web search
Market mentions — track AI B2B startups, industry news, funding rounds
Close-lost competitive intel — weekly rollup of lost deals grouped by competitor and loss reason
Operational:
Day planner — upcoming meetings, deadlines, and priorities for the day
Overdue responses — tasks assigned to you across Docs, Slack, and Notion that need follow-up
Communication triage — top 5 things across Slack and email you should pay attention to
Analytics:
Product usage trends — week-over-week metrics from Mixpanel or Pendo
Feature launch impact — measure how a release affected WAUs, MAUs, or other key metrics
Usage + backlog correlation — combine analytics data with Linear/Jira to recommend what to build next
Signal Outputs
Each Signal produces a block in your stream containing:
Summary — A concise overview (under 280 characters) of the key findings, shown at the top of the card
Detailed analysis — The full output: tables, breakdowns, grouped insights, and deeper analysis
Deep links — Samepage automatically includes clickable links back to the original source data (Jira tickets, Gong calls, CRM records, Zoom recordings, etc.) so you can jump directly to context. You don't need to ask for links in your prompt — they're included by default. Note: a few integrations (Slack, Granola, Mixpanel, and Pendo) don't provide source URLs, so links aren't available for those tools. See the supported integrations page for details.
Actionable recommendations — When relevant, Signals suggest next steps or flag items needing attention
Daily Summary
In addition to individual Signal outputs, Samepage generates a Daily Summary — a short rollup of everything across all your Signals for the day, focused on decisions, blockers, approvals, and follow-ups that need your attention. This appears at the top of your daily view.
Output Preferences
By default, Signals inherit your formatting preferences from your user profile — so if you've set your preferred length, style, and tone there, every new Signal will use those settings automatically. You can override these at the individual Signal level if a particular Signal needs a different format:
Length — Concise (quick scan), moderate (default), or detailed (comprehensive analysis)
Style — Bullet points, mixed (bullets + prose), or paragraphs
Tone — Friendly, balanced, or direct
Because these preferences are already applied, you don't need to include formatting instructions in your prompt (e.g., "use bullet points" or "keep it concise"). Focus your prompt on what you want to know, and Samepage handles the how.
Signal Examples
Signal | Data Sources | What It Does |
Weekly Product & Engineering Progress | Linear, GitHub | Summarizes what shipped, what's in progress, and what's coming next |
Risk Radar: Slips, Stalls & Blockers | Jira/Linear, GitHub | Finds work items in progress >5 days without movement, failing CI, blocked items |
Customer Feedback → Feature Requests | Gong, Slack, Intercom | Synthesizes feature opportunities from calls, messages, and support conversations |
Competitive Landscape | Web search | Tracks competitor product releases, pricing changes, and news |
Close-Lost Competitive Intel | Salesforce/HubSpot/Zoho | Weekly rollup of lost opportunities grouped by competitor and loss reason |
Service Line-Item Trends | Zoho CRM | Which service add-ons are attaching to deals and which aren't |
V2 Impact on WAUs | Mixpanel/Pendo | Measures product launch impact using analytics data |
Meeting Activity Summary | Zoom, Granola, Gong | Key themes, action items, decisions, and customer feedback from recent meetings |
Overdue Responses I Owe Someone | Google Docs, Slack, Notion | Surfaces tasks assigned to you with suggested responses |
Creating a Signal
There are two ways to create a Signal:
Use a Suggested Signal
Samepage generates personalized Signal suggestions based on your connected data. When you create a new Signal, suggestions appear automatically. Each suggestion includes a pre-configured title, data sources, prompt, and time range — just click to preview and activate.
Suggestions are generated by analyzing your most recent data (last 48 hours) across all connected integrations and recommending Signals that would surface useful patterns. Samepage always includes at least one web-search-only suggestion for competitive or market monitoring. Suggestions refresh regularly so you always see relevant options.
Build from Scratch
Go to your stream and click to create a new Signal
Select your data sources — pick one or more connected integrations and/or enable web search
Write your instructions — describe what to look for and how to organize the output
Set a schedule — choose how often the Signal should run
Set the time range — choose how far back the Signal should look
Preview — click Preview to see a sample output based on current data
Activate — if the preview looks good, save and activate the Signal
Writing Good Instructions
The instructions you write are the single biggest factor in Signal quality. Tips:
Be specific about what to find. "Tell me about bugs" is vague. "Surface the top 5 recurring bugs from Zendesk that have been reported more than twice this week and aren't yet in Linear" is actionable.
Tell it how to organize the results. You don't need to specify formatting (bullets, tone, length) — that's handled by your profile preferences. But you should tell the Signal how to structure the content: "group by theme," "rank by frequency," "include a table with columns for issue, status, and assignee."
Don't worry about asking for links. Samepage includes deep links to source data by default wherever the integration supports them. Focus your prompt on what you want to find, not on requesting links.
Combine sources for cross-referencing. The most powerful Signals pull from multiple tools. "Compare feature requests from Gong calls with what's currently in the Linear backlog" or "Cross-reference Zendesk ticket themes with Jira issue status" surface insights humans would miss.
Start narrow, then expand. Begin with a focused scope (one team, one channel, one project) and expand once you're confident the Signal is useful.
Example Prompt
Here's a detailed prompt you can use as a starting point for a cross-functional weekly digest:
Review all activity from the past week and produce a structured summary organized as follows:
What Shipped — List features, fixes, and improvements that were completed or merged this week. Include links to the PRs or issues.
What's In Progress — Summarize the key work items currently being worked on. Flag anything that's been in progress for more than 5 days without movement.
Blocked or At Risk — Call out any work items that are blocked, stalled, or at risk of missing a deadline. Include enough context that I can decide whether to intervene.
Customer Signal — Surface any feature requests, complaints, or feedback that came through support or sales conversations this week. Group by theme if there are multiple.
Decisions & Follow-Ups — Highlight any decisions made in meetings or async discussions that affect product direction, and any open follow-ups that haven't been addressed.
Customizing Signals
Users can fully customize their Signals at any time:
Change data sources — Add or remove which tools the Signal analyzes
Edit instructions — Modify the natural language prompt that drives the analysis
Adjust schedule — Change how often the Signal runs
Set time range — Control how far back the Signal looks when analyzing data
Filter scope — Narrow the Signal to specific Slack channels, Jira projects, teams, or other filters depending on the integration
Duplicate — Copy an existing Signal's settings to create a variation
Pause/resume — Toggle a Signal inactive without deleting it
Product Analytics Signals (Mixpanel & Pendo)
Signals that use Mixpanel or Pendo work differently from other integrations. Instead of just writing a prompt and letting Samepage pull data, you're configuring a custom analytics report as part of the Signal setup. This is because analytics platforms require you to define what to measure and how to visualize it.
When you add Mixpanel or Pendo as a data source, the Signal form expands to include report configuration:
Mixpanel setup requires:
Project — Which Mixpanel project to query (required)
Event — Which event to measure (required, loaded from your Mixpanel data)
Chart type — Line chart, bar chart, table, or single metric
Grouping — How to bucket the data: by hour, day, week, or month
Count type — Total events or unique users
Cohort filter — Optionally limit to a specific Mixpanel cohort
Pendo setup requires:
Data source — Page Views or Feature Clicks
Page or Feature — Which specific page or feature to measure (required, loaded from your Pendo data)
Chart type — Line chart, bar chart, table, or single metric
Grouping — By day, week, or month
Count type — Total events or unique visitors
Once configured, the Signal runs the report, generates a chart visualization alongside the AI-written summary, and embeds both in the output. The AI summary interprets the chart data — calling out trends, anomalies, and week-over-week changes — so you get both the visual and the narrative.
This is different from most other integrations where setup is just "pick the source, write a prompt, go." With analytics Signals, the report configuration is what makes the output useful.
Web Signals
In addition to tool-based Signals, Samepage supports web Signals that monitor the public web. These don't require any integration — they use web search to track topics like competitors, industry news, market trends, and more. Web Signals are often the first Signal new users create during onboarding.
When web search is enabled, the Signal searches the public web for information matching your prompt and restricts results to your selected time range. You can combine web search with integration data — for example, a Signal that monitors competitor news via web search and cross-references it with close-lost reasons from your CRM.
How Signals Use AI
Signals use large language models to interpret raw data, identify patterns and changes, and generate structured outputs. Here's what to know:
Samepage uses enterprise-grade AI APIs to process your data
Customer data is never used to train AI models
Data is processed only to generate the Signal output
Each Signal run is independent — the AI doesn't retain data between runs (though it does reference prior Signal history for context on what's changed)
Signal outputs always include the AI's reasoning based on actual source data, not fabricated information. If there's no relevant data for a given run, the Signal will say so rather than guess.