--- title: "Agent | Cometly" description: "Chat with your marketing data. Cometly's Agent surfaces insights, optimizations, and reports — without you having to build another dashboard." source: "https://www.cometly.com/platform/agent" --- Platform / Agent # Ask anything. Get the answer Agent is grounded in your live attribution data, pipeline, CAC, LTV, journeys, and every event Cometly captures. Answers in seconds. No SQL, no analyst queue. Get started Which sources are actually driving closed-won revenue this month, and how does it shift across attribution models? Your top 4 sources drove $700K in closed-won ARR this month. Google Ads leads under Last Touch, but Google Organic and LinkedIn Ads climb under First Touch. Sales data by source · this month Booked demo → Closed-won ARR | Source | Booked Demo | SQL | Deal Created | Closed Won | | --- | --- | --- | --- | --- | | Summary | 1,500 +18.6% ↑ | 860 +14.2% ↑ | 305 +9.4% ↑ | $700K +24.6% ↑ | | Google Ads | 420 +22.4% ↑ | 241 +18.1% ↑ | 85 +12.7% ↑ | $196K +31.2% ↑ | | Direct | 270 +3.2% ↑ | 155 +2.6% ↑ | 55 +1.1% ↑ | $126K +4.5% ↑ | | LinkedIn Ads | 210 +14.1% ↑ | 120 +11.5% ↑ | 43 +8.2% ↑ | $98K +12.8% ↑ | | Email | 210 +12.4% ↑ | 120 +9.1% ↑ | 43 +5.7% ↑ | $98K +8.3% ↑ | | SMS | 143 +6.1% ↑ | 82 +3.4% ↑ | 29 −1.2% ↓ | $67K +2.8% ↑ | | Meta Ads | 120 +9.7% ↑ | 69 +6.3% ↑ | 24 +3.1% ↑ | $56K +4.2% ↑ | Sources Stripe HubSpot Meta Ads Google Ads LinkedIn Ads Save as report Share Which ad campaigns are my top performers right now? Across Meta and Google, 5 campaigns are driving most of the pipeline. Free Trial — Founder LAL leads at 6.4× ROAS and $184K closed-won, with Branded — Cometly Exact close behind. Top ad campaigns · this month Ranked by closed-won ROAS | Campaign | Spend | ROAS | Closed Won | | --- | --- | --- | --- | | Free Trial — Founder LAL | $18.2K | 6.4× | $184K | | Branded — Cometly Exact | $9.6K | 5.8× | $121K | | Demo Request — Retarget 30d | $12.4K | 4.9× | $96K | | Nonbrand — Marketing Attribution | $14.1K | 4.2× | $84K | | ABM — VP Marketing LAL | $7.2K | 3.6× | $48K | Sources Meta Ads Google Ads Stripe HubSpot Any traffic anomalies I should look into this week? Heads up — 3 posts spiked well above baseline this week, lifting organic sessions +186%. Your 2026 attribution guide is the biggest mover at +214%. Analyzing your data - Scanning organic session trends - Comparing to 8-week baseline - Attributing spikes to content Traffic anomalies · last 7 days Organic sessions vs 8-week baseline Marketing Attribution for B2B SaaS (2026 Guide) 4,812 +214% ↑ How to Send Stripe Revenue Back to Meta 2,940 +138% ↑ Multi-Touch vs Last-Click Attribution 2,110 +96% ↑ CAC Payback Benchmarks for SaaS 1,760 +72% ↑ Server-Side Tracking After iOS 18 1,240 +54% ↑ Analyzing traffic patterns… Sources Google Organic Bing Organic HubSpot ## What can I do for you, Grant? Get instant, data-backed answers about your business from every connected system. Ask anything… Standard Create a report Optimize ads Troubleshoot setup What's my ROAS over the last 30 days? Top 10 campaigns by conversions this month. Which sources brought the most contacts last week? How Agent works ## Ask in plain English. Get the answer, with the math ### Grounded in your live data Agent pulls from your live Cometly data, same metrics your dashboards use. Numbers reconcile to Stripe and your CRM. Agent Live Closed-won this month — reconciled across sources: Stripe $312,480 HubSpot $312,480 Cometly dashboard $312,480 Same metric. Same number. Every surface. ### No SQL required Ask in plain English. "Which channels produced our highest-LTV customers last quarter?" "What's our CAC payback by Q4 cohort?" Answers come back in seconds. Which channels produced our highest-LTV customers last quarter? Generated query SQL ``` SELECT source, AVG(ltv_365) FROM customers WHERE first_touch BETWEEN '2025-07-01' AND '2025-09-30' GROUP BY source ORDER BY AVG(ltv_365) DESC; ``` You write English. Cometly writes the query. ### Generates charts and tables Agent outputs the answer plus a chart or table you can save to a dashboard. Iterate the question and the visual updates instantly. ARR by source — Q3 Chart $68K Meta $88K Google $54K LinkedIn $42K Organic $36K Email Chart Table Pivot Switch view instantly ### Cites sources Every answer cites the underlying metric, time window, and filters used. You can verify the math, not just trust it. CAC payback improved to 8.2 months [1], down from 11.4 [2]. [1] cac_payback_months Plan: All · Q3 2025 [2] cac_payback_months Plan: All · Q2 2025 Click any citation to inspect the rows. ### Handles SaaS context Knows what MRR, ARR, LTV, payback, churn, and cohort mean, and how they're computed in your data. No prompt engineering required. What's our net-new ARR from the Q3 cohort, by first-touch source? Parsed by Agent Metric net_new_arr Cohort 2025-Q3 Dimension source (first_touch) Window 90d post-acquisition ### Save as report Turn any answer into a saved dashboard widget. Schedule it as a digest. Share it with your team. ARR by source — Q3 Saved Growth dashboard ARR $1.4M CAC $4.8K New widget ARR ×Src #growth-weekly Mondays · 8:00am Scheduled FAQ ## Agent on Cometly data: what teams verify before trusting answers Accuracy, guardrails, metric definitions, and how answers reconcile to dashboards. Book a call and we'll run questions against a workspace like yours. Talk to sales → Is Agent just ChatGPT on top of my data? Agent uses LLMs but is grounded in your structured Cometly data, not free-text retrieval. It generates structured queries against your attribution layer, so answers are deterministic and reconcile to your dashboards. How accurate are the answers? Answers come from the same metrics your dashboards use, computed against your live data. Every answer cites its underlying metric, filters, and time window so you can verify. Does it know my SaaS metrics? Yes. Agent is trained on the SaaS metric vocabulary, MRR, ARR, LTV, payback period, CAC, cohort, source-specific attribution. Ask in plain English and it picks the right computation. Can I save answers as dashboards? Yes. Turn any answer into a saved dashboard widget with one click. Schedule it as a digest, share it with your team, embed it in Notion. Can my agents use this through MCP? Yes. The Cometly MCP server lets Claude, GPT, and custom AI agents query your attribution data through the same Agent layer. Build internal AI workflows that use your real funnel. ## Stop waiting on the queue. Ask. Get the answer Ask any funnel question in plain English and get the answer in seconds — grounded in your live attribution data, with the chart, the math, and the sources to back it up. No SQL, no analyst queue.