Strategy
T09
Scored

Gap finder

Compare yourself with two rivals and turn the gaps into tests.

THE RESEARCH QUESTION

Where are we under-investing compared with our two main rivals?

WHEN TEAMS USE IT

Before peak season

Check whether rivals have already opened early access or sale campaigns your calendar hasn't answered.

When leadership asks if you're out-spent

Tell volume gaps apart from message and offer gaps, with dated examples.

When planning next quarter's tests

Ground each test in a dated rival pattern, not in opinion.

THE PATH

How the AI gets there, step by step.

The path the best run took, with what came back at each stop and when.

Pull activity → List what was sent → Check the web → Open each as proof

Pull activity | Counts emails and ads for the window, and the days they went out. | 10 activity pulls, 1 blocked | 0:05–0:48 | Panoramata

List what was sent | Searches the emails and ads in the window, with dates and formats. | 10 searches of emails and ads, 3 blocked | 0:27–0:55 | Panoramata

Check the web | Checks the brand's site and other public pages. | 2 searches | 0:51–0:52 | Web

Open each as proof | Reads each email and ad, with a link to it. | 3 emails and 1 ad opened | 1:04–1:08 | Panoramata

TWO PATHS, SAME AI

What changes when the AI has the data.

Each dot is one step. Pick a model to compare its two runs.

RUNS RANKED

Every run on this job, best first.

Open a model to replay its run step by step.

Run this research on the competitors you watch.

Connect Panoramata to your AI and it takes the same path, on your brands.

Brands in the tests are anonymised. AI models are named as tested.

{"v":1,"usecase":"gap-finder","k":["runSlug","model","modelId","modelSlug","maker","setup","total","factsScore","proofScore","usefulScore","speedScore","classScore","tokenScore","factsFound","factsPool","factIds","linksPassed","linksCited","time","steps","tokens","credits","cost","rank"],"r":[["t09-claude-opus-5-5-with-panoramata","Claude Opus 5.5","claude-opus-5-5","claude-opus-5-5","Anthropic","pano","85.2","27.5","25.0","15.0","5.5","10.0","2.2","11","12","1,2,3,4,5,6,7,8,9,10,11","17","17","141","27","1951889","44","1.26","1"],["t09-claude-sonnet-5-5-with-panoramata","Claude Sonnet 5.5","claude-sonnet-5-5","claude-sonnet-5-5","Anthropic","pano","84.6","22.5","25.0","12.0","10.0","10.0","5.1","9","12","1,2,3,4,5,6,7,9,10","17","17","77","20","844677","38","0.38","2"],["t09-claude-haiku-5-5-with-panoramata","Claude Haiku 5.5","claude-haiku-5-5","claude-haiku-5-5","Anthropic","pano","72.1","17.5","20.0","12.0","9.3","10.0","3.3","7","12","1,2,3,4,5,7,10","8","10","83","31","1296074","54","0.16","3"],["t09-claude-haiku-5-5-web-only","Claude Haiku 5.5","claude-haiku-5-5","claude-haiku-5-5","Anthropic","web","53.3","2.5","25.0","9.0","9.1","0.0","7.7","1","12","12","7","7","85","19","563031","0","0.11","4"],["t09-claude-opus-5-5-web-only","Claude Opus 5.5","claude-opus-5-5","claude-opus-5-5","Anthropic","web","30.7","5.0","6.7","9.0","5.5","0.0","4.5","2","12","6,11","4","15","139","25","960016","0","0.71","5"],["t09-claude-sonnet-5-5-web-only","Claude Sonnet 5.5","claude-sonnet-5-5","claude-sonnet-5-5","Anthropic","web","25.0","0.0","0.0","6.0","9.0","0.0","10.0","0","12","","0","3","86","11","431875","0","0.19","6"]]}
{"v":1,"L":["setup|Loaded the Panoramata tools","pano|Pulled activity stats","pano|Searched ads","pano|Read the brand summary","pano|Searched emails","web|Searched the web","web|Read a page: News or blog article","web|Read a page: Third-party tracker","pano|Pulled daily send and launch counts","pano|Opened an email","pano|Opened an ad","web|Read a page: Panoramata public page","web|Read a page: Social network page","pano|Checked workspace","pano|Looked up the brand","pano|Found competitors"],"S":{"t09-claude-sonnet-5-5-with-panoramata":"6:0 7:1! 8:1 8:1 11:1! 11:2 12:2 15:1 19:3 20:3 20:3 21:4 21:4 22:4 23:2 29:4 30:4 31:2 32:2 32:4","t09-claude-sonnet-5-5-web-only":"6:5 7:5 12:6 12:6 13:6 13:7 25:6 26:7 27:7 27:7 39:5","t09-claude-opus-5-5-with-panoramata":"2:0 5:1 5:1 6:1 10:8 11:8 11:8 12:8 13:8 14:8 27:2 28:2 29:2 35:2 36:4! 37:4 38:4 42:4! 45:4! 48:3! 51:5 52:5 55:4 64:9 65:9 68:9 68:10","t09-claude-opus-5-5-web-only":"2:5 3:5 3:5 7:6 8:7 19:5 19:5 20:5 24:7 24:5 25:5 28:5 29:6 30:6 39:5 39:5 40:5 44:11 45:11 46:7 61:6 62:6 63:5 72:12 73:6","t09-claude-haiku-5-5-with-panoramata":"4:0 6:13 6:14! 6:14 6:14 10:14 10:15 10:1 11:1 14:1 14:8 15:8 15:8 15:8 18:2 19:2 19:2 19:4 19:4 20:4 34:3 34:3 35:3 35:2 35:4 36:4 42:2! 43:2! 43:2! 43:2! 46:2!","t09-claude-haiku-5-5-web-only":"6:5 6:5 6:5 6:5 7:5 7:5 11:7 11:6 12:5 12:11 26:5 27:7 27:7 40:6 41:7 41:7 53:7 54:7 54:6"},"N":{"t09-claude-sonnet-5-5-with-panoramata":[["17 of 17 links passed.","Every link opens the right item in Panoramata."],["0 invented links.","The facts were credited because the judge confirmed the numbers in Panoramata: 258/153/227 ads, 81/84/22 emails, exactly 5 promotional emails for Rival 1 and 0 for Brand G, the Oct 28 early-access ad cluster, and creator ads at about 25% of Rival 2's ads."],["Classification:","10. It is the only run that reports the exact promotional email count for Rival 1 (5). It also flags the gap between stats and search counts (81 vs 28) honestly."]],"t09-claude-sonnet-5-5-web-only":[["0 of 3 links passed.","The three third-party tracker pages open but are dated Mar 7 to 15, 2026, about 7 months before the window."],["0 invented links.",""],["Classification:","0. It has no window counts and no email data. The run is honest that every gap is a hypothesis, and it recommends an email audit before any email test."]],"t09-claude-opus-5-5-with-panoramata":[["17 of 17 links passed.","Every link is a public Panoramata shared page showing the right brand, and every dated item falls inside the window."],["0 invented links.","The judge confirmed the cadence numbers in Panoramata: 34 vs 12 emails from Oct 21 to Nov 1, and halves within 1 to 2 of the stated 46/35 and 29/55. Its count of 6 promotional emails for Rival 1 is one more than Panoramata's category count of 5."],["Classification:","10. It is the only run that finds the October cadence drop and Rival 2's members-only early access, and it ties each of the three tests to a gap."]],"t09-claude-opus-5-5-web-only":[["4 of 15 links passed.","These are the two press stories on Rival 1's fall celebrity campaign (Sep 2 and Oct 2), Brand G's undated strategy page, and Rival 2's site with a live members-first early access (blocked by the automatic check, but opened through WebFetch). The other 11 are dated outside the window."],["0 invented links.",""],["Classification:","0. It is clear that it could not measure the 60 days. Its strongest in-window finding is the gap in fall celebrity campaigns."]],"t09-claude-haiku-5-5-with-panoramata":[["8 of 10 links passed.","All are public Panoramata shared pages. The two that fail are Brand G's past sale emails (Dec 2025 and Jun 2026), which are dated outside the window, although the run labels them as history."],["0 invented links.","Counts confirmed in Panoramata: 258/153/227 ads and 81/84/22 emails. Its count of '8+' promotional emails for Rival 1 is too high (Panoramata has 5)."],["Classification:","10. It finds the main offer gap and the early-access ad cluster, but misses Rival 2's member early access and the cadence drop. Creator share is left uncounted because of a rate limit."]],"t09-claude-haiku-5-5-web-only":[["7 of 7 evidence links passed.","These are three third-party email archive pages dated Oct 2026, two copies of an undated ad-tracker page for Rival 1, an undated placeholder ad page for Brand G, and an undated article on Rival 2. A 2019 article was used only as background and is not counted."],["0 invented links.",""],["Classification:","0. It says honestly that Brand G's data is missing. Its welcome-flow comparison rests on a 3-email sample."]]}}