25+
technical & entity signals scored per audit
Position or lose.
No position, you lose. Right position, you win.
Every category has a leader and everyone else. AI doesn't read your deck — it just repeats whatever position you've actually earned, good or bad. Where do you stand right now? Positioning Lab measures it, across every model that matters, and shows you exactly what it takes to win.
Free, live, no card — or plans from $79 to run the full audit.
Where do you stand?
Right now, today, across ChatGPT, Claude, Perplexity, and Gemini — you have a position. You just don't know what it is yet.
Do you want to win?
Then stop guessing. The category leader isn't always the best product — it's the one AI keeps citing as the answer.
How do you win?
With Positioning Lab. Measure the position. Fix the gaps. Take the spot.
The scoreboard
Winning isn't a vibe — it's a number you can defend. A reproducible methodology for tracking AI visibility, built on captured prompts, scored signals, and engine-by-engine evidence. No black boxes.
4
frontier engines probed every run
5
composite scores per brand, 0–100
∞
audit history retained per workspace
Platform
Winning isn't a one-time score. It's a position you hold.
Audits, monitoring, and reporting in a single workspace. Built so a marketing team and an engineering team can use the same numbers to defend the same position.
Workspace
Purpose-built, reproducible, engine-agnostic.
Every audit is a snapshot of how AI sees your brand at a moment in time. Stack them across weeks and you have a defensible curve. Run them across competitors and you have a category benchmark. The platform handles the orchestration — your team handles the strategy.
Audits
Probe ChatGPT, Claude, Perplexity, and Gemini with five prompt classes. Capture every answer. Score every signal.
Monitoring
Schedule re-scans. Get drift alerts when an engine flips on you.
Reporting
Branded HTML and JSON export, a print-ready view (Cmd+P → Save as PDF), and a public-share URL for stakeholders.
How a probe runs
Six steps from unknown to undeniable.
Every score traces back to a captured artifact: a prompt, an answer, a header check, an entity record. You can audit the auditor.
01 / Capture
Site + entity ingest
Crawl the site. Pull headers, schema, sitemap, llms.txt, OpenGraph, robots policy. Resolve the brand against Wikidata and Wikipedia.
02 / Probe
Multi-model prompting
Run five prompt classes against ChatGPT, Claude, Perplexity, and Gemini. Persist every raw response with timestamps, model versions, and citations.
03 / Score
Composite scoring
Five composite scores (Overall · Clarity · Differentiation · AI citability · Entity authority) computed from the captured signals. Weights documented.
04 / Surface
Prioritised remediation
Top three fixes ranked by expected score impact. Each with current copy, rewrite, and the reason it moves the number.
05 / Track
Longitudinal comparison
Re-scan on demand or schedule. Diff every signal. Detect engine drift before it costs you pipeline.
06 / Export
Reportable output
Branded HTML for stakeholders, structured JSON for engineering, a print-ready view for Save-as-PDF (Cmd+P), and a public link for clients. Same data, four surfaces.
Full lifecycle
From unknown position to won position.
The complete trip, six stages, one continuous run. Nothing skipped.
positioning lab · running
headers
schema
sitemap
llms.txt
chatgpt
claude
perplexity
gemini
0
overall score
#1 · +12
#2 · +8
#3 · +5
Δ +6 vs last scan
HTML
JSON
PDF
>
starting the run — resolving site + entity…
The front door
Free live probe. $49 to keep it private.
Run ChatGPT, Perplexity, Claude, and Gemini against your domain live — free, no card required. Want it private from the start instead of public by default? $49, one time, paid up front.
Run yours — free
no login required · results in the live view
Models
Four judges. One position to defend.
No vendor preference. No single-model bias. Buyers move between engines — your measurement should too.
OpenAI
ChatGPT
Largest consumer surface area. Recognition here predicts general-market awareness.
Anthropic
Claude
Preferred by technical and operator audiences. A higher bar — recognition signals real authority.
Perplexity
Perplexity
Citation-native. Surfaces the exact sources AI is pulling from when discussing your brand.
Google
Gemini
Tied to Google's index and AI Overviews. A leading indicator for Search Generative Experience visibility.
Trust
A claimed win means nothing. A provable one does.
Every claim in a Positioning Lab report is backed by a captured artifact. We retain the raw payloads. You can re-derive the score.
Auditable
Reproducible. Inspectable. Versioned.
Model versions, prompt text, and timestamps are captured on every probe. Site signals are stored as the raw response, not a summary. If a score gets challenged, you can re-derive it from the artifact.
Reproducible
Same prompt, same model, same week — same answer captured for both your team and ours.
Versioned
Every audit is timestamped and immutable. Comparison reports diff signal-by-signal.
Portable
Export to JSON. Pipe into your warehouse. Own your AEO data.
Who it's for
Teams that refuse to lose the category.
Founders defining a category. Marketing teams owning AI visibility as a metric. Agencies running it at scale.
Founders & GTM
Define the category
Verify what AI says when a buyer asks for "the best X." Track the gap to incumbents. Close it deliberately.
In-house marketing
Own AEO as a metric
Five composite scores, monthly trend lines, evidence on every claim. Walk into the QBR with numbers, not narrative.
Agencies & consultancies
Operate at scale
Run audits across up to 10 monitored brands per workspace. Schedule re-scans. Branded HTML/JSON export with a print-ready view. Read-only audit history API on Agency tier.
Scope
What we don't do.
If you need any of the below, we're the wrong platform.
× Bespoke websites or web design
× Outbound / SDR / lead-gen builds
× CRM implementation, RevOps consulting, pipeline systems
× “AI strategy” decks
× SEO services — link building, content production, ranking work
Recent teardowns
Some brands are winning. Some don't know they're losing.
This is the same probe-and-score pipeline every audit runs — applied publicly.
northlake.io
Overall 58 / 100
Named correctly by 3 of 4 models, but never cited as a source — no llms.txt, thin entity graph.
fernwood-analytics.com
Overall 71 / 100
Strong Wikidata presence carries entity authority; category description drifts across models.
brightloop.ai
Overall 44 / 100
Confused with a same-named competitor by Perplexity in 4 of 5 prompt runs.
Illustrative examples — fictitious brands, real scoring mechanism.
Get started
Stop wondering where you stand. Find out, and go win.
Run your first audit, get five composite scores, raw answers from four frontier engines (ChatGPT, Claude, Perplexity, Gemini), technical findings, and a prioritised remediation plan.