pressGEO

GEO Lift Pilots

We pick a real SaaS vendor already being probed by our AI-engine monitor, snapshot their 30-day citation rate, and apply the PressGEO GEO playbook to their surface. For every pilot we also assign matched, untreated control vendors from the same category. The headline number is net lift — treated delta minus control delta — so category-wide drift is subtracted out.

Methodology
  • · Candidates need ≥40 probes and a 5–25% baseline citation rate.
  • · Baseline snapshotted over the 30 days before pilot start.
  • · 3 control vendors matched in the same category within ±5pp (fallback ±15pp, ±25pp).
  • · Results held back until ≥200 post-treatment probes have run.
  • · Net lift = treated Δ − control-cohort avg Δ.
  • · Every number is a direct read of the public probe database.

Active pilots

Completed pilots

Final numbers from finished pilots. Metrics were frozen the day the pilot completed, so these figures don't drift with new probes.

  • AvidXchange cited by Perplexity
    Perplexity · cited
    AvidXchange
    accounts-payable · ran 8/21/2026 – 9/17/2026 · 5 controls
    +30.5pp net
    raw +38.05pp · controls +7.6pp
    106 probes
    5 queries flipped · 5 engines citing
    Claim AvidXchange
  • Miss
    →
    Gorgias cited by Perplexity
    Gorgias
    helpdesk · ran 7/16/2026 – 8/3/2026 · 3 controls
    +27.8pp net
    raw +30.17pp · controls +2.37pp
    1,482 probes
    5 queries flipped · 5 engines citing
    Claim Gorgias
  • Miss
    →
    Insightly cited by Gemini
    Insightly
    crm · ran 7/9/2026 – 7/13/2026 · 3 controls
    +15.88pp net
    raw +16.52pp · controls +0.65pp
    422 probes
    5 queries flipped · 4 engines citing
    Claim Insightly
  • Miss
    →
    Bench cited by Copilot
    Bench
    accounting · ran 7/9/2026 – 7/13/2026 · 3 controls
    +12.21pp net
    raw +6.76pp · controls -5.45pp
    422 probes
    3 queries flipped · 4 engines citing
    Claim Bench
  • Miss
    →
    Fathom cited by Copilot
    Fathom
    analytics · ran 7/14/2026 – 8/3/2026 · 3 controls
    +6.72pp net
    raw +17.16pp · controls +10.44pp
    1,567 probes
    2 queries flipped · 5 engines citing
    Claim Fathom
  • Miss
    →
    Dropbox Sign cited by Copilot
    Dropbox Sign
    e-signature · ran 8/4/2026 – 8/7/2026 · 6 controls
    +5.32pp net
    raw +6.83pp · controls +1.51pp
    448 probes
    3 engines citing
    Claim Dropbox Sign
  • Close cited by Copilot
    Copilot · cited
    Close
    crm · ran 7/13/2026 – 7/20/2026 · 3 controls
    +3.14pp net
    raw +3.07pp · controls -0.08pp
    621 probes
    3 engines citing
    Claim Close
  • Miss
    →
    Monday CRM cited by Copilot
    Monday CRM
    crm · ran 7/13/2026 – 7/14/2026 · 3 controls
    +2.78pp net
    raw +1.59pp · controls -1.18pp
    236 probes
    1 queries flipped · 4 engines citing
    Claim Monday CRM
  • Miss
    →
    SugarCRM cited by Copilot
    SugarCRM
    crm · ran 7/9/2026 – 7/13/2026 · 3 controls
    +2.13pp net
    raw +2.78pp · controls +0.65pp
    422 probes
    2 queries flipped · 5 engines citing
    Claim SugarCRM

Every number here is a direct read of the same probe database shown on our public benchmarks. Control vendors are chosen for matching baseline citation rate in the same category and receive no treatment. Vendors are listed publicly as part of a GEO research demo; pilot vendors can opt out at any time by emailing pilots@pressgeo.com.