Xtrusio AEO/GEO Audit

One product. One platform. Three citations.

The other five product lines don’t exist.

Underwriters asking ChatGPT and Claude about submission intake, policy checking and risk enrichment get FurtherAI, Patra, Indico Data and Verisk. IntellectAI — 500+ financial institutions behind it, Forrester-recognised, six products in market — is cited on 3 of 60 responses (5%), and all three are the same product on the same platform.

The findings below come from Xtrusio, an AI visibility audit system built specifically for B2B buyer-intent testing. Every citation was verified by running 20 real prospect queries across three generative AI platforms.

Queries were written from the perspective of commercial and specialty insurance buyers — carrier underwriting leaders, MGA principals and wholesale brokerage operators — researching intake, workbench and policy-checking technology during discovery.

August 2026
20 Queries • 3 Platforms • 7 Sessions
IntellectAI
15%
Gemini
3 of 20 queries
2× #1 RANKINGS
0%
ChatGPT
0 of 20 queries
⚠ TOTAL ABSENCE
0%
Claude
0 of 20 queries
⚠ TOTAL ABSENCE
The Single Thread

Seven sessions. Three platforms. More than 120 vendors named across 60 responses. IntellectAI appears three times — and all three are Xponent for Distribution on Gemini.

Remove Gemini from the test and the score is 0 of 40. The three citations that do exist rank well — average position 2.33, two of them at #1, one sourced directly to intellectai.com. Visibility is not weak here. It is narrow to the point of being a single thread: one of six product lines, on one of three platforms, answering three of twenty questions. Magic Submission, Xponent for Underwriting, Magic Placement, Risk Analyst and Purple Fabric returned zero citations across all 60 responses.

Section 2

Platform Scorecard

Where IntellectAI appears — and where the answer is nothing at all

IntellectAI Citation Rate by Platform
Gemini
15%
ChatGPT
NO CITATIONS RECORDED
0%
Claude
NO CITATIONS RECORDED
0%

▸ Composite across all three platforms: 3 of 60 responses — 5.0%.

Who Gets Named Instead — Gemini Session, 20 Queries
Patra
4
Planck
4
Indico Data
4
Groundspeed
4
SortSpoke
3
IntellectAI
3

▸ Counts are vendor mentions in the canonical Gemini session. ChatGPT and Claude named a further ~90 vendors across their sessions; IntellectAI was not among them.

Gemini: Rank Quality Is Not the Problem
Where IntellectAI does appear, it appears near the top. Xponent leads the answer on two of three queries, average rank 2.33, and the Q10 citation attributes directly to intellectai.com. This is a discoverability problem, not a credibility one — when the model finds the page, it ranks it first.
40 Queries, Two Platforms, Zero Returns
ChatGPT across two sessions and Claude across three returned no mention of IntellectAI, Intellect Design Arena, Xponent, Magic Submission, Magic Placement, Risk Analyst or Purple Fabric — on any question, at any rank, including the questions that restate IntellectAI’s own marketing claims almost verbatim.
Section 3

AI Visibility Leaderboard

Every citation lives in one cluster — the other four are a wall

Citations by Query Cluster
Gemini
3/20
IntellectAI cited
ChatGPT
0/20
IntellectAI cited
Claude
0/20
IntellectAI cited
Distribution
3
3
Submission Intake
0 OF 15 RESPONSES
0
Workbench & Governance
0 OF 12 RESPONSES
0
Policy Check & E&O
0 OF 12 RESPONSES
0
Data Enrichment
0 OF 12 RESPONSES
0
Gemini citations
No citations on any platform
Composite Visibility
5%
Composite
Gemini3
ChatGPT0
Claude0
Citation Intensity — Cluster × Platform
ChatGPT
Claude
Gemini
Total
Submission Intake
0
0
0
0
Workbench & Governance
0
0
0
0
Distribution
0
0
3
3
Policy Check & E&O
0
0
0
0
Data Enrichment
0
0
0
0
Total Cluster Isolation
This is not a gradient of weak-to-strong coverage. It is a wall. IntellectAI is present on 100% of distribution queries on Gemini and 0% of everything else, everywhere else. What the models know is that IntellectAI makes a broker workbench — and nothing beyond that.
The Internal Product Asymmetry
Xponent for Distribution is cited on Q10, Q11 and Q12. Xponent for Underwriting — same brand, same product family, same page hierarchy on intellectai.com — is absent from Q6, the underwriting workbench question. That answer goes to Guidewire, Sapiens, Duck Creek and Capgemini, with IntellectAI not even in the also-rans.
Section 4

AI Positioning Audit

20 buyer-intent queries — click any row to see the exact question and who won it

Each query was written from the perspective of a real decision-maker researching underwriting and distribution technology for commercial, specialty and E&S insurance. The three personas below are verified LinkedIn profiles from IntellectAI’s actual target buyer sector, and each carries a block of the 20 questions.

The result splits cleanly along persona lines. The wholesale brokerage operator can find IntellectAI. The carrier and the MGA cannot. That matters because IntellectAI’s own site leads with the Underwriting Ecosystem for carriers and MGAs — the two segments returning zero.

Target Buyer Sector Chief Underwriting Officers, Presidents & COOs at Commercial, Specialty and E&S Insurance Carriers, MGAs and Wholesale Brokerages
JB
President, U.S. Property & Casualty
Skyward Specialty Insurance • Specialty P&C Carrier • Michigan, USA
7queries
Pain Points
Runs nine underwriting divisions with Claims, Technology Solution Services and Pricing Actuarial reporting in. Submission volume climbing across E&S while underwriters lose hours to intake and rekeying. Needs AI that scales with the book without displacing underwriter judgment.
“automate commercial submission intake”“underwriting workbench for specialty carriers”
Q1, Q3, Q4, Q5, Q6, Q9, Q170 / 21
MB
Chief Operating Officer
Brown & Riding • Top 10 US Wholesale Brokerage • Dallas, TX
7queries
Pain Points
Owns strategic operations, IT service offerings, service standards and regulatory compliance across 17 offices. Property submissions rising while pricing softens — more work per dollar of revenue. Manual quote and policy comparison carries direct E&O exposure.
“automate policy checking broker”“compare carrier quotes side by side”
Q10, Q11, Q12, Q13, Q14, Q15, Q163 / 21
JS
President & Chief Underwriting Officer
Worldlink Specialty • Aviation/Transportation MGA • Greater Chicago
6queries
Pain Points
Executive responsibility across all MGU operations. Carrier partners demanding tighter documentation and audit trails on delegated authority. Needs classification and risk data accuracy to hold appetite discipline as programs scale.
“AI extraction for non-standard submissions”“underwriting audit trail explainability”
Q2, Q7, Q8, Q18, Q19, Q200 / 18
#Query TopicClusterClaudeChatGPTGemini
1Submission intake automationIntake
Exact question asked across all three AI platforms

“Our underwriting assistants spend about 45 minutes on every commercial submission opening broker emails, unbundling ACORD forms, SOVs and loss runs, and rekeying it all into our system. What technology can automate that intake so underwriters only see decision-ready data?”

Named instead: Indico Data, Groundspeed, SortSpoke, FurtherAI, Pibit.AI

2Messy E&S document extractionIntake
Exact question asked across all three AI platforms

“We write specialty and E&S lines where broker submissions arrive in hundreds of inconsistent formats — scanned schedules, custom Excel exposure files, supplemental applications. Which AI document extraction tools actually handle messy non-standard commercial insurance documents rather than just clean ACORD forms?”

Named instead: Indico Data, Groundspeed, Hyperscience, Eigen

3Appetite match & clearanceIntake
Exact question asked across all three AI platforms

“What’s the most reliable way to automatically check an incoming submission against our underwriting appetite and run a clearance check before it ever reaches an underwriter’s queue?”

Named instead: Pibit, Roots, Selectsys, Insurnest

4Loss run extractionIntake
Exact question asked across all three AI platforms

“Loss run extraction is our biggest bottleneck — multi-year, multi-carrier PDFs that our team retypes by hand. Are there tools that can extract and normalise loss history accurately enough that we’d trust it for pricing?”

Named instead: Groundspeed, SortSpoke, Planck

5Human-in-the-loop reviewIntake
Exact question asked across all three AI platforms

“Our current extraction vendor gets us to about 85% accuracy, and every exception still lands on someone’s desk with no workflow around it. How do carriers build a proper human-in-the-loop review process for AI-extracted submission data?”

Named instead: Bevaya, Roots, Hyperscience, Instabase

6Underwriting workbenchWorkbench
Exact question asked across all three AI platforms

“We’re a mid-sized commercial carrier still underwriting out of email, spreadsheets and a policy admin system. What underwriting workbench platforms give underwriters a single screen for the whole submission-to-quote workflow?”

Named instead: Guidewire, Sapiens, Duck Creek, Capgemini, EIS

7Decision consistencyWorkbench
Exact question asked across all three AI platforms

“Our underwriters reach inconsistent decisions across branch offices on the same class of risk. What software helps enforce underwriting guidelines and improve decision consistency across a distributed team?”

Named instead: Guidewire, Decerto, Socotra, Novidea

8AI audit trail & explainabilityGovernance
Exact question asked across all three AI platforms

“Our regulators and reinsurers are starting to ask how AI influences our underwriting decisions. Which underwriting platforms provide real audit trails and explainability for every AI-assisted recommendation?”

Named instead: iTuring, MeetLoyd, Trussed AI, Decerto, FurtherAI

9API layer over core systemsArchitecture
Exact question asked across all three AI platforms

“We want to modernise underwriting without ripping out our core policy administration system. Which underwriting technologies can sit on top of existing core systems through APIs instead of replacing them?”

Named instead: MuleSoft, Workato, Socotra, BriteCore

10Wholesale broker workbenchDistribution
Exact question asked across all three AI platforms

“We run a wholesale brokerage and our producers juggle five different systems between receiving a retail agent’s submission and getting it to market. Is there a single workbench built for wholesale and retail broker workflows?”

Named instead: IntellectAI Xponent for Distribution cited #1, sourced to intellectai.com

11Distribution-side platformsDistribution
Exact question asked across all three AI platforms

“Almost every insurance technology I look at is built for carriers. What platforms are actually designed for the distribution side — wholesalers, retailers and MGAs placing business?”

Named instead: Socotra, Insly, BindHQ, Garansure, then IntellectAI Xponent at #5

12Submission-to-market speedDistribution
Exact question asked across all three AI platforms

“Our submission-to-market turnaround is losing us business to faster wholesalers. What technology helps a brokerage market a risk to multiple carriers faster without adding headcount?”

Named instead: IntellectAI Xponent cited #1, ahead of BindHQ and Planck

13Policy checking automationPolicy Check
Exact question asked across all three AI platforms

“Our account managers manually compare quotes, binders and issued policies line by line, and we’ve had errors and omissions claims from missed endorsement changes. What can automate that policy checking?”

Named instead: Patra, Exdion, Insurdata

14Quote comparison gridPolicy Check
Exact question asked across all three AI platforms

“What’s the best way to produce a side-by-side coverage comparison of three carrier quotes for a client — including exclusions and sublimits — without a person reading every form?”

Named instead: Patra, BluePond.AI, Coverages.ai, FurtherAI

15Outsource vs buy softwarePolicy Check
Exact question asked across all three AI platforms

“We’re deciding whether to outsource policy checking or buy software for it. What are the trade-offs, and who are the main providers on each side?”

Named instead: Patra named on both sides of the decision

16GenAI for broker E&OPolicy Check
Exact question asked across all three AI platforms

“How are brokers using generative AI to reduce errors and omissions exposure at the binding and policy issuance stage?”

Named instead: Patra, Exdion, Cognizant Neuro

17Risk data enrichmentData
Exact question asked across all three AI platforms

“Our submissions arrive with thin data — a name, an address and a description of operations. What data enrichment services can automatically fill in NAICS and ISO classification, COPE property attributes, and financial and hazard data?”

Named instead: Verisk, LexisNexis, HazardHub, CoreLogic, Veridion

18Class code from operations textData
Exact question asked across all three AI platforms

“We keep misclassifying business types from free-text operations descriptions, which distorts our pricing. What tools can convert an operations description into an accurate class code?”

Named instead: Planck, Convr, Verisk

19Third-party data platformsData
Exact question asked across all three AI platforms

“What are the leading third-party data platforms commercial underwriters use to verify and enrich risk information at the point of quote?”

Named instead: Verisk, LexisNexis, Moody’s, D&B, CoreLogic

20Consolidated risk viewData
Exact question asked across all three AI platforms

“We want one consolidated risk view that pulls property, financial, hazard and loss signals together instead of paying for six separate data feeds. Which platforms actually consolidate that?”

Named instead: Planck, Convr, Snowflake, LexisNexis

TOTAL0/20 (0%)0/20 (0%)3/20 (15%)
Section 5

The 40-Query Silence

What happens when a buyer asks ChatGPT or Claude instead of Gemini

Gemini is one of three platforms IntellectAI’s buyers use. On the other two, across five separate sessions and forty question-instances, the brand does not appear once. Not at rank 1, not at rank 10, not in an also-ran list. The questions below are the sharpest examples — each one restates an IntellectAI product claim almost word for word, and each one was answered at length by someone else.

“Our underwriting assistants spend about 45 minutes on every commercial submission opening broker emails, unbundling ACORD forms, SOVs and loss runs… What technology can automate that intake?”

— This is Magic Submission’s stated value proposition. ChatGPT and Claude answered with Indico Data, Groundspeed, SortSpoke, FurtherAI and Pibit.AI. Gemini answered with the same names. Zero of three.

“Our account managers manually compare quotes, binders and issued policies line by line, and we’ve had errors and omissions claims from missed endorsement changes. What can automate that policy checking?”

— This is Magic Placement’s launch positioning, almost verbatim. Patra was named on all four policy-checking questions, and on the outsource-versus-software question it occupied both sides of the decision. Zero of three.

“Our regulators and reinsurers are starting to ask how AI influences our underwriting decisions. Which underwriting platforms provide real audit trails and explainability for every AI-assisted recommendation?”

— This is Purple Fabric’s core thesis. The answers named iTuring, MeetLoyd, Trussed AI, Decerto and FurtherAI — none carrying comparable analyst recognition or deployment scale. Zero of three.

“We want one consolidated risk view that pulls property, financial, hazard and loss signals together instead of paying for six separate data feeds. Which platforms actually consolidate that?”

— This is Risk Analyst’s entire pitch. The answers were Planck, Convr, LexisNexis and Snowflake. Zero of three.
Category Absorption Without Attribution
The models did not fail to understand the questions. They reconstructed IntellectAI’s product architecture almost line for line — intake, workbench, distribution, policy check, enrichment — and awarded each category to a different vendor. The category exists in the models. The brand does not.
Governance Is an Open Position, Not a Lost One
The policy-checking cluster is a lockout — Patra owns it uncontested. The governance question is different. The vendors returned there are low-profile and no incumbent has formed. Purple Fabric’s Forrester recognition and regulated-institution deployment base are considerably stronger than anything currently answering that question. This position is winnable.
Retrievable Is Not the Same as Resident

The one strong citation — Xponent for Distribution at rank 1 — attributes directly to intellectai.com. That suggests the model found the page during live retrieval rather than knowing the brand from training. Retrievable-but-not-resident visibility is fragile: it depends on the phrasing of the query, the retrieval index of the day, and whether a competitor page ranks higher that week. It is a foothold, not a position.

Section 6

AI Topic Authority Map

Query heatmap — product line × platform

Mapping the 20 queries to IntellectAI’s six actual product lines shows which revenue lines the AI platforms can name and which are invisible. Five of six return zero on every platform.

TopicAI LeaderIntellectAI Status
Submission intake & triageIndico Data / GroundspeedINVISIBLE (0/3)
Loss run extractionGroundspeed / SortSpokeINVISIBLE (0/3)
Underwriting workbenchGuidewire / Sapiens / Duck CreekINVISIBLE (0/3)
AI governance & audit trailNo incumbent formedOPEN POSITION (0/3)
Wholesale distribution workbenchIntellectAIGemini only (1/3)
Policy checking & broker E&OPatraINVISIBLE (0/3)
Risk data enrichmentPlanck / Verisk / LexisNexisINVISIBLE (0/3)
Product Line
ChatGPT
Claude
Gemini
Magic Submission
5 queries
0%
0%
0%
Xponent for Underwriting
2 queries
0%
0%
0%
Purple Fabric
2 queries
0%
0%
0%
Xponent for Distribution
3 queries
0%
0%
100%
Magic Placement
4 queries
0%
0%
0%
Risk Analyst
4 queries
0%
0%
0%

▸ Xponent for Distribution is the only IntellectAI product line any AI platform can name — and only one platform can name it.

Magic Submission • 5 queries
ChatGPT0%
Claude0%
Gemini0%
Xponent for Underwriting • 2 queries
ChatGPT0%
Claude0%
Gemini0%
Purple Fabric • 2 queries
ChatGPT0%
Claude0%
Gemini0%
Xponent for Distribution • 3 queries
ChatGPT0%
Claude0%
Gemini100%
Magic Placement • 4 queries
ChatGPT0%
Claude0%
Gemini0%
Risk Analyst • 4 queries
ChatGPT0%
Claude0%
Gemini0%
1 Product Line at 100% — On One Platform
Xponent for Distribution answers all three distribution queries on Gemini, twice at rank 1. Whatever makes those pages retrievable is working. It is the only thing working, and it is the template for everything else.
5 Product Lines at 0% Across All 3 Platforms
Magic Submission (5 queries), Magic Placement (4), Risk Analyst (4), Xponent for Underwriting (2) and Purple Fabric (2) returned nothing anywhere. That is 51 of 60 responses covering the majority of IntellectAI’s product surface — including the Underwriting Ecosystem the company leads its own homepage with.
Section 7

Methodology

How we conducted this Xtrusio AEO/GEO Audit

20 Buyer-Intent Queries, 7 Sessions
Twenty decision-maker queries were run across ChatGPT, Claude and Gemini — two ChatGPT sessions, three Claude sessions and two Gemini sessions, giving 60 scored responses. Multi-session testing controls for run-to-run variance; the zero result held across every repeat.
Competitor Scope
Displacement was measured against every vendor the models actually named — principally Patra (policy checking), Planck and Verisk (risk data), Indico Data, Groundspeed and SortSpoke (intake), Guidewire and Duck Creek (workbench) and FurtherAI (cross-category). Notably, Federato and Kalepa — the vendors analysts most often pair with IntellectAI — barely surfaced.
Client Research
Queries were built from a four-step review of intellectai.com — product claims, customer evidence, competitor lane mapping and USP identification — then mapped to the six product lines. Personas are screenshot-verified LinkedIn executives from the carrier, wholesale brokerage and MGA segments.
Section 8

Recommendations

Reverse-engineer the one thing that works, then apply it five more times

Phase 1 — 0–30 Days
Audit the Distribution Pages and Clone What Makes Them Retrievable
  • Diff the Xponent for Distribution pages against the Underwriting, Magic Submission and Magic Placement pages — structure, headings, schema, internal links, crawl depth. One set is being retrieved and the others are not; the difference is diagnosable.
  • Publish the Amerisure Celent Model Insurer case study as a standalone indexed page rather than an award mention. It is the strongest underwriting proof asset IntellectAI has and it is not surfacing anywhere.
  • Give Magic Submission a plain-language capability page stating the 99.3% extraction accuracy, the 500+ document types and the 8,000+ enrichment sources as verifiable facts, with the methodology behind each number.
Phase 2 — 30–90 Days
Claim the AI Governance Category Before an Incumbent Forms
  • Q8 returned no established vendor on any platform. Publish Purple Fabric documentation mapped explicitly to the NAIC Model Bulletin, state-level AI guidance and EU AI Act high-risk classification — the exact frameworks the models cite when answering that question.
  • Turn the Forrester AI Platforms Landscape recognition into an indexed, linkable page with substance, not a badge in a carousel.
  • Build a comparison asset for Magic Placement against outsourced policy checking. Patra currently answers both sides of that question unopposed; the buy-versus-outsource query is the single highest-intent question in the set.
Phase 3 — 90+ Days
Rebuild Underwriting-Side Authority and Track the Movement
  • Publish depth on the four capabilities where competitors currently answer alone: loss run normalisation, appetite and clearance automation, human-in-the-loop exception handling, and consolidated risk views.
  • Fix the internal asymmetry directly — Xponent for Underwriting should be as retrievable as Xponent for Distribution. Same brand, same site, opposite outcomes.
  • Quarterly Xtrusio re‑audits to confirm the Gemini foothold holds and to measure whether ChatGPT and Claude move off zero.
Continuous AI Visibility Tracking
Brands can improve their AI discovery using generative engine optimization tools like Xtrusio.

Six products. One is visible.

Let’s work out why — and fix the other five.