• BUILT FOR SECTORS WHERE A WRONG SPECIFICATION IS A LIABILITY
HEALTH & SAFETY
MINING
CONSTRUCTION
GOVERNMENT
ENERGY
LOGISTICS
ERODED AUTHORITY
Unverifiable authority
Unstructured content gets paraphrased, blended with competitors or ignored. Machine-readable provenance is what makes a model trust and quote your data.
LOST PIPELINE
Absent from the shortlist
If an engine does not cite you when asked for qualified suppliers, you are removed from consideration before a person ever sees your site.
LIABILITY EXPOSURE
Hallucinated specifications
A model states the wrong pressure rating, tolerance or certification. The error travels into RFQs, safety reviews and warranty disputes, under your name.
• THE EXPOSURE
Buyers now ask an AI before they ask your sales team.
Engineers, procurement leads and purchasing agents query language models for suppliers, specifications and compliance data. What the model says becomes the brief.
• THE EVIDENCE
The GEO Scorecard shows what AI systems say about you, query by query.
Every Truth Audit returns a scorecard. Each answer is graded Cited, Misquoted or Absent, then priced as exposure so the fix can be sequenced by cost.
Share of Model
The percentage of AI answers that cite and recommend you. This is the primary metric Iddix reports.
Your company
Competitor A
Competitor B
9%
41%
27%
Sample data for illustration.
Sample data for illustration. Not a client result.
BUYER QUERY
WHAT THE MODEL SAYS
STATUS
Load rating, flagship product
Certification, key standard
Supplier shortlist, region
States a figure that does not match the datasheet
Names a competitor as the certified supplier
Lists the company and cites its product page
• Misquoted
• Absent
• Cited
• THE METHOD
Diagnose. Rebuild. Operate.
The Truth Layer is the structured, verified data infrastructure that language models and agentic systems read. Iddix architects it, deploys it and instruments it.
01
Diagnose
Query the major AI systems with the questions your buyers ask. Grade every answer and quantify the exposure.
02
Rebuild
Architect and deploy the Truth Layer: structured product, certification and entity data that machines can verify and cite.
03
Operate
Monitor Share of Model monthly and mitigate drift. Reporting is delivered through Iddix Portal.
• SERVICES
Three tiers. Fixed fee, scoped from the audit findings.
Every engagement starts with a paid diagnostic. Fees are quoted after a materials review, never hourly.
TIER 1
Truth Audit (GEO Scorecard)
The diagnostic and entry point. Establishes your exposure before any build decision.
-
AI answer testing on your buyer queries
-
Graded GEO Scorecard
-
Exposure quantified in financial terms
-
Prioritised remediation plan
TIER 2 · RECOMMENDED
Rebuilds your data so AI systems cite and recommend it, then instruments the result.
-
Everything in the Truth Audit
-
Structured data architecture, deployed
-
Corrected product and certification data
-
Monthly Share of Model reporting via Iddix Portal
Truth Layer
TIER 3
Extends the Truth Layer across the organisation’s digital estate and AI-facing workflows.
-
Everything in the Truth Layer
-
Site, content and data rebuild to specification
-
Agentic-system readiness
-
Ongoing operation and governance
Full AI Transformation
Paco Chávez
Strategic Growth Director & Brand Developer
• WHO STANDS BEHIND THE AUDIT
Accountable by name.
Every Truth Audit is led and signed off by the principal, not passed to a junior team.
01
Fixed fee, never hourly.
You know the cost of the diagnostic before it starts.
02
Every claim checked against source.
Findings cite the AI answer and the verified specification side by side.
03
You keep the scorecard.
Whether or not you engage further, the findings are yours.
04
Samples are labelled as samples.
We do not publish figures we cannot verify.
• FAQ
Questions buyers ask first.
GEO structures your digital footprint so AI systems such as ChatGPT, Perplexity, Gemini and Claude cite your data accurately and recommend you. It measures cited answers, not search rankings.
We run a defined set of buyer and procurement prompts across the major AI engines and record the percentage of answers that cite and recommend you accurately.
ChatGPT, Perplexity, Gemini and Claude as standard. Additional engines can be added to the prompt set.
It is a fixed fee, scoped to your product range and confirmed before work begins. We do not bill by the hour.