Find the gaps before the market does
The share price moves on what AI reads
- 1
AI extracts the figures and tone
Trading algorithms and AI tools parse your announcement as soon as it’s released, pulling out the figures and comparing them with your guidance.
- 2
Gaps and ambiguity get misread
These tools read exactly what’s on the page, so a headline figure shown only in a graphic can be missed, and wording that doesn’t match the result can be read the wrong way.
- 3
The market acts before your call
Data feeds and trades built on that read reach investors before your results call, and a correction after the fact rarely gets the same attention.
How Affinda Signal finds the gap
-
1What you wrote
Underlying EBITDA of $471.2 million, up 4.1%, in line with our expectations for the year.
Reads well to the team that wrote it.
-
2How AI interprets it
- Underlying EBITDA
- $471.2m
- FY26 guidance, Feb 2026
- $485m–$505m
- Against guidance
- −$13.8m, 2.8% below the range
- Tone of wording
- Positive
Wording and numbers disagree
-
3What Affinda Signal gives you
High severity
Underlying EBITDA is $13.8m below the bottom of February’s guidance range. The draft calls it in line with expectations.
What drove the shortfall against the guidance range you gave in February?
What every review checks
Consistency
Does it agree with itself, your accounts and your earlier announcements?
Market expectations
How does each result compare with your guidance and with consensus?
Narrative
Is each main claim supported by a disclosed number?
Completeness
Does it give investors what they need to assess the result?
Clarity
Is the meaning clear on a first read?
Ambiguity
Could investors reach different conclusions from the same words?
Machine readability
Can software extract every figure and label correctly?
Exchange queries
Is anything likely to prompt a query from the exchange?
Ranked findings before you lodge
- 1
Add your draft
Add your draft announcement and the documents that go with it, such as the financial report and investor presentation, with broker research and call script for optional additional context.
- 2
The review runs
AI agents read the draft against your prior disclosures and the market context, reconcile the figures and test how the market will read it.
- 3
Act on the findings
Your team gets ranked findings with evidence, reconciled figures and the questions analysts are likely to ask, with time to fix them before you lodge.
Ranked by what the market
will notice first
Every figure checked
against your accounts
For the board and audit and risk committee
An independent pre-release review, with evidence for each finding and a clear record of how AI was used.
For the CFO and investor relations
Every figure reconciled against the accounts and the presentation, and the likely analyst questions before the call.
Price-sensitive drafts stay in your hands
- Runs on a machine you nominate
- Your draft is price-sensitive until it is released. The review runs on your nominated machine, and your documents never reach Affinda.
- Processed in your region
- Choose Australia, the EU or the US. Your content is encrypted and processed by AWS Bedrock in that region, with zero data retention and no model training.
- Report and audit trail stay with you
- Findings, evidence and the record of how AI was used are kept on your machine.
Affinda is ISO 27001:2022 certified and SOC 2 Type II audited.