AI-RIS
Service · Report checking

Find the mistakes in your report before anyone else does.

Annual and sustainability reports, results reports for funders, reports for clients and regulators: they are long, written by many hands and full of numbers, and somewhere in each is a claim you can’t back up. We find it first, show you exactly where it is, and tell you how to fix it. For companies, consultancies, public institutions and non-profits.

Somewhere in your report, a number is wrong.

A figure starts in a spreadsheet, a branch office or a field team. It is copied into a department report, rounded in a regional one, and described a little more generously in the annual summary. By the time it reaches your board, your investors or your funder, it has quietly become a different number.

Nobody meant to mislead. It just happens. We have seen it from both sides: checking reports, and writing them.

The four mistakes hiding in almost every report
Credit that isn’t clearly yours Growth of the whole market, or your partners’ work, ends up reported as your own result.
Numbers that change meaning “Active customers”, “jobs created” or “people reached” is counted one way this year and another way next year.
Sources that don’t say that, or don’t exist One quotation in six misrepresents its source; AI-written drafts now cite papers nobody can find.
Results with nothing to measure against Achievements that don’t match anything you set out to measure at the start.
How it works
01

We find it

Our software reads every page. It redoes the sums, compares each number with the same number in your other reports, checks that every cited source exists and says what you say it says, and marks every claim of success.

02

We prove it

A second, separate check must find the proof for each problem: the page, the exact sentence and, where needed, an official source such as national statistics. No proof, no finding.

03

We judge it

A reviewer trained as a lawyer, with more than 20 years of writing and checking reports for the UN, USAID, AusAID and in business, goes through every problem, reads the charts the software can’t, and decides what really matters. Software finds; a person judges.

You’ll know what to fix first
1
Could cost you trustClaims a reader could challenge: success that isn’t proven, credit that isn’t yours, numbers that can’t be right.
2
Gaps in how you keep trackThe same thing measured differently in different reports, or never measured at all.
3
Fix before you publishWrong percentages, footnotes in the wrong place, charts that disagree with their own tables.

We tell you how sure we are about each one. When we’re not sure, we ask a question instead of pointing a finger.

What we’ve caught in public reports · names and amounts removedlevel
A cost per person reported as down 28% on last year. It was compared with a figure last year’s report never published; against the one it did publish, the cost had gone up.1
A cost per person that doesn’t match its own footnote: the footnote’s sum comes out 4% higher.3
A programme budget printed with two digits swapped between header and text, and the slip copied into the next year’s report.3
Private investment understated a thousand times: millions printed as thousands.3
Extra income claimed for 17% more farmers than the income results show, with the number of women taken from a different result. The report’s own table shows it.1
A profit margin printed as 4.9% in a table and 4.8% elsewhere. Worked out from that same table, it is 4.8%.3
One key measure, free cash flow, defined in at least three different ways in the same annual report.2
What a finding looks like
Annual review · July 2019 · Section D
Value for money

Efficiency refers to how well programme inputs are converted into outputs. The majority of the programme’s efficiency indicators continued a positive downward trend. The first is total spend per beneficiary (access/use) of £■■■, compared to £■■■ in Year 4 and an expected £■■■ in the original business case. This represents a 28% decline from last year.

!
Annual review · the year before · Section D

The programme has a total spend per beneficiary of £■■■, compared to £■■■ the year before and an expected £■■■ in the original business case.

From our own check of the public annual reviews of a closed aid programme. Names and amounts removed, wording shortened; percentages as published.
FindingCould cost you trust
Where it isAnnual review 2019, section D, value for money. Repeated in the programme’s completion review.
What it says“£■■■, compared to £■■■ in Year 4 … a 28% decline from last year.”
What we foundThe comparison figure is not the one last year’s review published. That review gave a lower figure, so measured against it the cost went up, not down 28%.
How to fix itCompare with the figure you published last year, or explain why it changed. Then correct the completion review as well.
Pick your starting point
Check one report Before it goes to your board, investors, funder or client.
Check several years We compare your reports year by year and show where the numbers drifted.
Get it right from the start For a new project or reporting cycle: we help you decide what to measure and how, so every result can be proven later.
Every problem, in four lines you can act onfour lines
Where it is
The report and the page
What it says
The exact words, quoted
What we found
The page, table or official source that disagrees
How to fix it
What to write instead
A clear list of problems, most important first, each with a fix
Your report marked up, page by page
Practical advice on avoiding the same problems next time
Kept confidential. We point to facts, never blame people.
Send us one report. We’ll show you what a careful reader would have found. Send us a report or write to info@ai-ris.eu
AI-RIS
HR
Service · Report checking

Find the mistakes in your report before anyone else does.

Annual and sustainability reports, results reports for funders, reports for clients and regulators: they are long, written by many hands and full of numbers, and somewhere in each is a claim you can’t back up. We find it first, show you exactly where it is, and tell you how to fix it. For companies, consultancies, public institutions and non-profits.

Somewhere in your report, a number is wrong.

A figure starts in a spreadsheet, a branch office or a field team. It is copied into a department report, rounded in a regional one, and described a little more generously in the annual summary. By the time it reaches your board, your investors or your funder, it has quietly become a different number.

Nobody meant to mislead. It just happens. We have seen it from both sides: checking reports, and writing them.

The four mistakes hiding in almost every report
Credit that isn’t clearly yours Growth of the whole market, or your partners’ work, ends up reported as your own result.
Numbers that change meaning “Active customers”, “jobs created” or “people reached” is counted one way this year and another way next year.
Sources that don’t say that, or don’t exist One quotation in six misrepresents its source; AI-written drafts now cite papers nobody can find.
Results with nothing to measure against Achievements that don’t match anything you set out to measure at the start.
How it works
01

We find it

Our software reads every page. It redoes the sums, compares each number with the same number in your other reports, checks that every cited source exists and says what you say it says, and marks every claim of success.

02

We prove it

A second, separate check must find the proof for each problem: the page, the exact sentence and, where needed, an official source such as national statistics. No proof, no finding.

03

We judge it

A reviewer trained as a lawyer, with more than 20 years of writing and checking reports for the UN, USAID, AusAID and in business, goes through every problem, reads the charts the software can’t, and decides what really matters. Software finds; a person judges.

You’ll know what to fix first
1
Could cost you trustClaims a reader could challenge: success that isn’t proven, credit that isn’t yours, numbers that can’t be right.
2
Gaps in how you keep trackThe same thing measured differently in different reports, or never measured at all.
3
Fix before you publishWrong percentages, footnotes in the wrong place, charts that disagree with their own tables.

We tell you how sure we are about each one. When we’re not sure, we ask a question instead of pointing a finger.

What we’ve caught in public reports · names and amounts removedlevel
A cost per person reported as down 28% on last year. It was compared with a figure last year’s report never published; against the one it did publish, the cost had gone up.1
A cost per person that doesn’t match its own footnote: the footnote’s sum comes out 4% higher.3
A programme budget printed with two digits swapped between header and text, and the slip copied into the next year’s report.3
Private investment understated a thousand times: millions printed as thousands.3
Extra income claimed for 17% more farmers than the income results show, with the number of women taken from a different result. The report’s own table shows it.1
A profit margin printed as 4.9% in a table and 4.8% elsewhere. Worked out from that same table, it is 4.8%.3
One key measure, free cash flow, defined in at least three different ways in the same annual report.2
What a finding looks like
Annual review · July 2019 · Section D
Value for money

Efficiency refers to how well programme inputs are converted into outputs. The majority of the programme’s efficiency indicators continued a positive downward trend. The first is total spend per beneficiary (access/use) of £■■■, compared to £■■■ in Year 4 and an expected £■■■ in the original business case. This represents a 28% decline from last year.

!
Annual review · the year before · Section D

The programme has a total spend per beneficiary of £■■■, compared to £■■■ the year before and an expected £■■■ in the original business case.

From our own check of the public annual reviews of a closed aid programme. Names and amounts removed, wording shortened; percentages as published.
FindingCould cost you trust
Where it isAnnual review 2019, section D, value for money. Repeated in the programme’s completion review.
What it says“£■■■, compared to £■■■ in Year 4 … a 28% decline from last year.”
What we foundThe comparison figure is not the one last year’s review published. That review gave a lower figure, so measured against it the cost went up, not down 28%.
How to fix itCompare with the figure you published last year, or explain why it changed. Then correct the completion review as well.
Pick your starting point
Check one report Before it goes to your board, investors, funder or client.
Check several years We compare your reports year by year and show where the numbers drifted.
Get it right from the start For a new project or reporting cycle: we help you decide what to measure and how, so every result can be proven later.
Every problem, in four lines you can act onfour lines
Where it is
The report and the page
What it says
The exact words, quoted
What we found
The page, table or official source that disagrees
How to fix it
What to write instead
A clear list of problems, most important first, each with a fix
Your report marked up, page by page
Practical advice on avoiding the same problems next time
Kept confidential. We point to facts, never blame people.
Send us one report. We’ll show you what a careful reader would have found. Send us a report or write to info@ai-ris.eu