IRO Landscape · Datamaran / SaaS / B2B / AI

Turning an industry's sustainability reports into answers

Large European companies must now publish an audited sustainability report. The hardest part is deciding what goes in it. We built an AI benchmark that shows what the whole industry reports, so teams can do it in minutes instead of weeks.

RoleSenior Product Designer

FocusAI · Industry benchmark

Year2026

Why now

Datamaran helps nearly 500 large companies, like Visa, Amazon, Dell, Cisco and PepsiCo, manage sustainability risks and reporting. Three things made this the right moment:

A new law: The EU's CSRD makes thousands of companies publish an audited sustainability report every year.
New data: The first CSRD reports are out, so we can finally see what each industry reports.
AI: Language models can read and classify those long reports at scale.

What an IRO is

Before writing their report, companies list how they affect the world, and how the world could affect them. Each item is an IRO: an Impact, a Risk or an Opportunity. For a chocolate maker:

Type	Example
Positive impact	Our training program raises cocoa farmers' income
Negative impact	Our plastic wrappers add to packaging waste
Risk	A plastic tax could raise our costs
Opportunity	Plastic-free packaging could win new customers

The problem

Talking to clients across logistics, payments, tech and consumer goods, one question came up more than any other:

“Is my analysis complete, and what am I missing?”

Benchmarking by hand: Analysts read competitors' reports one by one and copied every IRO into spreadsheets.
No answer for the board: Leadership asked “are we in line with our sector?” and nobody had evidence.
Fear at audit time: A missing IRO found late by the auditor means rework and a delayed report.

The vision

Give every sustainability team the knowledge of an analyst who has read every report in their industry.

The solution

The IRO Landscape is an AI benchmark built into Datamaran's materiality assessment, with three pillars: a map of the industry, an AI assistant and a co-creation assistant.

The foundation: AI that reads the market

All three pillars stand on Datamaran's AI, which reads more public text than any team could:

Source	What the AI reads	Why it matters
Corporate reports	Annual, sustainability and CSRD reports from about 10,000 companies	What competitors say about themselves
Regulations	Over 19,800 laws and initiatives in more than 190 countries	What companies must do now, or soon
News	Global online media	What the public and investors talk about
Social media	Public posts and conversations	How people react to issues and companies

For the IRO Landscape, the key source is audited CSRD reports. The AI finds every IRO, classifies it by topic, type and value chain, and picks up new reports as they come out.

1. Landscape Map

A heatmap of the whole industry: topics down the side, IRO types across the top, and the share of peers that report each one. One click shows the peer IROs, and an overlay marks where the user's own list has gaps.

IRO Industry Landscape heatmap: sustainability topics by IRO type, with the share of peers that report each one — The map: topics by IRO type, with the share of peers that report each one.

Detail panel for Water and marine resources risks: 78% industry disclosure, 11 of 14 peers, and a disclosure gap for the user's company — Each cell opens the detail: how many peers report it, and whether the user has a gap.

Disclosing companies tab for Water and marine resources risks: five peers listed, with Merck's two reported IROs on the right — Disclosing companies: the peers that report the topic, and the exact IROs each one wrote.

Heatmap with the company overlay on: dots mark gaps and topics already covered — Compare with my company: gaps and covered topics marked on the map.

2. AI Assistant

Users can simply ask about their industry. Instead of walls of text, the assistant answers with the right component: a chart, a list of peers, a gap card or a ready-to-add IRO.

The assistant works next to the map: ask, compare with peers and draft IROs.

AI Assistant opened next to the heatmap with suggested questions — The assistant opens next to the map, with suggested questions.

AI Assistant answering which companies report water risks, with a list of peers and their IRO counts — Answers come as components, with a shortcut to draft IROs.

3. Co-Creation Assistant

For every gap, the assistant drafts an IRO adapted to the user's company. They refine it by chatting or editing, and add it to their list in one click.

Designing for trust

Reports are audited, so every number has to hold up. Every answer links to its source, “no data” is never shown as 0%, and the AI never makes decisions for the user:

Impact	Example	What the AI does
Low	Answer a question, show a chart	Acts on its own, with sources
Medium	Draft or edit an IRO	Suggests, the user confirms
High	Mark a topic as material	Never acts, the user decides

Did we solve the problem?

Yes. We tested it with 8 sustainability managers and 300 real client questions. What used to take weeks of reading reports now takes minutes, without leaving the assessment.

Clients use it actively during their assessments: they add IROs to their analysis straight from the IRO Landscape, and those IROs end up in the final analysis they present to their board and auditors.

It also opened a new business line: Datamaran now sells the IRO Landscape as a separate product.

87%of drafted IROs accepted as is or with minor edits
100%of assistant numbers match the heatmap