OneNym.idDiscuss deployment
English

SECURITY SERVICES / INVESTIGATE · ASSESS · PREPARE

Matched to your situation,
starting with the facts.

Whether a breach is confirmed, something merely looks off, or nothing has happened yet changes both the work required and what must happen first. For organizations that use Microsoft 365, Google Workspace, cloud platforms and generative AI, OneNym.id follows a defined process for each of the three situations. This page explains everything: what we do from the first call to the final report, what you receive, and what we do not do.

This work is also the starting point before any deployment of the OneNym.id privacy platform. We first establish the facts and your current state, then identify which data and boundaries need protecting. We propose the product only when the results show it is needed, and we accept investigation or assessment work on its own.

THREE SITUATIONS

Which situation are you in?

STARTFirst call (free, 30–60 minutes)

The first question we ask is: did something specific happen, or do you want to confirm that you are safe? The answer decides which of the three paths below we follow.

PATH 1 · BREACH

A breach is confirmed or ongoing

For exampleA hijacked account, fraudulent payments or invoices, a ransom demand, or someone outside telling you your data is circulating.

What we do

  1. Contain
  2. Preserve evidence
  3. Investigate
  4. Support reporting

First response 24–72 hours, investigation 1–3 weeks

See the steps ↓
PATH 2 · SUSPICION

Something looks off, nothing is confirmed

For exampleAn unfamiliar sign-in, a suspicious email or alert, activity on a former employee's account, or confidential data that may have been entered into a generative AI tool.

What we do

  1. Preserve logs
  2. Focused investigation
  3. Breached or not
  4. Report

1–2 weeks

See the steps ↓
PATH 3 · PREPARE

Nothing has happened, and you want to be ready

For exampleYou have expanded Microsoft 365 or generative AI use, a customer or insurer has asked about your security, or an audit or listing preparation is coming.

What we do

  1. Assess
  2. Prioritize fixes
  3. Response plan
  4. Yearly review

2–4 weeks

See the steps ↓

2 → 1If path 2 finds a breach, we switch to path 1 and carry over all the work done so far.

1 · 2 → 3After path 1 or 2, path 3's assessment and response plan help prevent a repeat.

If you are not sure which applies, that is fine. We decide together on the first call.

PATH 1 · BREACH

A breach is confirmed or ongoing

This is for a breach that is confirmed or still happening. The priorities are to stop the damage from spreading and to keep the evidence before it disappears. What to do right now, and what not to do, is on the incident response page.

STEP 1First assessment and authorization (same day)
We confirm what happened, which systems are involved, what has already been changed, and whether cyber insurance or a lawyer is involved. Work starts once a contract is signed and you have authorized in writing which tenants and accounts we may access, and until when.
STEP 2Containment
We revoke sessions and tokens, reset passwords, enforce multi-factor authentication, remove suspicious connected apps and mail forwarding, and reissue API keys. Every action is logged with a timestamp so the investigation can tell our changes apart from the attacker's.
STEP 3Evidence preservation
Logs disappear once their retention period ends. Entra ID sign-in logs, for example, are kept for 7 days on the free tier and 30 days with P1 or P2, and the Microsoft 365 audit log for 180 days on the standard tier. What generative AI tools keep also depends on the plan. We export everything first, record a hash, who exported it and when, and store it in a dedicated encrypted workspace.
STEP 4Investigation and timeline
We examine identity (sign-ins, changes to multi-factor authentication and permissions, consents to connected apps), then data (downloads, external sharing links, mail forwarding), then generative AI use (what was entered or uploaded, connectors, how API keys were used). Every event goes into a single timeline with its source log and a confidence level.
STEP 5Scope of impact and reporting support
We establish which data, and how many people's data, may have left your control. If personal data is involved, Japan's Personal Information Protection Commission expects a preliminary report (about 3–5 days) and a final report (30 days, or 60 days for unauthorized access and other malicious cases), and the affected people must be notified. We deliver the facts needed for that decision in time to meet those deadlines. Whether to report is decided by you and your lawyer.
STEP 6Report and prevention
You receive a one-page summary for executives, a technical report and a prioritized list of fixes. After 30 days we check that the fixes are working.

PATH 2 · SUSPICION

Something looks off, nothing is confirmed

This is for when you do not know whether anything was compromised. The goal is to determine, from evidence, whether a breach happened. The earlier you call, the more logs remain and the more certain the answer can be.

STEP 1Interview and authorization (1–2 days)
We ask what you noticed, when, and which accounts and services are involved, then agree the scope and period of the investigation and get your written authorization.
STEP 2Log preservation
As in path 1, we export and store the logs before they disappear. Until then, please avoid deleting accounts or making large configuration changes, because that destroys evidence.
STEP 3Focused investigation
Starting from what you noticed, we check the sign-ins before and after it, permission changes, data access and generative AI input for the accounts involved. Because the scope is narrow, this is shorter than path 1.
STEP 4Determination
The result is one of three: breached, not breached, or cannot be determined because the records no longer exist. Each answer comes with the logs behind it. If it cannot be determined, we explain why and which settings would make it possible next time.
STEP 5Next steps
If there was a breach, we move straight to path 1. If there was not, we still explain the weakness that triggered the concern and how to fix it.

PATH 3 · PREPARE

Nothing has happened, and you want to be ready

We review your current state before anything happens, and set up a way of working so that nobody hesitates when something does. It is also the first step before deploying the OneNym.id platform: the results help us decide together which data you should hold, and how much of it.

What we check

01

Identity and access

How many people hold admin rights, how widely multi-factor authentication applies, Conditional Access, accounts of former employees and external users, and shared accounts.

02

Data sharing

External sharing links in SharePoint, OneDrive and Google Drive, links that anyone with the link can open, and automatic mail forwarding.

03

Generative AI use

Which teams use which AI tools, the rules for entering business data, whether your contracts exclude your data from model training, how connectors and API keys are managed, and business use of personal accounts.

04

Connected apps

Apps that users have consented to, third-party integrations with broad permissions, and integrations no longer in use.

05

Logging and detection

Whether audit logging is on, whether retention is long enough, whether enough would be recorded to investigate an incident, and who actually reads the alerts.

06

Where personal data lives

Where customer and employee personal data is held, and who can reach it. Within an agreed scope, this can be combined with an exposure assessment that tests the paths to personal data.

Steps and deliverables

STEP 1Scoping (week 1)
We learn which services you use, how your team is organized and what concerns you, then agree in writing on scope, methods, timing and data handling.
STEP 2Configuration and log review (weeks 1–3)
We work with read-only access wherever possible and do not change your settings. Your business does not need to stop.
STEP 3Findings and priorities
We rank each finding by impact and effort, and split them into what to do this week, this month, and as a planned project.
STEP 4Incident response plan
We write a playbook fitted to your organization: who to call, who decides, the first 24 hours, and how to export the logs if you ever end up in path 1 or 2. A tabletop exercise is available on request.
STEP 5Yearly review
Once a year we update the assessment and the playbook to reflect new services, new AI tools and changes in your organization.

Organizations that meet the criteria receive an AI & Cloud Security Assessment Certificate they can show to business partners. What we check, and what is required for a certificate, is published in full as our assessment criteria (Japanese and English).

AI & Cloud Security Assessment Criteria →

HOW WE WORK

How we work in every situation

Written authorization
Work starts only after we agree in writing which tenants and accounts we may access, with which permissions, and until when. We do not touch any system you have not authorized.
Evidence and data handling
Logs and data we collect are kept in a dedicated encrypted workspace, only named staff can access them, and they are deleted after the period set in the contract.
How we use AI
We use AI to sort large volumes of logs, build timelines and draft reports, only in environments contracted not to train on your data. Every point the AI raises is tied back to the original log, and only findings a person has verified go into the report.
Facts kept apart from inference
Reports separate what the logs confirm, what we infer, and what could not be determined. We never present an unknown as a finding.
What we do not do
We do not give legal judgements, such as the final call on reporting obligations or who is liable. Malware analysis on PCs and servers goes to a forensics partner. We never investigate systems we are not authorized to access, and we do not re-enact attacks.
Services we cover
Microsoft 365, Google Workspace, Azure, AWS and Google Cloud, generative AI such as Claude, ChatGPT and Copilot, major SaaS products, and web services and APIs you build yourself.

PRICING

How pricing works

Fees depend on scope and size, so we quote after the first call. Pricing is structured as follows.

First call
30–60 minutes, free of charge. It does not commit you to anything.
Paths 1 and 2: investigation
A fixed fee for the first assessment, then a fee based on days worked. We give an estimate of the days before starting and talk to you before going beyond it.
Path 3: assessment and plan
A fixed fee for an agreed scope. The yearly review is quoted separately.

If you have cyber insurance, your insurer may require a specific investigation firm. Please check with them before engaging anyone.

FAQ

Frequently asked questions

Can we contact you if we don't know whether we were breached?
Yes, that is path 2. We determine from the logs whether a breach happened. The sooner you call, the more logs remain and the more certain the answer.
Do we have to stop work during the investigation or assessment?
In most cases, no. Our work is mostly read-only. In path 1, containment such as password resets and revoked sessions may require some users to sign in again.
Can you tell whether information entered into generative AI has leaked?
If the business plan's audit logs exist, we can usually see who entered or uploaded what, and when. When personal accounts were used, the company has no records, so we rely on interviews and checks of the devices involved.
Can the results be used to report to the regulator?
We organize the facts a report needs: when it happened, which data, how many people, the cause and the measures taken. Filing the report and the legal judgement remain with you and your lawyer.
Can we engage you without using the OneNym.id product?
Yes. We accept investigation or assessment work on its own, and propose the product only if the results show it is needed.
How quickly can you respond?
We do not currently offer a guaranteed response time. If you choose “Incident response (urgent)” on the contact form, your message arrives marked urgent and we review it first.

CONTACT

Start by telling us what is happening

The first call is free, and you do not need to know which situation applies. If a breach is ongoing, use the urgent button.