FireTail Research · Edition 01 Edited · Q2 2026 · Reading time 16 minutes 77 days to August 2 Web version →
FireTail Research · Edition 01 · Published Q2 2026

Most enterprises will arrive at the August 2 deadline without the controls the law assumes they have.

A year of incidents, shadow tools, and ungoverned deployments has compressed the AI security problem into a single deadline. This is the report of what we found in the lead-up to it.

13%
of organisations IBM studied reported a breach of an AI model or application. Another 8% said they did not know.
IBM · Cost of a Data Breach 2025
97%
of those breached had no AI access controls in place. The gap is foundational, not a tooling problem.
IBM · Cost of a Data Breach 2025
77 days
to August 2, 2026 — the date the EU AI Act starts applying to high-risk AI systems across the bloc.
EU Regulation 2024/1689 · Article 99
How to read this

Nine chapters. Pick where to start.

The report builds in order if you have the time. Otherwise, jump to whichever chapter answers the question your board is asking on Monday.

Chapter 01 The shift

AI adoption has outpaced security by every measure that matters.

The first wave of generative AI was about prompts. Then employees brought their own tools. Then vendors shipped agents. Each wave landed before the controls for the previous one had finished being written. The 2025 IBM data is the cleanest read of where that has left enterprises: 13% have already had a breach involving an AI model or application, and 97% of those organisations had no AI access controls in place when it happened.

63%
of breached organisations have no AI governance policy in place — or are still drafting one.
IBM · Cost of a Data Breach 2025
20%
had a breach involving shadow AI — unsanctioned employee use of tools like ChatGPT or personal Copilot accounts.
IBM · Cost of a Data Breach 2025
$4.63M
is the average shadow-AI breach — $670K above the global average. Highest cost-per-record of any vector this year.
IBM · Cost of a Data Breach 2025
The new shape of the kill chain
Time to breach · 4 seconds · Human actions required · 1
↳ Attacker Malicious input Email, document, ticket
Stage 02 AI agent ingests Content enters context window
Stage 03 MCP tool call Executes injected instruction
↳ Breach Data exfiltrated No user interaction required

The three waves, very compressed

The first wave was conversational. ChatGPT inside the browser, employees pasting in code or contracts, security teams writing acceptable-use memos that almost nobody read. The data leakage problem is real — 27% of organisations report that more than 30% of the data flowing through AI tools contains private information — but the controls for this layer are at least understood: DLP, CASB, browser policy.

The second wave was vendor AI — Copilot in Microsoft 365, Gemini in Workspace, AI features inside every SaaS your business already paid for. Different problem. The model is reading your real data. If anyone can write instructions into that data, they can write instructions for your AI. EchoLeak is the documented case; we walk through it in Chapter 03.

The third wave is agents. Tools, MCP servers, autonomous task execution. The category did not exist as an enterprise concern eighteen months ago. It is now in production at most large organisations — with, in most cases, no inventory, no logging, and no policy specific to it.

Why the controls did not keep up

A representative read from the IBM cohort, with our own field observations.

97% missing access controls. Of organisations that reported an AI-related breach, almost none had functional AI-specific access controls. This is the single most common gap in the dataset.
47% personal accounts. Almost half of employees using generative AI at work are doing so through their personal accounts — outside the visibility of any enterprise tool. Netskope tracks the top quartile at 2,100 data-policy violations per month.
34% audit, 12% govern. Even where governance policies exist, only a third include audits for unsanctioned AI. Gartner puts the share of organisations with dedicated AI governance structures at 12%.
16% attacker AI. Attackers are not waiting. IBM's 2025 cohort saw AI on the attacker side of one in six breaches — mostly AI-generated phishing (37%) and deepfake impersonation (35%).
Chapter 02 Pattern

Most AI breaches in 2025 used trust relationships your organisation had already approved.

Four featured incidents below — chosen because together they describe the four shapes the AI security problem actually takes today: shadow tooling, supply chain trust, model-context manipulation, and OAuth overreach. The full tracker, with every incident we have documented this year, lives on the breach tracker page.

700+
organisations had Salesforce data accessed in the UNC6395 Drift OAuth incident — from one compromised third-party AI integration.
Mandiant / Reco · August 2025
241 days
mean time to identify and contain a breach in 2025. The nine-year low — still long enough for an attacker to do everything they wanted to.
IBM · Cost of a Data Breach 2025
60%
of AI-related security incidents in the IBM cohort originated in the AI supply chain — compromised apps, APIs, or plug-ins.
IBM · Cost of a Data Breach 2025
Incident Shape Reach Severity Period
UNC6395 — Drift / Salesforce OAuthStolen OAuth tokens from a trusted AI integration. The attacker did not exploit a vulnerability; they used legitimate third-party access. OAuth overreach 700+ organisations Critical Aug 2025
EchoLeak — Microsoft 365 CopilotCVE-2025-32711. Hidden prompt in a document instructs Copilot to forward mail. Privately patched before public release — mechanism is general. Indirect prompt injection Vendor-wide Critical Disclosed 2025
Cursor IDE RCE chainTwo vulnerabilities in the AI coding assistant. Source code, API keys, and cloud credentials exposed on developer machines. Supply chain Per-developer Critical Aug 2025
Shadow AI breach (composite)The IBM cohort case. Employees feeding sensitive data to personal ChatGPT or Copilot accounts. 20% of breached organisations report this shape. Shadow AI 1 in 5 organisations High Year-round, 2025
Live tracker
Every AI security incident we have documented — updated weekly.
Filterable by vector, sector, and year. Updated by the FireTail research team.
Open the breach tracker

The pattern is that the perimeter did not disappear. It mattered less. Most of these breaches went through trust relationships your organisation had already approved — an OAuth grant, a SaaS integration, a document the AI was authorised to read. Endpoint security and user-focused monitoring were not the right tools for any of them.

Reco · AI & Cloud Security Breaches: 2025 Year in Review
Chapter 03 Slowed down

EchoLeak is the most documented example of what a working AI exploit actually looks like.

Of the four featured incidents, EchoLeak is the one with a published CVE and a documented mechanism. Microsoft caught and patched it through coordinated disclosure, so there is no record of in-the-wild exploitation. The mechanism is still general: any AI tool that reads attachments, emails, or shared documents is reachable through the same shape. What follows is the breach in six steps, four seconds total.

Attack categories
Six families · one walkthrough live
MCP supply chain Coming
Tool poisoning · agent marketplace, 2026
API misconfiguration Coming
Unauthenticated endpoint · Trend Micro data
Shadow AI breach Coming
OAuth overreach · IBM Cost of Breach 2025
Training data poisoning Coming
Slow burn · model behaviour drift
Multi-agent cascade Coming
Agent-to-agent propagation · 2026 pattern
Prompt injection · live

Zero-click data exfiltration via indirect prompt injection

Based on CVE-2025-32711 "EchoLeak" — Microsoft 365 Copilot, 2025
1 / 6
Elapsed: T−72 hours · Human actions: 0
Chapter 04 The permission problem

Most AI breaches succeed not because the model failed but because of the permissions you granted it.

EchoLeak worked because Copilot had permission to read mail and permission to send it. The permissions were granted at deployment and never re-examined. That is the structural problem this chapter is about — not specific to agents, not specific to MCP, just the gap between how IAM teams think about access and how AI deployment teams think about it. Below: the same employee, side by side with the AI assistant deployed to support her.

Human employee

6 systems · scoped
Maya Chen
Senior counsel · 4 yrs
Legal drive
Read / write — legal/*
Outlook
Send / receive
Teams
Chat · calls
Workday
View HR · self only
Salesforce
Read only
Intranet
Read only
Scoped, audited quarterly

AI agent — same organisation

8 systems · broad OAuth
Legal AI Agent
Deployed March 2025
All SharePoint
Full tenant read
All mailboxes
Send on behalf of
Salesforce API
Read + write CRM
GitHub repos
All private repos
Jira / Confluence
All projects
Slack workspace
All channels + DMs
HR system API
Employee records
Finance ERP
Invoice processing
Broad OAuth · rarely audited · no least-privilege review

An agent with tool access and stored credentials is not the same risk profile as a chatbot. Most AI acceptable-use policies do not make that distinction. The result is that procurement, legal and security all signed off on "the same product" — except the product quietly became a different one.

IBM · Cost of a Data Breach 2025 · 63% of breached organisations had no AI governance policy

Why this looks normal to your SIEM

The security stack most enterprises run was built around an attacker who had to escalate to act. You log the escalations — that is most of what your SIEM does for a living. The agent does not escalate. The permissions were granted at deployment. The actions it takes are the actions a service account was approved to take, and your SIEM correctly does not alert on those.

So "lateral movement" has collapsed into a single tool call. The motion your defenders were trained to detect simply no longer happens. The agent does not move across systems. It already lives in all of them.

Three changes that close most of the gap

None of these are exotic. They are standard hygiene, applied to a layer most teams have not yet added to their threat model.

Scope every agent to least privilege. Treat each one like a service account that needs quarterly review. The permissions granted at deployment day are almost never re-examined; that is where the broad OAuth grants accumulate.
Log every tool call, not just the prompt. The breach timeline lives in the chain: which tool, which parameters, which return, which next action. If you are only logging conversation turns, you have a transcript — not an audit trail.
Alert on out-of-role access. A legal agent reading GitHub at 2 a.m. is the kind of thing that should fire. Until you write rules for cross-system access, your existing detections will not see it.
Chapter 05 Regulator

August 2, 2026 is the date the EU AI Act starts applying to high-risk AI systems across the bloc.

From that date, Articles 9 through 17 apply to providers of high-risk AI systems and Article 26 applies to deployers. Article 50 transparency rules apply too. Failure to comply carries fines up to €35M or 7% of global annual turnover. The data from the year before says most organisations are not ready, and the obligation that is most consistently missing is the one every other obligation rests on — an inventory of what AI you actually have.

Readiness by requirement

% of orgs meeting
Risk management system
34%
Data governance
41%
Technical documentation
48%
Human oversight mechanisms
29%
Logging & audit trail
38%
AI inventory
22%
Incident reporting
31%

Readiness by sector

% of orgs meeting
Financial services
54%
Healthcare & pharma
46%
Enterprise SaaS
38%
Government / public sector
28%
Manufacturing
21%
Source · FireTail Research, getaiactready.eu cohort, n=412

Inventory is the prerequisite for every other obligation in the Act. Solve it first. Risk management, oversight, logging, incident reporting — none of them function until you can produce a complete picture of every AI system in your organisation. It is also the highest-leverage compliance work available in the time you have left.

FireTail Research, 2026
77
days to August 2, 2026 — the date Articles 9 through 26 begin applying to high-risk AI systems.
Regulation 2024/1689
22%
of organisations in the FireTail cohort can produce a complete inventory of every AI system in production. The foundation every other obligation rests on.
FireTail Research · 2026
€35M
or 7% of global turnover, whichever is higher — the upper bound for non-compliance with high-risk AI requirements.
EU AI Act · Article 99
Chapter 06 Comparatives

Breach cost and peer ranking are the two numbers that actually move the AI security budget.

Is the exposure expensive enough to act on, and are we behind our peers? The table answers the first. The grid answers the second. A note on the risk-reduction column: percentages reflect FireTail's review of incidents where the control was present versus absent. They are floors, not promises. The point is the gap between vectors.

Threat vector Avg breach cost Primary control Effort Risk reduction
Prompt injection (indirect) $4.88M Input sanitisation layer + prompt shield before LLM ingestion Low · 2–4 wks
~82%
MCP supply chain poisoning $4.63M Internal MCP registry + version pinning + automated trust scoring Med · 4–8 wks
~74%
API misconfiguration / exposure $3.86M AI API gateway with auth enforcement, rate limiting, anomaly detection Low · 1–3 wks
~91%
Shadow AI / OAuth overreach $4.63M AI asset discovery + OAuth audit + agent-specific acceptable use policy Med · 3–6 wks
~67%
Training data poisoning $5.20M Data lineage tracking + pipeline integrity + behaviour monitoring High · 8–16 wks
~58%
Multi-agent cascade failure $6.10M+ Agent isolation boundaries + inter-agent message validation + kill-switch High · 12+ wks
~70%

The economic argument for AI runtime monitoring is, frankly, the easiest part of this report. IBM puts the saving at $1.9M per breach, with containment 40% faster. The harder part is that you cannot monitor what you have not yet inventoried. The order of operations matters more than the budget.

IBM · Cost of a Data Breach 2025
Sector benchmark

Where your sector sits

If your score is below your sector median, you are carrying above-average risk relative to your competitive set. That is the comparison your cyber insurer is making, and increasingly the one your regulator is too. The number you walk into the audit committee with is not your absolute score — it is the gap.

Financial services
54/ 100 median
AI inventory
Runtime monitoring
EU AI Act readiness
3rd quartile · ahead of median
Healthcare
46/ 100 median
AI inventory
Runtime monitoring
EU AI Act readiness
2nd quartile · at the median
Enterprise SaaS
38/ 100 median
AI inventory
Runtime monitoring
EU AI Act readiness
2nd quartile · below median
Government
28/ 100 median
AI inventory
Runtime monitoring
EU AI Act readiness
1st quartile · trailing

Source · FireTail Research, getaiactready.eu cohort · n=412

Chapter 07 Operations

The first ninety days of AI security work returns more than the next twelve months combined.

Twelve actions across three horizons. Horizon 1 is what the security team can start without procurement. Horizon 2 needs some new tooling. Horizon 3 is architecture work that takes a fiscal year. Run them in order — the first item tends to reveal the rest.

Chapter 08 The honest read

Five questions reveal where your AI security posture actually sits.

The questions are the ones FireTail uses in initial assessments — the same ones most security teams quietly skip when they review their own posture. Two minutes. No login. Run it again in ninety days and watch the trajectory.

AI Security Maturity Assessment
5 questions · 2 minutes · instant score

0

Chapter 09 · The board card

The report compressed to one page.

This is the artefact CISOs walk into the audit committee with. The numbers on the right are from the IBM 2025 cohort — the canonical reference. Read out the questions, point at the deadline, leave the briefing with the directors. The detailed case is in the eight chapters above.

A note on how FireTail thinks about this

The argument of this report is that inventory is the lever — the prerequisite for every other control, the foundation under every other Article. FireTail builds inventories. The platform discovers every AI endpoint, agent and integration across your stack — authenticated or not, documented or not — and produces the map every other control depends on. Fifteen minutes from connect to coverage. The rest of what we do is built on top of that.