Proof of Concept

See exactly what your stack let through.

A StrongestLayer POC is not a detection bake-off. It is a diagnostic of your actual threat surface. Deploy in 15 minutes, observe for one week, and walk away with a 44-subtype threat map, a stacked-technique analysis, and a FAIR-grade risk number your CFO can use.

15 min to deploy Zero training period Zero user disruption 1 week to a full report
Step 1: Deployment

Fifteen minutes. Zero training period.

Three steps: authenticate, pull users, configure settings. The process is the same on Microsoft 365 or Google Workspace. Read-only access, no MX changes, no new agents. Protection starts at full strength on Day 1. No baseline learning period. No soak time before the model is useful.

01
Authenticate
Global Admin (Microsoft 365) or Super Admin (Google Workspace) grants read-only access via the Integration Hub. Microsoft: three Graph scopes, User.Read.All, Directory.Read.All, Group.Read.All. Google: equivalent read-only Directory API scopes. Revocable in one click from your admin console.
2 to 3 minutes
02
Pull users
Directory sync pulls user profiles: first name, last name, email, department, title. That is all. Passwords, mail content, calendar, and file storage are never accessed. Sync completes automatically.
2 to 5 minutes
03
Configure settings
Enable email analysis for the organization. Optional during the POC: deploy Inbox Advisor to users, or leave it off so users see nothing change. The reasoning engine is live the moment you save.
3 to 5 minutes
Why this matters
No training period is an architecture proof, not a convenience.
Baseline-behavioral detection tools require a training period because their models learn what "normal" looks like in your tenant, then flag deviation. If a targeted attack lands during the baseline window, it shapes the model's definition of normal. StrongestLayer reasons about intent and technique, not deviation from norm, so it needs zero baseline. A 15-minute deploy is a direct consequence of reasoning-based detection, not a UX polish. Ask your incumbent why they need ten days. The answer tells you their architecture.
Step 2: The POC itself

POC mode means zero disruption.

For the observation period, StrongestLayer runs in read-only mode alongside your existing SEG. We see what you see. We generate detections in parallel. Nothing is quarantined, nothing is released, nothing is changed. Your users never notice. Your admins never touch it.

What happens during the week

Email analysis runs on your inboxes via read-only API access to Microsoft 365 or Google Workspace. The reasoning engine records a verdict for every message. Your existing mail path, your third-party gateway if you run one, and your native Microsoft Defender or Google Workspace protection all continue to own delivery. Inbox Advisor stays off for the POC so users see nothing change.

  • No mail flow changes. MX records unchanged. Third-party SEG, if present, still front-line. Microsoft or Google SEG still enforces delivery. Zero risk of broken delivery.
  • Read-only access. Directory scopes only: User, Directory, Group. Never read: passwords, mail content, calendar, file storage, personal content.
  • No admin workload. No tuning, no whitelisting, no false-positive triage. Set it, then forget it until the report.
  • No user impact. Inbox Advisor stays off. No banners, no prompts, no warnings. Users experience their inbox exactly as they do today.
  • Revocable in one click. Remove the integration from your admin console at any time. All API tokens revoke immediately.
Access pattern during the POC window
Inbound mail
Internet
3rd-party SEG
Proofpoint, Mimecast, etc.
Microsoft/Google
Owns delivery
User inbox
No banners, no prompts
StrongestLayer
Reasoning engine pulls directory data via read-only API. Records a verdict for every message. Never touches mail content storage, calendar, files, or credentials.
Read-only. Out-of-band. No risk to delivery. Your existing stack continues to own the delivery decision for the duration of the POC. StrongestLayer sits beside the mail provider, not in front of it, so a verdict disagreement never becomes a business continuity issue.
Step 3: The report

The report is the education.

At the end of the POC period, you get a deliverable your team can walk into a board meeting with. Four sections. Each one changes how your team thinks about email threats, not just how many they caught.

Most POC reports show you a detection count. Ours shows you a new map.

A tally of "things your incumbent missed" is a commodity deliverable. Every email security vendor can produce one. The StrongestLayer POC report is built to answer a harder question: where is your real threat surface, and why can your current stack not see most of it? The sections below are rebuilt from a real customer engagement, illustrative counts preserved.

118
Threats reached user inboxes
86%
Tier 3+ (Advanced detection required)
17/30
Matrix cells populated
$3.9M
Annualized risk exposure

The 44-subtype threat matrix

5 tiers (sophistication) × 6 categories (attack type). Cell = subtype count observed during the POC.
Category
T1 Commodity
T2 Emerging
T3 Established
T4 Advanced
T5 Apex
Total
Credential HarvestingPhishing pages, AiTM, OAuth, QR
3M365 Lures
32Lookalike
26Multi-Stage
2Homograph
63
Business Email CompromiseExec, vendor, thread hijack
1Vendor Pay
3Phone Callback
4
Malware DeliveryMacro, HTML smuggle, cloud-hosted
2Cloud Hosted
2
Brand ImpersonationDocuSign, Microsoft, banking
2DocuSign
16Zero-Day Pattern
14Redirect Chain
1Homoglyph
33
Social EngineeringAdvance fee, calendar hijack
1Advance Fee
1Calendar Hijack
2
Consumer ScamAdvance fee, romance, health
4Advance Fee
6Health Scam
2Romance
2Get-Rich-Quick
14
Column total
5
11
51
48
3
118
All 118 threats bypassed the existing gateway. 86% were Tier 3 or higher, including three Apex threats using Unicode homograph evasion. Traditional SEGs plateau at Tier 2 detection. The 102 threats at T3+ would have remained undetected without reasoning-based analysis. The matrix teaches the buyer what their real surface looks like, not just what one vendor caught.
Report section: Technique stacking

One attack. Five techniques. Four defense layers defeated.

Across 5,000 catalogued detections, 56.8% of attacks use four or more evasion techniques simultaneously, averaging 4.11 per detection. Each technique defeats a different defense layer. No single rule catches the full chain. The report shows you the chains, not just the endpoints.

T5 APEX

SUMUP Cyrillic Homograph, 5 techniques across 4 layers

Unicode Homograph
Defeats Layer 1
Authentication
SUMUP Impersonation
Defeats Layer 4
Content / NLP
Multi-hop Redirect
Defeats Layer 3
URL Sandbox
False Urgency
Defeats Layer 5
Human Judgment
German Localization
Defeats Layer 4
EN-trained NLP

Cyrillic characters bypass keyword filters at the encoding level. No signature or ML model trained on Latin text detects the mismatch. Each subsequent technique defeats a different fallback layer. The stack is not accidental, it is engineered to cover every plausible defense a mid-market SEG can raise.

T4 ADVANCED

DocuSign S3 Payload, 4 techniques, multi-vector attack

DocuSign Impersonation
Defeats Layers 1 + 4
Auth + NLP
Amazon S3 Payload
Defeats Layer 2
Reputation (trusted infra)
CAPTCHA Gate
Defeats Layer 3
URL Sandbox
Clipboard Hijacking
Defeats Layer 5
User sees nothing
 
 

Cloud-hosted payload on Amazon S3 is trusted infrastructure; reputation engines allowlist amazonaws.com. CAPTCHA prevents sandbox detonation. Clipboard hijacking delivers malware without any visible download. Four techniques, zero rule overlap with the SUMUP variant.

56.8%
Of attacks use 3+ techniques
4.11
Average techniques per detection
0
Rule overlap across variant families
Report section: Financial impact

Risk in dollars. Methodology your CFO already trusts.

The ROI page uses the FAIR Institute methodology: probability-adjusted loss events, calibrated to your tenant size and observed threat mix. This is the same framework used by Fortune 500 risk committees and cited in Verizon DBIR and IBM Cost of a Data Breach. Not a vendor estimate. Not a guess.

$11
Return per $1 invested, median case
$3.9M
Annualized risk exposure (median)
$360K
Annual investment (30K seats × $12/user)
$3.5M
Net annual benefit
Risk exposure range, conservative to aggressive
$2.3M
Conservative
$3.9M
Median (base case)
$5.4M
Aggressive
POC threats detected118
Annualization factor (1-week POC)52×
Projected annual threats6,136
User engagement rate (Verizon DBIR)10%
Compromise rate per engaged threat (FAIR)3.5%
Annualized risk exposure (median)$3.9M

Sources: FAIR (Factor Analysis of Information Risk) methodology. Category success rates derived from FBI IC3 2024 Report (BEC 5.4%), Sophos State of Ransomware 2024 (Malware 4.1%), Verizon DBIR 2024 (Credential Harvest 3.2%), IBM Cost of a Data Breach Report 2024. Base loss values reflect industry-standard incident cost estimates scaled by evasion sophistication tier. Conservative estimate applies 0.6x modifier; aggressive applies 1.4x modifier.

The takeaway

What your team walks away with.

Even if you do not buy StrongestLayer, the POC changes what your team knows. Four things, each of which outlives the engagement.

01: Mental model

A 44-subtype map of your real threat surface

Most security programs still triage to two buckets: phishing and BEC. Our report shows you the 6 categories and 44 subtypes your inbound traffic actually contains, and which cells of the matrix are lit up in your tenant.

02: Technique fingerprint

Evidence that no single rule catches the chain

The stacking analysis shows which attack families chain 4 or more techniques together, and which layers each technique defeats. This becomes a benchmark your team can use against any vendor, now or in the future.

03: CFO-grade number

FAIR-methodology risk exposure in dollars

A probability-adjusted annualized risk number, with conservative, median, and aggressive bounds. Built on industry-standard loss data. Not "trust us," not a vendor estimate. The same framework your risk committee already uses.

04: Priority list

A remediation sequence ranked by evasion sophistication

Top threats ranked by Apex and Advanced tier first, with MITRE mappings, indicators, and the specific defense layer each one evades. Your team can take this into a sprint regardless of the vendor decision.

Ready to see your own numbers?

Fifteen minutes of your admin's time, one week of observation, a full report your team can walk into a board meeting with. No commitment past the POC window.

Typical engagement: 15 min deploy, 7 days observation, Report delivered in week 2