Guides / SonicWall

Find the AI tools your staff use, from SonicWall syslog

SonicWall appliances guard thousands of professional-services offices, the exact environments where a dictation app with silent training terms is a genuine liability. The syslog those boxes already emit is the audit's raw material.

Layer: firewall / content filter
Identity: usr= when the SSO agent runs
Format: key=value syslog
Typical shop: one IT person, many obligations
Run a free audit on a SonicWall export
The one-admin reality

An audit built for the IT department of one

SonicWall shops usually mean a single admin carrying security alongside everything else. This loop is sized accordingly.

One hour, total

Pull a syslog slice from wherever it already lands, upload, read. No new infrastructure, no agents on partner laptops, no change-control meeting.

Evidence, not opinions

When a managing partner asks "are we exposed?", the answer becomes a dated PDF instead of a hedge. That difference is the whole product.

Reuses what exists

The syslog server you stood up for compliance years ago finally earns its keep: the audit reads raw key=value lines exactly as SonicWall writes them.

The lines

SonicWall syslog, decoded for the audit

Web-filter and connection events carry the four values that matter. A realistic slice, noise dimmed:

id=firewall sn=18B1690XXXXX time="2026-09-16 10:12:44" fw=203.0.113.10 pri=6 c=1024 m=97 usr="LAW\h.brenner" src=10.8.2.41:52104:X0 dst=104.18.7.90:443:X1 dstname=speechnotes.example arg=/dictate Category="Business" usr="LAW\h.brenner" src=10.8.2.41:52161:X0 dstname=api.speechnotes.example arg=/v1/stream usr="LAW\c.mensah" src=10.8.2.57:49330:X0 dstname=chatgpt.com arg=/ Category="Information Technology" usr="" src=10.8.9.12:51780:X0 dstname=quillbot.com arg=/summarize Category="Education"

The keys that matter

  • dstname= the hostname, the match key. arg= paths are dropped at parse time.
  • usr= the SSO-agent identity. Empty usr= lines fall back to the src IP.
  • time= keeps windows honest across appliance and collector clocks.
  • Category= the CFS label, kept only for the before/after comparison it invites.

What these four lines already say

  • An associate is streaming dictation to an AI service, app and API host both visible.
  • The CFS categories call it "Business" and "Education". Category filtering was structurally never going to flag it.
  • The empty-usr= line on the guest segment still counts as a source.
Worked scenario

A law firm and the dictation problem

A 45-lawyer firm, one SonicWall, one IT manager, SSO agent on the domain. The trigger was an e-discovery vendor asking about AI use. Sample data, brief format.

Matter: unknown AI exposure

The IT manager exports 30 days of web-filter syslog: 150,000 lines. Upload, match, 16 AI tools found. Twelve are unremarkable. Four are dictation or summarization tools, which in a law firm means client-confidential material is the input.

Findings of fact

The per-user table shows two dictation apps concentrated among four associates, plus a summarizer being fed briefs by a paralegal team. One tool trains unless opted out; two say nothing about training at all.

Our wider terms analysis says silence is the norm, 85.5% of AI tools state nothing about training. Law firms cannot file "the norm" as due diligence.

Holding

The firm sanctions one legal-grade dictation vendor after a terms review, blocks the consumer tools at the appliance, and adds an AI clause to the staff handbook.

The audit PDF goes into the e-discovery vendor response and the professional-indemnity renewal file, both of which had asked the same question.

Costs

One $99 report, roughly ninety minutes of the IT manager's time, and a partner meeting that ran short because the tiles page needed no explanation.

SSO agent states

How attribution behaves in each deployment state

The usr= field depends on the SSO agent. Every state still yields a valid inventory.

Agent on, domain users

Names on every line for domain-joined machines. The per-user table supports the exact conversation the law firm needed.

Agent on, gaps

VPN users, terminal servers and non-domain devices may log empty usr=. Their src IPs still count as sources, so nothing disappears.

No agent at all

Whole export attributes by IP. Pair the report with your DHCP reservations and the reading is nearly as sharp.

Output

What the report returns for a SonicWall export

Report sectionWith SonicWall syslog
Summary tilesFull totals: tools, high-risk, training exposure, abusive-purpose, unsanctioned.
Tool tableEvery matched dstname with category, risk, sovereignty and dated training verdicts.
Per-user tableusr= identities plus IP-only sources, side by side.
Sanctioned splitThe firm's approved list against observed use, the professional-duty number.
Blocked evidenceCFS-blocked events stay visible as enforcement proof if included in the export.
CSV + PDFFull-report tier: the PDF for partners and insurers, the CSV for the risk register.

Uploads are read once and discarded; arg= paths never reach the report. Reports live 90 days and can be deleted earlier, which belongs in the firm's own data-handling note.

Appliance edges

SonicWall details worth knowing

CFS categories vs the register

  • The scenario's dictation tools sat in "Business". Category filters protect against known bad, not unknown AI.
  • Keep the Category= values in your export: the mismatch table is your budget justification.

DPI-SSL is optional here

  • Without deep inspection, dstname= still carries the hostname from SNI, which is all the matching needs.
  • No need to expand decryption, and no client-privilege questions from doing so.

Which events to slice

  • Web-filter and connection events carry hostnames; system and VPN chatter does not.
  • Filtering the slice by those message classes keeps a month lean.

Size expectations

  • A 50-person office's month lands well inside the 2,000,000-line / 25 MB full-report cap.
  • The free preview's 5,000 lines suit a spot-check afternoon, not a monthly baseline.

See the withheld-rows preview for yourself

The sample evidence pack shows the full document; the free preview shows the same with masked rows you can count. Both are one click from here.

Open the sample PDF
Professional duties

Why professional-services firms audit first

The law firm's pattern generalizes to accountants, advisors and consultancies: confidentiality is the product, and AI tools are where it leaks quietly.

Privilege and confidentiality

Dictation, summarization and drafting tools all ingest client material by design. The report's training verdicts, dated, are the due-diligence artifact for that exposure.

Insurers now ask

Professional-indemnity questionnaires increasingly include AI-usage questions. A log-based inventory answers in a sentence plus an attachment.

Clients ask sooner

E-discovery vendors, corporate clients and outside counsel guidelines all probe AI use now. Firms with a quarterly audit answer from the file, not from memory. The evidence-chain angle continues on the compliance officer page.

Quarterly is the professional-services cadence: one report per quarter, filed with the matter-management or risk system. Plans and packs are on the pricing page.

FAQ

SonicWall export questions

Which SonicWall log events does the audit need?

Web-filter and connection events carrying dstname=, as raw key=value syslog. No reformatting; collector slices upload as-is.

We do not run the SSO agent. Is the audit still useful?

Yes. Attribution falls back to source IPs, and the tool inventory is identical. DHCP reservations make the IP table readable.

Does DPI-SSL need to be on?

No. Hostnames arrive via SNI in dstname= regardless, and the audit needs nothing deeper.

Can the export include blocked traffic?

It should. Blocked AI events document enforcement, and attempt volume against blocks signals demand for a sanctioned alternative.

Is client-confidential data exposed by uploading?

The audit reads hostnames, identities and timestamps; arg= paths are dropped at parse. The upload is discarded after the run and reports are deletable early.

How fast can one admin do this end to end?

The law firm took about ninety minutes including reading time. The export is the slow part; matching takes minutes.

The category mismatch table

Where CFS filed the tools the audit found

From the law firm's report, sample data. This one table justified the audit to the partners in a glance.

Tool foundCFS category at the timeAudit verdict
Dictation app ABusinessBlock: trains unless opted out, audio input
SummarizerEducationControl: terms silent on training
ChatbotInformation TechnologyControl: enterprise tier acceptable
Legal-grade dictationBusinessAllow: contractual no-training terms

Same CFS category, opposite verdicts, twice. Category filtering cannot make that distinction; per-tool verdicts with dated terms checks can.

Handbook clause

The three-sentence AI clause the audit supports

The firm's handbook addition was short because the audit does the enforcement measuring. A pattern worth borrowing:

The rule

"Client material may only be entered into tools on the approved AI list." One sentence, no technology lecture, enforceable because the list exists.

The path

"Requests for new tools go to IT and are answered within two business days." Fast sanctioning is what keeps associates from not asking.

The check

"Usage is reviewed quarterly from network logs." The audit is the review; the sanctioned split is the metric; nobody is surprised later.

Building the approved list itself, tier by tier, is covered on sanctioned vs unsanctioned AI.

Answer the partner's question with a PDF

Thirty days of syslog, one upload, and "are we exposed?" gets a dated, specific, fileable answer.

Run the free audit