Discovering, governing, and monetizing AI use in Microsoft SMB tenants.
The question isn’t whether your customer’s employees are using AI. It’s whether you’re the one who found out first.
By the time a partner sits down to talk about a Copilot license, the AI conversation has usually already started — just not with IT in the room. Employees are pasting client emails into free ChatGPT accounts, dropping proprietary code into personal Cursor or Claude sessions, and running meeting recordings through consumer transcription apps. None of it shows up in a firewall log, because every one of those tools is a trusted TLS endpoint that every URL filter waves through as "productivity."
Lead with the plumbing. Permissions, Conditional Access, MFA. Technically correct, and it asks a prospect to get excited about your compliance checklist before they've agreed there's a problem worth solving.
Lead with Shadow AI discovery. It starts from something the prospect already privately suspects is true about their own business, and offers to make it visible — on their terms, before someone else finds it for them.
Governance and consumption are the same problem, not two separate ones. Whatever AI surface a customer eventually turns on — Copilot, a sanctioned ChatGPT Enterprise seat, an internal agent — inherits whatever permissions and data hygiene already exist in the tenant. An assessment that surfaces unsanctioned AI use is also the exact prerequisite work for making any AI purchase pay off. Governance readiness is consumption readiness.
The figures above come from a mix of primary research (Gallup, McKinsey, PagerDuty) and secondary community synthesis. Use the primary-sourced ones with confidence in an executive conversation; treat the synthesized and benchmark figures as directional, and always re-check the current edition of any cited report before quoting an exact number to a customer.
The following categories show up on nearly every SMB discovery scan, well before any purpose-built discovery tool is even switched on. Treat this as the starter checklist for the conversation, not an exhaustive taxonomy.
| Category | Common tools | Typical exposure |
|---|---|---|
| General LLM assistants | ChatGPT, Claude, Gemini, Copilot Chat (personal), Perplexity, Meta AI | Client emails, contracts, HR data pasted into personal accounts |
| AI coding assistants | GitHub Copilot (personal), Cursor, Windsurf, Codeium, Claude Code | Proprietary source code, secrets, connection strings |
| Meeting notes / recording | Otter.ai, Fireflies.ai, Fathom, Read.ai, Zoom AI Companion | Confidential meeting content, financials, personnel decisions |
| Writing / editing | Grammarly (personal), Wordtune, QuillBot | Legal, HR, client-facing content routed through personal accounts |
| Productivity / knowledge | Notion AI, ClickUp AI, Coda AI, Mem | Company docs, project plans, roadmaps |
| Design / marketing | Canva AI, Adobe Firefly, Midjourney, Ideogram | Brand assets, client campaigns, PII embedded in image inputs |
| Sales / RevOps AI | Gong, Apollo.io AI, Clay, Lavender | Full CRM records synced to an unmanaged AI vendor |
| Enterprise search / RAG | Glean, Guru, You.com Enterprise, Dust | Broad SharePoint/Drive access under a personal or loosely-scoped grant |
| Web-scraping AI | Diffbot, Browse.ai, Bardeen, PhantomBuster | Undocumented data-extraction pipelines with no security review |
| Voice / avatar AI | ElevenLabs, HeyGen, Synthesia, Descript | Executive voice cloning, deepfake and impersonation risk |
Microsoft publishes a formal deployment model in Microsoft Purview called Prevent data leak to shadow AI. It's worth anchoring to because it's the reference architecture a customer's auditors, cyber insurer, and Microsoft account team already expect to see reflected in a serious engagement.
Defender for Cloud Apps has had Cloud Discovery for years; the AI-specific category plus the Entra Internet Access network-layer signal together approximate visibility partners historically had to buy from a dedicated CASB vendor.
Two enforcement paths: Conditional Access plus a Defender for Cloud Apps session policy for browser sessions, or Intune / managed-browser enforcement at the device level. Community best practice: don't block everything on day one — close the highest-risk consumer apps first and monitor the rest while the sanctioning conversation happens.
Purview DLP now has an AI-app category as a first-class condition. A reasonable starter rule: location = devices onboarded to Defender for Endpoint; condition = a Confidential label or a sensitive info type; action = block upload to the AI-app category; mode = audit first for two weeks, then enforce.
Configure Audit (Premium) retention — a full year, not the 180-day default — before an incident, not after. Communication Compliance can capture AI interaction logs where the tool integrates with Microsoft's information-protection stack; for third-party tools, coverage depends entirely on the vendor.
Microsoft MVP Nikki Chapple frames the same architecture as two decisions, and it's the cleanest version of this story for an executive conversation:
Decide where data is allowed to go. This is Defender for Cloud Apps, Entra, and Intune working together to build the sanctioned/monitored/blocked list.
Decide what data is allowed to leave, regardless of destination. This is Purview DLP, sensitivity labels, and DSPM for AI.
Microsoft provides the control plane. The customer's organization defines the control model. Your job as the partner is to help them write that model down — nobody else in the room is going to.
Before quoting a Shadow AI assessment, run the same five-capability audit on your own tenant first. Selling discovery you haven't personally survived is the fastest way to under-scope a SOW — dual-track it: run the snapshot on yourself, then sell it to customers.
| # | Capability | Why the assessment needs it | Typical SMB-partner starting point |
|---|---|---|---|
| 1 | Identity and access visibility | Who is using what, under which account | Usually already have it — Entra ID P1 ships in Business Premium |
| 2 | Application discovery | The engine that finds unsanctioned AI tools | Confirm — needs Defender for Cloud Apps |
| 3 | Data classification and DLP | To say what data went where — this is what sells the remediation | Partial — sensitivity labels ship in Business Premium; deeper DLP needs the Purview add-on |
| 4 | Policy authoring | Write the AI Acceptable Use Policy and the DLP rules | Have the skill — the AUP itself is almost always missing |
| 5 | Ongoing monitoring | The step that converts a one-time report into recurring revenue | Confirm — needs the licensing in place at the customer or partner-tenant level |
Frame Shadow AI as an ongoing exposure, not a one-time compliance exercise. Get a named executive sponsor and confirm scope, seat count, and data sensitivity.
Enable Cloud Discovery, Entra Internet Access, and DSPM for AI. Deploy the Purview browser extension to a pilot device group. Onboard endpoints to Defender for Endpoint if not already done.
Passive collection, no user disruption. A 14-day minimum is the community standard; extend to 21 days if usage is unusually cyclical (month-end close, seasonal work).
Categorize discovered apps, attribute usage by user and department, quantify sensitive-data exposure, map to risk ratings, and prioritize remediation.
A focused deck: exposure heatmap, top apps, top exposed users, business impact, remediation roadmap, licensing gaps, and the next-step SOW menu.
Fixed-fee AUP, oversharing remediation, sanctioned-app onboarding, and the Managed AI Governance retainer.
| Phase | Elapsed time | Partner effort | Skills required |
|---|---|---|---|
| Executive intake | 1–2 days | ~2 hrs | AI-fluent seller / vCIO |
| Discovery tool configuration | 2–5 days | 4–8 hrs | M365 admin comfortable in Defender + Purview |
| Telemetry collection | 14–21 days | Passive | — |
| Analysis + report | 3–5 days | 8–16 hrs | Security analyst / consultant |
| Executive readout | 1 day | 2 hrs + prep | vCIO / AI practice lead |
| Total | 4–6 weeks | 16–28 hrs | 1 practice lead + 1 M365 admin |
Community-reported SMB assessment pricing clusters around $1,500–$7,500 fixed fee for a 20–200 seat customer, climbing past $15K for regulated or mid-market accounts. Independent MSP benchmarking (Cynomi) puts the share of assessment clients who convert to a follow-on retainer at 50%+ — and that conversion rate rises further when the assessment surfaces a specific, named risk rather than a generic score. Treat both figures as planning inputs, not guarantees, and validate against your own delivery cost.
Departments crossed with the sensitivity of the data leaving them.
Top apps by user count, top apps by risk.
Highest-risk individuals, with a separate executive-only version.
What any future AI product would be able to see on day one.
Current state versus recommended state, mapped to the Suite add-ons in Section 6.
Quick wins (0–30 days), foundational (30–90 days), strategic (90+ days).
Customized to the sanctioned tools and the customer's vertical, not boilerplate.
Priced by service line — see Section 10.
This is a composite, representative walkthrough built to teach the pattern of a common SMB deal — the numbers are illustrative of what a real 14-day scan at this seat count typically turns up, not a specific named customer's data.
This is the objection that stalls the deal if it's handled as a forklift upgrade to E5. It isn't one. Microsoft sells a Defender Suite and a Purview Suite specifically scoped to Business Premium, and the two can be bundled at a meaningful discount when purchased alongside Copilot. The pitch is not "buy more Microsoft" — it's "the tool that powers this assessment is the same tool that governs Copilot after it's turned on."
You're about to spend real money on 50 Copilot seats. Before you turn it on, we need to know two things: what your team is already pasting into ChatGPT and Claude today, and whether Copilot's own grounding is going to surface data your labels weren't ready for. We do that with a four-week Shadow AI and Copilot Readiness Assessment. The Purview Suite add-on we'll turn on for those 50 seats does triple duty — it powers this assessment now, it governs Copilot once it's live, and it gives you your first documented AI compliance evidence ahead of your next cyber-insurance renewal.
The customer doesn't get to flip on Copilot until the oversharing findings are remediated — not because the partner is being obstructive, but because AI output layered on top of ungoverned data will erode the customer's confidence in the product itself before it ever has a chance to prove its value. Show the value of getting this right first; then sell the protection that keeps it right.
"You need to upgrade to E5." A Business Premium customer hears this as "double my Microsoft bill," and the deal stalls right there.
Same Copilot engine, right governance stack. The Copilot AI experience is the same on Business Premium and E5 — the difference is the governance tooling in the base SKU, not the AI feature set. Business Premium + Copilot + the Suite add-ons gets there without the forklift.
Exact feature coverage between the Business Premium Suite add-ons and full E5 Compliance shifts as Microsoft updates its licensing tables. Confirm the specific capability — especially anything DSPM- or Copilot-assessment-related — against the current Microsoft licensing comparison before a proposal goes out.
The instinct after a discovery scan is to block everything unsanctioned. Resist it. Blocking without sanctioning drives usage further underground — onto personal devices and personal networks the partner has zero visibility into — and it burns the goodwill that made the customer receptive to the conversation in the first place. The credible position is narrower and more durable: bring the tools people already trust under management, rather than take them away.
If your team relies on Claude or ChatGPT, we're not taking it away. We're bringing it inside the fence — enterprise SSO, SCIM provisioning, an admin console, a signed data processing agreement, and audit logging. Once that's in place, the tool your team already trusts is no longer a liability on your books; it's an asset with a paper trail.
Move them to an enterprise tier — Claude Enterprise, ChatGPT Enterprise, or another vendor's enterprise plan, matched to their preference.
SSO via SAML, SCIM for lifecycle management, and an admin console configured for data retention, model access, and audit export.
The browser extension, DLP AI-category policies, and Defender for Cloud Apps monitoring — the same tools, pointed at a different sanctioned destination.
Monthly reporting on usage, cost, and DLP incidents under the same Managed AI Governance retainer described in Section 10.
Keep this as a living document — expand it quarterly. Pricing and SSO/SCIM support change often enough that this table intentionally omits dollar figures; confirm current terms directly with each vendor before quoting a sanctioned-app onboarding SOW.
| # | App | Category | Entra SSO | SCIM | Admin console | Notable data-boundary / governance |
|---|---|---|---|---|---|---|
| 1 | OpenAI ChatGPT Enterprise | LLM assistant | ✅ SAML | ✅ | Global Admin console | Data excluded from training by default; SOC 2 |
| 2 | OpenAI ChatGPT Business | LLM assistant | ✅ SAML (Business tier) | ✅ | Workspace admin | Data-exclusion by default |
| 3 | Anthropic Claude Enterprise | LLM assistant | ✅ SAML | ✅ (JIT/SCIM) | Console organizations | Audit logs, custom retention, RBAC |
| 4 | Anthropic Claude Team | LLM assistant | ✅ SAML | Partial | Team admin | Smaller-org tier |
| 5 | Google Gemini Enterprise | LLM assistant + agents | ✅ via IdP config | ✅ | Google Cloud console | Separate token/compute billing |
| 6 | Gemini in Google Workspace | Assistant | ✅ (Workspace SSO) | ✅ | Workspace admin | Bundled in Business Standard/Plus/Enterprise plans |
| 7 | Perplexity Enterprise Pro | Research / search | ✅ SAML | ✅ | Enterprise admin | SOC 2 |
| 8 | GitHub Copilot Business / Enterprise | AI coding | ✅ SAML via GitHub Enterprise Cloud | ✅ | GitHub org / enterprise settings | IP indemnification (Business+); public-code filter |
| 9 | Cursor Teams / Enterprise | AI coding | ✅ SAML 2.0 | Partial | Cursor org | Privacy Mode retention controls |
| 10 | Codeium / Windsurf | AI coding | ✅ SAML (Teams/Enterprise) | ✅ | Windsurf admin | Codebase-indexing scope configurable per repo |
| 11 | Glean | Enterprise search + assistant | ✅ SAML/OIDC | ✅ | Glean workspace admin | Permissions-aware connectors; tenant data boundary maintained |
| 12 | Notion AI | Productivity | ✅ SAML | ✅ | Notion workspace admin | DPA available; enterprise data segregation |
| 13 | Grammarly Business / Enterprise | Writing | ✅ SAML | ✅ | Grammarly admin console | Enterprise data doesn't train models |
| 14 | Otter.ai Enterprise | Meeting AI | ✅ SAML | ✅ | Otter admin | Meeting retention and sharing controls |
| 15 | Fireflies.ai Business | Meeting AI | ✅ SAML | Partial | Fireflies admin | Storage-region controls |
| 16 | Zoom AI Companion | Meeting AI | ✅ Zoom SSO / Entra | ✅ | Zoom admin | Off by default; no third-party training on customer data |
| 17 | Read.ai / Fathom Business | Meeting AI | Varies; some SAML | Partial | Vendor admin | Consent / recording notifications |
| 18 | Copilot Studio / custom agents | Agent platform | ✅ via Microsoft 365 identity | N/A | Power Platform admin center | Full Purview + Entra governance, natively |
| 19 | Microsoft Copilot Chat (free) | LLM assistant | ✅ Entra | N/A | M365 admin center | Enterprise Data Protection at the commercial data boundary |
| 20 | Canva Enterprise (Magic AI) | Design | ✅ SAML | ✅ | Canva admin | Enterprise-wide brand kit + admin controls |
| 21 | Adobe Firefly for Enterprise | Design | ✅ SAML | ✅ | Adobe Admin Console | Commercial-safe training-data statement |
| 22 | Midjourney | Design | ⚠️ Discord-based, limited SSO | ⚠️ | Server-level moderation | Recommend blocking for regulated data |
| 23 | HeyGen / Synthesia Enterprise | Video / avatar | ✅ SAML | Partial | Vendor admin | Voice-cloning consent + retention controls |
| 24 | ElevenLabs Enterprise | Voice AI | ✅ SAML | Partial | Vendor admin | Voice-clone authorization controls |
| 25 | Gong / Clari Copilot | Sales AI | ✅ SAML | ✅ | Gong / Clari admin | CRM data-boundary controls |
| 26 | Apollo.io AI | Sales AI | ✅ Growth/Enterprise | Partial | Apollo admin | CRM sync scope controls |
| 27 | Clay | Data enrichment AI | ✅ SAML (Enterprise) | Partial | Clay workspace | Multi-source enrichment audit trail |
| 28 | Jasper Enterprise | Marketing AI | ✅ SAML | Partial | Jasper admin | Brand voice + guardrails |
| 29 | Writer Enterprise | Enterprise gen-AI | ✅ SAML | ✅ | Writer admin | Custom private models; data isolation |
| 30 | Mistral Le Chat Enterprise | LLM assistant | ✅ SAML | Partial | Mistral admin | EU-data-residency option |
For each partner-sanctioned app, the onboarding pattern is the same six steps:
Microsoft's partner-facing propensity tooling in Partner Center — surfaced through the AI Business Solutions & Security Partner Experience (ASPX) — exposes weekly-refreshed, per-tenant signals that map directly onto a Shadow AI conversation. If your organization has Partner Center access, these are worth building into a weekly call-list routine rather than treating as background reporting.
| Signal | What it tells the partner | The Shadow AI conversation move |
|---|---|---|
| Free Copilot Chat MAU (unlicensed) | Users on free Copilot Chat this month | These users are proving demand — and are very likely also using ChatGPT/Claude under personal accounts. Sell the assessment. |
| Free-to-paid whitespace % | Ratio of free monthly active users to paid | A high ratio signals both Copilot license upsell potential and Shadow AI exposure |
| Copilot-eligible seats | The Copilot-eligible base at the tenant | Total addressable seats for both the license conversation and the governance retainer |
| Users hitting free-tier usage limits | Free-tier users being throttled | The strongest demand signal available — "your own users hitting a wall" is the opener |
| Adoption status | An estimate of the tenant's Copilot health | A "failure to adopt" flag is usually a governance problem, not a licensing one — an assessment opportunity |
| Data security maturity | A tiered estimate from Minimal to Advanced-Healthy | The sharpest single opener — Minimal and Limited-Unhealthy tenants are the top of the weekly call list |
Rank tenants by a blend of (a) users hitting free-tier usage limits, (b) unlicensed Copilot Chat activity, and (c) a weak data-security maturity signal. Take the top 15–20 for the week's call list. The customer's own numbers make the strongest possible opener: "You have people already hitting a usage ceiling on the free tier. They're trying to do AI work today — the only question is whether it's happening inside your tenant or inside their personal accounts."
The assessment is the land. What makes the practice durable is that every rung above it has a natural trigger from the one below — nothing has to be sold cold. Pricing ranges below are editorial estimates meant as a starting point; test them against your own cost model before quoting.
The land. A discovery-led engagement that converts into a specific finding ("here are the 14 AI tools your staff used last month, and here is the client data that left your environment") rather than a generic audit pitch.
Sold immediately after the assessment. A customized policy referencing the sanctioned tools, prohibited data classes, and — critically — an enforcement design. Never sell a policy without a control attached to it.
The recurring engine: ongoing enforcement, policy cascade, blocking of unsanctioned tools, identity-gated access, and a monthly usage/risk report the customer can hand to an auditor or insurer. Best packaged as a tier on an existing MSP contract rather than a standalone SKU.
Triggered by a cyber-renewal date or an enterprise customer's security questionnaire. "If your application asks whether employees share data with AI systems, right now the honest answer is 'we don't know' — here's how we get you an answer you can support."
For firms that need security leadership but can't justify a full-time hire. The MSP market's own data backs the trend: Cynomi's 2025 benchmark found the share of MSPs and MSSPs offering vCISO services rose from 21% to 67% year over year.
Short, role-based modules on what not to paste into public AI tools, with completion tracking. "Training completion records are exactly what your insurer and your auditor ask for."
| Objection | The reframe |
|---|---|
| "Our staff isn't using AI." | The bet is on naming exactly how many are, before an audit — or a breach — does it for you. A large share of AI-using employees are on personal accounts, not sanctioned ones. |
| "We already have DLP." | Traditional DLP watches for structured data — credit card numbers moving to a file-sharing site. It does not watch for a paragraph of client content pasted into a browser tab. That's exactly the gap the Purview browser extension exists to close. |
| "Our firewall blocks the risky stuff." | ChatGPT, Claude, Gemini, and Perplexity are all trusted TLS endpoints categorized as productivity tools by every standard URL filter. A firewall has no reason to block any of them. |
| "Our team is small — it's fine." | A 40-seat firm shows roughly the same statistical distribution of AI use as a 400-seat firm — it just has less resilience per incident when something goes wrong. |
| "We'll deal with it after we turn on Copilot." | Then Copilot inherits every oversharing permission that already exists, and the customer's first real AI experience is AI surfacing content they didn't know was exposed. Show the value first — but gate it behind the governance work. |
| "We'll just block everything." | Blocking without sanctioning pushes usage further underground, onto personal devices you can't see at all. Sanction one or two tools, monitor the rest, block only the highest-risk consumer apps. |
| "An AUP is enough." | A policy with no enforcement behind it is a wall with no gate. Sell the policy with the control attached, every time. |
| "We don't own Purview or Defender for Cloud Apps." | The Business Premium Purview Suite add-on is a fraction of the Copilot spend, and often discounted further when bundled with it. It powers the assessment today and governs Copilot tomorrow. See Section 6. |
Ten asset concepts worth productizing on top of this guide — each one shortens the sales cycle for a rung on the ladder in Section 10.
A full week-by-week delivery method mapped to Microsoft's four-step model, with tool-configuration walkthroughs, a report template, and an executive-readout skeleton.
The living matrix in Section 8, expanded quarterly with SSO/SCIM detail, DPA availability, and data-training defaults per vendor.
A modular template — sanctioned-tools list, prohibited data classes, monitoring statement, enforcement steps, framework crosswalk.
Cover, exposure heatmap, top apps, top exposed users, business impact, remediation roadmap, licensing gap, timeline, appendix.
The Section 9 method, turned into a weekly 30-minute routine that converts propensity data into a ranked call list.
One per service line in Section 10, with deliverables and exclusions pinned so the assessment doesn't quietly become an unpaid Purview deployment.
Monthly and quarterly cadence: AI usage summary, newly discovered apps, DLP incident summary, sanctioned-app health, license utilization, risk-register updates.
The Section 7 material as a partner-branded one-pager. Converts the biggest objection into a three-minute pre-read that lands the meeting.
Business Premium vs. Business Premium + Suites vs. E5, with the Shadow AI capability delta line by line.
HIPAA / GLBA / SOC 2 / ISO 42001 / EU AI Act supplements to the base playbook, capturing the extra evidence requirements and where the premium sits.
Vertical-specific work in a regulated customer is genuine additional scope, not padding — and it's usually visible enough that the customer recognizes it as real. Community-observed pricing premiums of roughly 20–40% above generalist rates in these verticals reflect that extra work, not markup for its own sake.
| Vertical | Framework | What's different |
|---|---|---|
| Healthcare | HIPAA | PHI-specific sensitive-info-type tuning, a signed BAA for every sanctioned AI vendor, and an ePHI-specific incident-response overlay. |
| Finance | GLBA, SOC 2 | Client financial data classification, customer-identification data classes, and non-public information DLP. |
| Legal | Privilege & ethical walls | Attorney-work-product classification, matter-scoped access control, and ethical-wall enforcement wherever an AI agent could otherwise cross one. |
| Public sector | Government / defense | Data-residency and sovereignty considerations; sanctioned-app selection narrows to vendors with a compliant government-cloud boundary. |
| EU / global | EU AI Act, ISO 42001 | Formal AI system inventory, risk classification, and human-oversight documentation — genuine implementation-grade compliance work. |
"Nobody in our shop is doing this." Even a single unlicensed Copilot Chat user, or one hit in a two-week Defender for Cloud Apps scan, closes this objection in days, not weeks.
"We already have DLP." DLP was built to watch structured data leave through a file share, not to watch a paragraph get pasted into a browser tab.
"Our firewall blocks the risky stuff." Every major consumer AI assistant is a trusted, categorized productivity endpoint. Firewalls let it through by design.
"We're too small to matter." Smaller firms show the same adoption pattern as large ones — with less resilience when something goes wrong.
"We'll deal with it after Copilot is live." By then, Copilot has already inherited every oversharing permission in the tenant.
"Blocking is the answer." Blocking without sanctioning drives use further underground, to devices and networks the partner can't see at all.
"An AUP is enough." A policy with no enforcement mechanism behind it protects no one.
The first paid conversation is Shadow AI discovery, not a Copilot quote.
Any AI product a customer later turns on inherits whatever data hygiene already exists.
Sanction it. Don't block it. Bring it under Entra and Purview management instead.
Discover, Block, Protect, Govern — every finding maps to a step.
The Copilot engine is the same; the Suite add-ons are the SMB-native governance path.
Assessment → AUP → Managed Governance → Insurance readiness → vCISO. Every rung triggers the next.
Before selling a Shadow AI assessment, complete one on your own tenant.
Free-tier usage limits and weak security-maturity scores are the strongest openers you have.
Never sell an AUP without an enforcement mechanism attached to it.
What to build first, in rough priority order:
Every linked item below was checked at the time this guide was written. Product pages, pricing, and licensing tables change on Microsoft's own schedule — re-verify before quoting a customer. Items without a live link are cited by name only; confirm the current URL before publishing them externally.
Statistics with a live link above are attributed to the specific report that produced them. A handful of figures throughout this guide — the ~75% shadow-AI-use estimate, the $400K exposure benchmark, and the SOW pricing ranges in Sections 4 and 10 — are community or editorial estimates rather than primary survey data. They're useful for planning and framing; treat them as directional in front of a customer, and lead with the primary-sourced numbers when precision matters.
Prepared by Ken Lince — Sr. Director, Cloud Engineering, TD SYNNEX · ken.lince@tdsynnex.com
Re-verify pricing, licensing tables, and product names before re-delivering this content — this space moves quickly.