- One of three scenarios. Pick the one that stretches you most, not the one you already know.
- Option 1: Remote access / VPN replacement for a regulated customer
- Option 2: Web filtering and SWG rollout for a hybrid workforce
- Option 3: Bring your own real customer engagement (minimum 3 CF1 products)
- Which Cloudflare One products to propose. This is the whole point of the exercise.
- We will not tell you which products to use
- Pick from the full CF1 portfolio based on the scenario requirements
- Be ready to justify why each product is in (or out of) your proposal
- A High Level Design (HLD). One architecture diagram showing users, CF edge, identity, private apps, and policy enforcement points.
- A migration plan. Phased rollout with week by week milestones and a named owner per phase.
- A presentation deck. 5 to 8 slides max. You present to the room in 15 minutes (10 pitch, 5 Q&A).
The customer runs two legacy VPN concentrators at its primary HQ. The main appliance is 6 years old, out of vendor support, and has failed over twice in the last quarter. Remote employees and contractors all connect through these boxes to reach internal applications including a web based line of business application, a SharePoint on prem farm, a Windows jump host used to access production databases, and a GitLab server for the engineering team.
The firm is closing an acquisition in 90 days that will add ~480 users across two additional offices. The acquired company has its own VPN and no appetite for merging into the customer's existing appliance. The CISO has told the board she will not extend the current VPN to the acquired entity. She wants a Zero Trust model for both populations before integration.
Separately, the compliance team is unhappy. Contractors (roughly 60 people, mostly engineers engaged through a staffing partner) get the same network level access as full time staff once they are on VPN. There is no per application access control and no per user audit trail for sessions. The auditors flagged this in the last review and expect a remediation plan in writing within 60 days.
The IT team is small. 9 engineers including 2 dedicated to end user compute. They cannot absorb a large hardware project. They want something they can stand up in weeks, not a 12 month programme.
- Per application access for all internal apps: web LOB application, SharePoint farm, Windows jump host (RDP), GitLab (HTTPS and SSH).
- Identity based policy integrated with the existing IdP (Azure AD), including MFA enforcement at the access layer, not just at the device level.
- Separate access tiers for full time employees vs third party contractors. Contractors must never reach beyond their assigned applications.
- Device health and posture must influence access decisions (managed vs unmanaged; disk encrypted; OS up to date).
- Full session audit trail. Who accessed what, from where, on what device. Exportable for compliance audit.
- Outbound only connection from data centre to the cloud service. No new inbound firewall rules, no public exposure of internal apps.
- Scales to ~900 users within 6 months without adding appliances or HA complexity.
- VPN decommission path. A written plan showing how the old appliance is retired, in phases, with rollback at each stage.
The customer runs a legacy on prem web proxy at its primary data centre. Every office site backhauls internet through that data centre over MPLS. Hybrid staff (about 55% of the office workforce) VPN in from home to get filtered internet, which makes Zoom calls stutter, Teams hang, and generates a steady stream of helpdesk tickets.
In the last 4 months the SOC has investigated three incidents where finance and operations staff browsed to compromised websites and downloaded malicious content. One incident reached the accounting server before EDR caught it. The CFO signed off on two weekends of incident response billing and has made it clear he will not do it a third time.
The proxy only sees HTTP/S traffic. It does nothing about DNS level threats such as C2 beacons, DNS tunnelling, and fast flux domains. The current DNS setup is split. On prem AD DNS forwards to an ISP resolver with no filtering. The SOC has no DNS log visibility at all.
There is no written Acceptable Use Policy enforcement. HR has a policy document, but IT has nothing that actually blocks adult content, gambling, or personal cloud storage. HR flagged two internal cases last quarter of inappropriate browsing that went undetected for weeks because no one was looking.
The CIO wants a single web security story for every user, on every network, on every device, including the frontline kiosks, which today sit on a dedicated VLAN with no filtering at all. The board wants the project live within one quarter.
- DNS layer filtering covering malware, C2, and cryptomining, applied to every user, on network and off network, without backhaul.
- Web and URL filtering over HTTPS with content categories for AUP: adult, gambling, personal cloud storage, uncategorised high risk.
- TLS inspection for HTTPS traffic, with a do not inspect list for certificate pinned apps (banking, healthcare portals, Zoom, Teams).
- File type and upload controls to prevent exfiltration of office documents to personal cloud storage over web.
- Isolation for risky or uncategorised browsing. Finance and operations users should be able to open questionable links without the endpoint rendering them.
- Frontline kiosk policy. A stricter location based policy (vendor allow list only) for the 350 kiosks.
- Elimination of the proxy backhaul. Hybrid workers must get policy enforcement from wherever they are, not via VPN.
- Unified log and analytics. SOC needs a single place to investigate DNS and HTTP events, with retention that meets internal audit.
- Governance model. Named owners for DNS threat policy vs AUP policy, defined review cadence, documented exception process.
- Phased rollout. One quarter to full deployment, including a log only pilot before enforcement.
- The anonymised customer profile. Industry, region, size, regulatory context. No real names, no real logos, no account identifying detail. A reasonable pseudonym is fine.
- The actual business problem. Stated as the customer would state it, not as a Cloudflare feature list. "We have three VPN appliances out of support" is a problem. "We need Cloudflare Access" is not.
- Three to five concrete pain points. The ones your champion would nod at if you read them out loud.
- A hard constraint or two. Something real that shapes the design. Identity provider, regulator, existing investment that cannot be thrown away, a deadline.
- A success criterion. One sentence on what "done" looks like for this customer.
- Aim for at least three Cloudflare One products in your proposal. It keeps the story interesting and pushes you beyond a single feature.
- Be ready to share why each product made the cut. The fun part is defending your thinking in Q&A.
- Keep the customer fully anonymous. No real names, no real logos, nothing that would identify them to anyone in the room.
- Skip padding. If you're adding a product just to hit the count, swap it for one you genuinely believe in.
- Public information (industry reports, published case studies) is welcome as colour. Anything shared with you in confidence stays out.