Fit

Who REAP is for

Every network changes. Every change is a risk the ticket cannot see. Every team is smaller than its estate. Whether REAP is the right next step for yours comes down to three questions, and the honest answer to each includes when to use the tools you already have first.

Restored is not the same as fixed.

Completed is not the same as verified.

Configured is not the same as compliant.

The argument

Why every company, and why not every company today

Every network changes: a firewall rule, a route, an acquisition, a cloud account nobody told the network team about. Every change is a risk the ticket cannot see: the change record says what was intended, not what happened to the paths around it. And every team is smaller than its estate: the engineer who knows why the plant’s core was built that way is one person, and that person is on holiday when it matters. Those three facts are true of every company that runs a network, which is why evidence-backed network assurance is for every company.

Whether REAP is the right next step today is a narrower question. It depends on whether the problems the three doors solve are the problems costing you money now, and on whether you can bring what a proof of value needs. The decision tree below is the same one we use in a first conversation.

The decision tree

Three questions, one door

Answer them about your own estate. Each yes names the door to start with; the last step is for the team that answers no three times.

  1. Do you run three or more network vendors, or one vendor plus a hybrid cloud?

    If yes, start with Incident Reasoning. The incidents that take longest are the ones that cross a boundary: between vendors, between the data center and AWS (opens in a new tab) or Azure (opens in a new tab), between the network and the ticket that changed it. A single vendor’s assistant stops at the vendor boundary; REAP’s Investigation Window runs one investigation across all of them and writes the evidence to the ticket.

  2. Did your last P1 follow a change that had closed as successful?

    If yes, start with Change Assurance, even with one vendor. The change record said the work was done; nobody compared the network before and after. Change impact analysis before the window and a verdict written to the change record afterwards are the first things a pilot delivers, on a real change your team runs.

  3. Can you prove today that failover will hold when it is needed?

    If no, start with Resilience Assurance. Standby sync, running-versus-startup drift and contextual CVE (opens in a new tab) exposure are findings in week 3 of a pilot; continuous failover-readiness proof is in development, and backup-path capacity is a pilot finding where interface telemetry is available.

  4. No three times? Use your vendor’s tools first, and here is when to come back.

    If one vendor, few sites and a change process you trust describe your estate, your vendor’s assistant and your monitoring are the right next step, and the free guides on this site (the change-verification checklist, the failover test plan, the drift review, the operations baseline) will take you a long way. Come back on the day one of three things happens: an incident crosses a boundary your tools cannot see across; a change closes as successful and breaks something anyway; an auditor or a customer asks you to prove that failover works, and the proof is a screenshot.

By team

How teams of different shapes start

The door is the same decision tree; what differs is who runs the loop and what a pilot needs from you.

TeamWhat usually costs the mostStart withWhat you bring to a pilot
Enterprises: many sites or business units, three or more vendors, change control and auditIncidents that cross business units and vendors; change windows nobody verified; redundancy the board assumesIncident Reasoning on one domain; Change Assurance on one real change window in the same pilotA champion and an economic buyer at kickoff and readout; the ITSM, one telemetry source and configurations; read-only credentials, brokered through your vault where you run one
Mid-sized companies: 3 to 10 network engineers, many sites, no 24x7 NOCThe senior engineer as the single point of failure; incidents that open at 2 a.m. with no context; changes verified by waitingIncident Reasoning: incidents open with what changed and the evidence behind itOne domain of 50 to 250 devices, ServiceNow (opens in a new tab), Jira (opens in a new tab) or ManageEngine, and a decision date
MSPs and managed-network providers: many customer estates, one consoleEscalations to senior engineers that tier 1 cannot resolve; proof to each customer of what was doneIncident Reasoning across 3 to 5 customers in a 60-day partner pilotOne production domain per customer; per-operator customer scoping is in development
Not yet: one vendor, few sites, a change process you trustUsually nothing the three doors solve this quarterYour vendor’s tools and the free guides; the self-assessment to find your stageNothing yet: bring us the first incident that crosses a boundary

What a pilot needs

Three things to bring, and what you keep

A proof of value is 45 days on one domain, measured against your baseline, ending in BUY, EXTEND or STOP in front of the person who signs.

Week 0

Read-only access, through a connector inside your network

Least-privilege credentials for discovery, telemetry and reasoning, brokered through CyberArk (opens in a new tab) where you run it; an outbound TLS tunnel and no inbound firewall rule. You can revoke any credential at any time.

Weeks 1 to 4

One domain and its last incidents

50 to 250 devices, one or two sites, and five to ten incidents your team already solved, whose root cause you keep to yourself until the replay. The replay is the test.

Week 6

A baseline and a decision date

Your numbers from week 0 (detect, resolve, recurrence, verified changes, cost per hour) and a business review on a date you set. What you keep: the topology knowledge, the runbooks, the change policy and every approval.

Straight talk

When REAP is not the right next step

Some teams should build rather than buy: if a platform team already builds and runs your own agents, build, and use the buyer’s guide as a design review. Some policies decide the order of events: if your policy requires your own cloud account or your data center, say so in the first conversation; both are by arrangement and scoped per customer, and the timeline is agreed before a pilot is scheduled. Some estates are not ready for an investigation to help: without a ticket system or any telemetry there is nothing to reason over, and the operations baseline is the place to start. And some teams want a reference first: REAP runs today with design partners and has no public customer references yet; named references follow only with measured results and written agreement, which is the honest answer and the one we would want from a vendor.

Questions

Fit: straight answers

We have one vendor. Is REAP for us?

Ask the second and third questions. A change-caused P1 or a failover you cannot prove is a reason to start with one vendor; a single-vendor estate with no such history is better served by its vendor’s tools today, and the guides on this site are free.

We already run our vendor’s AI assistant.

Keep it; it knows its own devices well. REAP reasons across every vendor plus the tickets and configuration history between them, and the assistant’s answer becomes evidence in the investigation. The page REAP and your vendor’s AI assistant is written for the colleague who asked.

We run OT networks in plants.

Then the question is usually the third one. Failover on the plant floor is rarely proven before it is needed, and a change made in IT reaches OT through a path nobody drew. Read-only is the starting posture, which is the posture OT requires.

What does it cost?

An annual subscription priced by the size of the managed environment (device and site bands), the capability tier and the deployment model; a paid 45-day proof of value, credited in full against the first-year agreement. The pricing page has the bands.

We are not ready. What should we do first?

Measure the seven numbers of the operations baseline, verify your next change with the checklist, and test one failover with the test plan. Each is a free template; each is what a pilot would do first anyway.

Three statements. Check them against your own network.

If two are true, REAP is built for your estate. The self-assessment takes 10 minutes and returns a stage, your 3 largest gaps and a first door; a 45-day proof of value proves the next stage on your own network.