Write-ups from deployments we have actually run, organised by section. Specific problems, the reasoning behind the approach, and what it cost to get wrong. Client names are withheld under confidentiality.
Page 6 of 8
Most published checklists are lead generation with a form at the end. A real one already exists, is free, is maintained, and is ordered by what actually stops attacks — this maps it onto what a growing business already runs, and is honest about which items are hard.
Availability is sold in nines, and each additional nine costs roughly an order of magnitude more than the last. Most businesses asking for high availability need fast, tested recovery instead — and the two are different engineering problems with different price tags.
Most dashboards are built to show everything, which is why nobody looks at them. A dashboard is a diagnostic tool with one question to answer, and the discipline that makes it useful is deciding what to leave off.
The build-or-buy question is usually asked about the wrong thing. It is not whether custom software is better, but whether this particular workflow is where your business differs from everyone else's — because that is the only case that justifies owning it.
Both are logged-in areas where someone outside the business sees data from inside it. They differ in tenancy, in what a mistake costs, and in the single control that matters most: whether one account can ever see another's records.
A contact centre is one of the few businesses where infrastructure failure stops revenue in the same minute. That single property changes how every technology decision should be made.
Choosing a topology is the architecture decision. Getting the deployment right is a different set of problems: protocol choice, addressing, certificate lifecycle, split routing and the rollback path you will want on the day it goes wrong.
Screen pop, click to call and automatic call logging are three separate integrations that get bought as one feature. Each has a different failure mode, and the one that decides whether agents trust it is number matching.
Zero trust is an architecture, not a product, and the vendors selling it as one have made it harder to adopt. What it actually requires is a shift from trusting a network location to verifying every request.
Enabling multi-factor authentication is a Tuesday afternoon. Choosing which factors, covering the accounts that matter, and handling the recovery path without recreating the weakness are the parts that decide whether it holds.
Most infrastructure is not designed. It accumulates, one urgent decision at a time, until the estate is a record of past deadlines rather than a deliberate architecture. Planning is mostly the discipline of deciding before the deadline arrives.
Voice systems are attacked continuously and automatically, and the objective is almost always money. Toll fraud is billed by a carrier who delivered exactly the calls that were requested, which makes it a loss you cannot dispute your way out of.
We typically respond within 4–8 business hours.
