Same Floor, Every Afternoon: What Managed WiFi Should Change for a Singapore Professional Services Firm

In short: The complaint was always the same floor, always in the afternoon, and always unreproducible by the time anyone looked. Nobody owned the wireless network, because nobody had ever been asked to. What fixes that is not new hardware — it is somebody owning the control plane, so the next dropout leaves a record behind it.

Why Does Office WiFi Matter More to a Singapore Professional Services Firm Than Anyone Budgeted For?

Picture a Singapore professional services firm of somewhere between 80 and 150 people: an accounting practice, a law firm, a consultancy, an architecture studio. The work it sells is time and judgement, and almost all of that time now passes through a laptop on a wireless connection. Client calls happen on Teams or Zoom from a meeting room or a desk. Documents live in a cloud document management system rather than on a file server in the back room. Partners work from home two days a week and expect the office to be at least as reliable as their flat.

The office itself rarely grew by design. The firm started on one floor of a building in or near the CBD, took a second floor when a practice group was hired in, a third a few years later, and eventually spilled into part of a floor in the building next door because nothing contiguous was available when the lease came up. Each expansion had a fit-out, a deadline and a budget, and the wireless network was one line in each of them.

That history matters because it is how most wireless estates in firms of this size are actually built: not as one network, but as several decisions made at different times by different people, each reasonable on the day it was made.

The consequence is that WiFi becomes load-bearing without ever becoming anyone's job. It carries the firm's revenue — every billable hour spent on a call or in a cloud application depends on it — and it is managed the way a coffee machine is managed: somebody notices when it breaks.

What Does the Network Actually Look Like on the Fourth Floor?

What follows is an illustrative composite, not a named client, but the shape will be familiar to anyone who has run IT in a firm of this size.

Three batches of access points, bought by three different people

The first floor's access points were specified by the fit-out contractor seven years ago. The second batch was bought by a previous office manager, who picked a well-reviewed business-grade range and installed them with a local cabling vendor. The third batch, covering the newest floor and the space next door, was ordered by the current IT lead during a busy month, partly because it matched what was already on one floor and partly because it was in stock.

Nobody chose a design. Each batch was sized for the floor in front of it, on the day it was installed, for the headcount the firm had at the time.

No central controller, and channels nobody has looked at

Some of those access points are managed individually through their own web interface. Some were adopted into a controller running on a small PC that was set up during one fit-out and has not been logged into for a long time. Nobody can say with confidence which channel each access point is using, whether the 2.4 GHz radios on adjacent floors are stepping on each other, or whether the channel plan that made sense before the firm next door installed its own network still makes sense now.

Channels matter more than most people expect. In a dense office building, access points on the same or overlapping channels — including the neighbours' — share airtime. When there are more devices and more of them are busy, everyone waits longer to transmit. None of this shows up as an error message. It shows up as a video call that freezes for three seconds.

The mesh extender somebody added

At some point a partner complained that the corner meeting room had weak signal, and a consumer mesh extender appeared on a shelf in the corridor. It helped for a while. It also added another radio to an already crowded floor, and many extenders of that kind relay traffic over the same airwaves they serve clients on, so they can quietly add contention as well as coverage. Nobody documented it, and nobody is sure whether it is still doing anything useful.

An IT lead whose actual job is everything else

The firm has one IT lead, or an office manager who has become the IT lead by default. Their week is onboarding and offboarding, laptop replacements, a Microsoft 365 licence question, the printer on level three, a partner's phone that will not sync, and a vendor renewal. Wireless is not in their job description because nobody wrote it in. When a complaint arrives, they do what anyone reasonable would do: walk over, look at the signal bars, restart the access point if it seems to be misbehaving, and move on.

And the complaint keeps arriving. It is always the same floor. It is almost always the afternoon. And by the time the IT lead walks over, everything looks fine.

Why Can't Anyone Fix a Fault That Only Happens in the Afternoon?

The scenario above produces three problems. They are worth separating, because the first one is the reason the other two never get solved.

An intermittent fault with no history is unfixable by definition

Troubleshooting depends on evidence. A permanent fault supplies its own evidence: the thing is broken, and it stays broken while you look at it. An intermittent fault does not. It appears at 3:40 on a Thursday, disappears at 3:55, and leaves nothing behind unless something was recording.

In an estate with no central controller, nothing was recording. There is no record of how many devices were associated with the access point in the corner meeting room at 3:40, no view of how busy the channel was, no log of whether that access point restarted, and no way to know whether the partner's laptop was clinging to a distant access point on 2.4 GHz instead of roaming to a closer one on 5 GHz. The IT lead is left with the one thing that is always available: a person's recollection that "it dropped again."

That is not a competence problem. A skilled network engineer, walking onto that floor at 4:00 with no history, would be in exactly the same position. The fault cannot be fixed because it cannot be observed, and it cannot be observed because nothing was set up to observe it. Restarting the access point feels like progress because the complaint stops — until next Thursday.

Capacity decisions get made by guessing

The same absence of data affects every forward-looking decision. The firm is about to add a practice group of twelve people to the third floor. Does that floor need another access point? Two? None? Are the existing ones overloaded in the afternoon, or is the problem one mis-channelled unit? Should the large meeting room — which fills up for all-hands sessions — have its own access point?

Without client counts per access point, band distribution and a view of utilisation over a normal working week, every one of those questions gets answered by instinct, by a vendor's recommendation, or by buying another batch of hardware and hoping. Hardware bought that way can make things worse: more access points on a crowded set of channels is not the same as more capacity.

The cost shows up as billable hours, not on an invoice

The third problem is the one that keeps the other two alive. Poor wireless in a professional services firm rarely appears as a line item. Nobody receives a bill for it. It appears instead as a fee earner reconnecting to a client call, a meeting that starts six minutes late because the screen share will not load, a document that fails to sync, and a partner who has quietly started taking important calls from home.

Each of those costs is real, but each is small, scattered and attributed to "the WiFi being flaky" rather than to a decision anyone made. So the budget conversation never happens, because nothing in the management accounts points at the problem.

What Does "Managed WiFi" Actually Mean in Singapore — and Is That the Right Question?

When a firm in this position starts looking for help, it will search for something like managed WiFi, WiFi as a Service or a managed WiFi provider in Singapore. Those are the words the market uses. It is worth being precise about what they usually describe.

In this market, "managed WiFi" most often means that someone else owns the hardware. A provider supplies the access points — sometimes bundled with a business broadband circuit, sometimes as a rental or subscription — installs them, and takes a support call when something goes wrong. For a firm that simply wants the problem to belong to someone else, that can be a reasonable arrangement.

Brocent's view is that hardware ownership is the less useful way to split the decision. The more useful split is who owns the control plane: the controller that holds the configuration, watches the access points, and records what happened.

Here is why that matters for the scenario above. The recurring afternoon complaint does not need new hardware to become fixable. It needs the access points to be watched by something that keeps a record, and it needs somebody whose job is to read that record. Once the access points on that floor sit in one controller, the next complaint arrives with evidence attached:

• Channel utilisation and interference on the access point serving that room at that time — whether the channel was congested, and whether neighbouring networks were part of it.
• Client counts and band distribution — how many devices were associated, and how many were stuck on 2.4 GHz.
• Uptime and dropout history — whether the access point went offline, entered a reboot loop, or lost power over Ethernet.

That changes the complaint from "it dropped again" into a diagnosable event with a time, a device and a cause that can be tested. It also changes who is accountable. An alert that becomes a ticket has an owner. A signal-bars check by a busy IT lead does not.

Owning the control plane is also what makes the other decisions tractable. Capacity planning stops being guesswork once there is a normal working week of client counts to look at. And the question of whether a floor needs a survey or simply a channel change can be answered from data before anyone spends money.

None of this means hardware is irrelevant. Poorly placed access points are a physical problem, and no controller moves them. But the order of operations matters: see first, then decide what to buy.

Comparison: Three Ways a Singapore Firm Can Get Managed WiFi

Telco Managed WiFi Bundled With Circuits

• Who owns the hardware: Usually the provider, supplied as part of the service or on a rental, and typically tied to the contract term.
• Who owns the control plane: The provider. What the firm can see of it depends on the provider and the contract — some offer a portal, some offer a support line.
• How it is priced: Often as part of a bundle with the circuit, so the cost of the wireless element on its own is not always visible as a separate unit.
• What to ask before signing: Who owns the access points at the end of the term; whether channel, client and dropout data is visible to you or only to the provider; whether the configuration can be handed over if you change circuit provider or move; and what happens to the wireless if the circuit contract ends.
• Fits: A firm that wants one bill and one number to call, and is comfortable that the wireless network moves with the circuit.

Buy Your Own and Self-Manage

• Who owns the hardware: The firm.
• Who owns the control plane: The firm, in theory. In practice, whoever set up the last controller — which, in the scenario above, is nobody currently employed.
• How it is priced: Capital outlay on hardware, then apparently nothing. The real cost is the IT lead's time and the billable hours lost while problems stay unresolved.
• What goes wrong: No continuous monitoring, no history when a fault is intermittent, firmware left un-upgraded because nobody wants to risk a Friday afternoon, and no record of who changed what.
• Fits: A firm with a dedicated network engineer who treats the controller as part of their job.

Managed Control Plane With Your Own Hardware (Brocent's model)

• Who owns the hardware: The firm. The access points remain its assets, and the configuration can be exported if it leaves.
• Who owns the control plane: Brocent operates a managed UniFi Network controller; the firm's access points are adopted into it, and a NOC watches them 24×7.
• How it is priced: Per access point, per month — a unit the firm can count and check, rather than a bundle it has to trust.
• What changes: Channel utilisation, client counts, band distribution and dropout history become visible per site; alerts become tickets; firmware is upgraded in batches inside a maintenance window; a monthly availability report arrives through Report Central.
• Fits: A firm that already owns UniFi access points, or is willing to standardise on them, and wants accountability for the network without handing over ownership of it. For a firm that would rather not own hardware at all, the same service is available with the hardware included.

What Does This Look Like in Practice Once the Access Points Are in One Controller?

For the firm in the scenario, the change is narrower than a "WiFi project" and more useful.

An honest inventory comes first

The first practical question is which of the three batches can be adopted. Brocent's managed wireless service runs on UniFi Network, so UniFi access points already on the walls can be registered straight into the controller. Anything else on the floor — an access point from a different vendor, the consumer mesh extender in the corridor — cannot, and the inventory makes that explicit. That is often the first time the firm has a single list of what is actually installed, where, and on which firmware.

For access points that cannot be adopted, the firm has a straightforward choice: replace them with units that can, or take the hardware-included model for that part of the estate. The retired extender usually does not need replacing at all once the channel plan around it has been looked at.

Every floor becomes visible in one place

Once adopted, the access points on every floor, in both buildings, appear in one controller under one set of policies. Wireless networks and VLAN separation are defined once and pushed per site or per floor, rather than configured unit by unit. A guest network with a portal keeps visiting clients off the internal network — which matters in a firm that holds client files — and staff can authenticate through RADIUS / 802.1X with their own accounts instead of a shared password that every departed employee still knows.

The monitoring view is the part that changes the afternoon complaint. Channel utilisation and interference are visible per access point. Client counts, band distribution and congestion are visible per site, which is the basis for deciding whether the third floor needs another unit before the new practice group arrives.

Alerts become tickets, not inbox noise

Offline access points, reboot loops and PoE faults are watched around the clock by Brocent's NOC and raised as tickets in the service desk, with someone accountable for closing them. The IT lead stops being the first person to find out that something is wrong — and stops being the only person who can do anything about it.

Firmware goes out in a window, with a backup first

Firmware upgrades are executed in batches inside an agreed maintenance window, never across the whole estate at once during working hours, and a configuration backup is taken before anything changes. The controller's configuration is backed up daily and retained for three years, and there is a change trail recording who changed what and when. For a firm whose clients occasionally ask how its systems are controlled, that record is useful in its own right.

A monthly availability report that says what to do

Each month an availability report arrives through Report Central, stating what happened, what it means and what is recommended. The kind of finding it contains is concrete — for example, an access point logging repeated short, unexpected restarts, traced to power-over-Ethernet delivery running at the edge of budget on an older, out-of-support switch, with a recommendation to move it to a supported port and replace the unit under warranty if the restarts continue. That is precisely the class of intermittent fault that is invisible without a record.

The two buying models, with the unit named

There are two ways to buy the same service, and the only variable is who owns the access points:

• Model A — bring your own access points. The firm's existing UniFi access points are adopted into the controller Brocent runs. Adoption, SSID and VLAN, guest portal and enterprise authentication, 24×7 monitoring with alert-to-ticket, firmware upgrades in a window, daily configuration backup with three-year retention, and the monthly report. The hardware stays the firm's, and the configuration can be exported on exit. Priced per access point, per month.
• Model B — hardware included. Brocent supplies the access points as part of a subscription: no capital investment, the network ready within an hour of subscribing, and hardware maintenance and replacement included, with a spare delivered and swapped within the agreed SLA at no extra cost for the repair. Controller, NOC, change process and report are identical to Model A. Custom-quoted, because it scopes by sites and access-point count.

For the firm in the scenario, the realistic answer is often a mix: adopt the batches that are already UniFi under Model A, and cover the rest either by replacing units or through Model B. The managed wireless network service page documents both models, the commitments behind them — 99% controller availability, no fixed access-point cap, and logical tenant isolation on a shared platform runtime — and what the service does not cover.

What Does the Money Actually Look Like?

The point of a per-access-point unit is that a firm can check it. Here is what it is, and what it is not.

The per-access-point unit

When Brocent's price source was read for this article on 17 September 2026, controller hosting was offered at US$1.20 per access point per month, on a 12-month contract, excluding tax — a year-end offer quoted with promo code YE26-UNIFIAP and running until 30 November 2026. Outside an offer window, the standing price for controller hosting is a custom quote. Offers change, so treat the dated figure here as context and check the current rate on the service page or the pricing page rather than trusting a blog post.

The arithmetic is deliberately simple. A firm with, say, 22 UniFi access points across four floors and two buildings would be looking at 22 units at that month's per-access-point rate — at the figure read on 17 September 2026, US$26.40 a month before tax. The number is not the argument; the fact that the firm can count it is.

This is not a claim to be cheaper than any alternative. A bundled telco arrangement may well cost less in some configurations, and self-managing costs nothing on paper. What differs is that the unit is published, the access points stay the firm's property, and the control plane — with its history — is watched by a NOC and reported on monthly.

The Network & Wireless add-on, for firms on a managed IT plan

Separately, firms on a Brocent Managed IT Support plan can add Network & Wireless maintenance, which looks after switches, gateways and access points together. For Singapore, that add-on was published at S$113.60 per month for up to 10 devices when the pricing source was read on 17 September 2026. It is a banded price rather than a per-device multiplier: it re-bands at 11–30, 31–75 and 76+ devices, and a firm with several floors of switches and access points will usually sit in a higher band. The current Singapore rate for each band sits on the network, server and wireless maintenance pricing page.

The two are different things. Controller hosting is priced per access point and stands on its own — no plan is required to buy it. The add-on is a maintenance line for plan clients that covers the wider network estate. A consultant will say which is the better fit for a particular building.

Where a survey belongs

A controller manages what is installed; it does not move an access point that was mounted in the wrong place. If the controller data shows that the afternoon problem on one floor is coverage rather than congestion or a faulty unit, that is when a wireless site survey earns its cost. Brocent's surveys are carried out by engineers holding Ekahau Certified Survey Engineer (ECSE) and Ekahau Certified Designer (ECD) qualifications using Ekahau tooling, and the survey page publishes man-day scoping — for example, a passive/active survey of existing access points across 5,000㎡ in one man-day.

The useful sequence for a firm like this one is usually: adopt what can be adopted, watch a normal working week, then decide whether the evidence calls for a survey, a channel change, a replacement unit or nothing at all.

Frequently Asked Questions

What does "managed WiFi" actually include?

It depends on who is using the phrase, which is why it is worth asking. Some providers use it to mean supplying and supporting the hardware. In Brocent's model it means the control plane is operated for you: access-point adoption, SSID and VLAN configuration, a guest portal and RADIUS / 802.1X authentication, 24×7 monitoring of uptime, reboot loops and PoE faults with alerts raised as tickets, visibility of channel utilisation, client counts and band distribution, firmware upgrades in batches inside a maintenance window, a daily configuration backup retained three years, a change trail, and a monthly availability report. Whatever a provider's definition, ask whether you will be able to see the history behind a complaint.

Is managed WiFi priced per site or per access point?

Brocent's controller hosting is priced per access point, per month — not per site and not per user — so the cost scales with what is actually installed, whether that is four floors in one building or split across two. The figure read on 17 September 2026 was US$1.20 per access point per month under a year-end offer on a 12-month contract, excluding tax; the standing price is a custom quote, and the current rate is on the service page. The Network & Wireless add-on for plan clients is different: it is a banded monthly price by device count, published per market.

Do we have to buy new access points, or can you work with what we already have?

You do not have to buy anything if your existing access points are UniFi — they can be adopted directly, and they remain your assets. If some of them are from another vendor, those units cannot be adopted into a UniFi Network controller, and the honest options are to replace them or to take the hardware-included model for that part of the office. An inventory at the start makes the split clear before anyone commits to spending.

Do we need a site survey first?

Not necessarily. If the access points are already installed, it is usually more useful to adopt them into the controller first and look at a normal working week of channel utilisation, client counts and dropout history. That data will show whether the problem is congestion, a faulty unit, a channel plan or genuine coverage. A survey is the right tool when the evidence points to placement or coverage, or before a fit-out when nothing has been mounted yet.

What happens when one floor has a problem?

The access points on that floor are already being watched, so an offline unit, a reboot loop or a PoE fault raises a ticket in the service desk with someone accountable for it — often before staff notice. When the complaint is intermittent, the controller's record of that access point around the time of the complaint is the starting point, rather than a walk to the floor after the problem has gone. Any fix that changes configuration goes through a change window with a backup taken first. If the controller itself is unreachable, UniFi access points keep forwarding traffic on the configuration already pushed to them; what is affected is management and statistics, not staff connectivity.

Is this different from WiFi as a Service?

WiFi as a Service usually describes a subscription in which the provider supplies the hardware, which is close to Brocent's Model B. The difference worth checking with any provider is not whether hardware is included but whether the control plane is visible and accountable to you: whether you receive a report, whether alerts become tickets with an owner, and whether the configuration can leave with you. If you would rather keep owning the access points, Model A gives you the same managed controller without the hardware subscription.

What if we move office?

Under Model A the access points are your assets, so they move with you, and the configuration is backed up daily. A mature site's policy can be cloned to the new office, so the new floor does not start from a blank configuration, and pre-provisioned access points inherit it when they are powered on. A new space with different walls and a different layout is exactly where a survey is worth doing before anything is mounted. Under Model B, a move is a scoping conversation with a consultant, since that model is quoted by sites and access-point count.

Where This Leaves the Office Manager

The afternoon complaint on the fourth floor was never really about the access points. It was about the fact that nobody owned the network, so nothing was recording, so nothing could be fixed. Buying a fourth batch of hardware would not have changed that. Somebody owning the control plane does.

That is also why wireless is a useful place for a firm to find out what "managed" actually means in practice. It is small, measurable and easy to check: someone watching the access points at 3 a.m., a change that goes through a window instead of a chat message, a monthly report that says what happened and what to do. Brocent's own hosted controller and its monthly report are one worked example of that — if you want to see the same approach applied to a multi-site business in another city, the Hong Kong serviced-office example walks through it.

But it is the same engine a managed IT plan runs on, applied to one corner of the office. For an 80-to-150-person firm, the wireless network is one of many things the IT lead is quietly carrying alone — alongside devices, identities, Microsoft 365, backups, security and the help desk. The better decision is usually to decide the plan first and let wireless sit inside it, rather than to buy a wireless service and still have everything else on one person's desk. Wireless is the sample; the plan is the meal.

The plan tiers, what each includes and how the managed IT support plans are priced per user are set out in full. When the Singapore rates were read on 17 September 2026, the Established tier was published at S$185.08 per user per month and the Growth tier at S$227.20 per user per month, with Enterprise quoted individually. Every other published rate — including the per-access-point figure and the Network & Wireless add-on — is on the pricing page. Send the number of floors, the number of access points and what you already own, and Brocent's Singapore team will price both the wireless and the plan against the office you actually have.