Support desk guide

Five Common ISP Support Calls: A Resolution Playbook

A vendor-neutral desk guide for internet outages, slow speeds, Wi-Fi problems, paid-but-offline customers, and repeated drops. Ask three questions, check the evidence, and choose a fix or a field visit.

Review CPE compatibility
Illustrative support worksheet

Find the fault before changing access

Call types
5
Questions each
3
Evidence checks
4
BalanceInvoice and service entitlementRecorded
SessionAccess state and timestampRecorded
SignalBaseline and current readingRecorded
CPEPower, uptime and WAN linkRecorded
Example worksheet, not a live product screen. Use your existing authorized billing and network tools.
Page type
Guides
Primary search
ISP support call playbook
Updated
2026-10-16

Start with the customer path

These five scenarios are a practical training set, not a measured ranking of call volumes. Use your own ticket history to decide which deserves the most attention. The goal is to resolve routine calls with evidence from your authorized tools, without opening a vendor router console for every caller.

Verify the caller and the affected service before discussing balances or changing settings. Record the start time and timezone, check known incidents, and confirm whether one device, one household, or several subscribers are affected. Never ask for a payment PIN, one-time code, or router administrator password.

CallFirst distinctionJump to the playbook
My internet is offBilling restriction or network fault?Outage checks
It is slowOne device, home Wi-Fi, or the access network?Speed checks
Wi-Fi password / can't connectCannot join Wi-Fi or joined without internet?Wi-Fi checks
I paid but I'm still cutVerified allocation or access restoration failure?Payment checks
It keeps droppingLocal power, Wi-Fi, access link, or shared incident?Drop checks

Follow the same path each time: balance and entitlement → current session → access signal → CPE status → customer traffic test. CPE means customer premises equipment, such as the router, outdoor radio, or optical network terminal (ONT). Mark unavailable evidence as unknown, not healthy. A last-seen timestamp is not a live reachability test.

1. “My internet is off”: unpaid or a fault?

Three questions to ask

  1. When did service stop, and did anything change just before it stopped?
  2. Are all devices affected, including a wired device if one is available?
  3. What do the power, WAN or optical lights show, and is there a payment or service-expiry message?

What to check

  • Balance: inspect the affected service's invoice, allocation, expiry, grace period, suspension reason, and manual hold. A balance alone does not prove a suspension occurred.
  • Session: check the current access session or lease and its timestamp. On PPPoE, distinguish discovery failure, authentication rejection, and an established session without usable traffic. Consult the PPPoE protocol stages when interpreting logs.
  • Signal: compare the wireless link or optical receive level with the installation baseline and that model's operating limits. Check whether neighbors on the same access equipment are affected.
  • CPE status: check power, uptime, WAN link, last contact, and recent configuration changes. Missing management contact can also mean the management path is unavailable.

Usual fix and confirmation

For a verified billing restriction, explain the exact service and period involved and use the existing approved payment or dispute process. Restore access only through the authorized policy path after entitlement is confirmed. Keep any manual hold until its authorized owner releases it.

For a fault, resolve a known shared incident first. If evidence points to the premises, guide the customer through a safe power and cable check. Capture evidence before any agreed restart; do not factory-reset equipment. Confirm the session and ask the customer to load a normal site or application after the change.

When to send a technician

Dispatch for a persistent premises power or WAN-link fault, lost optical signal, or a radio fault requiring physical inspection after safe checks. Escalate multi-subscriber outages to the network operations center (NOC), and billing disputes to finance. A financial restriction alone does not justify a truck roll.

2. “It is slow”: isolate the bottleneck

Three questions to ask

  1. Is every application slow, or only one site, stream, or download?
  2. Does it happen on every device, and how does a wired or near-router test compare?
  3. Is it constant or limited to certain times, and are uploads, backups, or other users active then?

What to check

  • Balance: confirm the purchased package and any documented restriction or expiry; do not assume a slow connection is unpaid.
  • Session: check the applied speed profile, current traffic, and session freshness against the entitled plan. A profile mismatch needs access-team review.
  • Signal: compare wireless signal, noise, retries, and link quality with the baseline; for fiber, compare optical readings with model limits. Check access and backhaul utilization at the reported time.
  • CPE status: inspect WAN link speed, errors, uptime, and available resource indicators. Compare home Wi-Fi performance with the access link rather than treating them as the same radio path.

Usual fix and confirmation

Run a short, repeatable test with the customer's consent: one capable device, a wired connection where practical, other heavy transfers paused, and the same test destination. Record time, method, download, upload, and latency. A single Wi-Fi speed test is not proof of access-network capacity.

Correct an evidenced profile mismatch through the authorized workflow. For home coverage, test closer to the router and agree a placement or coverage change. Escalate shared busy-hour congestion to the NOC for capacity work. Repeat the same test and the customer's original application after the fix; do not sell an upgrade as a substitute for diagnosing a fault.

When to send a technician

Dispatch for a reproducible premises cable, Ethernet negotiation, coverage, or alignment problem that needs a visit. Include baseline and current readings. Shared backhaul congestion needs network work, not separate customer appointments.

3. “Wi-Fi password / I can't connect”: joining is not internet access

Three questions to ask

  1. Can the device see the correct Wi-Fi network name, and what exact error appears?
  2. Can another device join that network and use the internet?
  3. Was the Wi-Fi name or password changed recently, or was the router replaced or reset?

What to check

  • Balance: verify entitlement, but remember a billing restriction does not by itself explain a wrong-password error.
  • Session: if other devices work, keep the investigation local to the failing device. If devices join but none can browse, inspect the WAN session, client address assignment, and DNS evidence.
  • Signal: compare the failing device near the router with its usual location. Confirm whether home Wi-Fi or the outdoor access signal is being measured.
  • CPE status: verify the Wi-Fi radio is enabled, the intended network name is present, and the device model supports the configured security mode. Keep security enabled.

Usual fix and confirmation

After verifying the caller's authority, guide them through the operator's secure Wi-Fi credential retrieval or change process. Do not copy passwords into tickets. If a saved profile is wrong, have the customer forget that network and rejoin with the current credential. Warn that changing a shared Wi-Fi password disconnects other devices, and confirm a recovery path before applying it.

Separate an association error from “connected, no internet”; the latter belongs in the outage playbook. Apple's Wi-Fi troubleshooting guidance illustrates checking the selected network, range, and other devices. Use the equivalent instructions for the customer's device. Confirm both joining and browsing after the fix.

When to send a technician

Dispatch when the router's radio appears faulty, an authorized configuration cannot be recovered remotely, or a coverage survey is needed. A problem isolated to one phone or laptop may require that device's support team instead.

Three questions to ask

  1. What payment reference, amount, time, and channel were used, without sharing a PIN or one-time code?
  2. Which customer account, service, and billing period was the payment intended to cover?
  3. Is the payment shown as completed or pending, and has service worked at any point since payment?

What to check

  • Balance: find trusted payment evidence, then check allocation to the correct invoice and service. Review partial payments, duplicates, reversals, expiry, and manual holds. A screenshot alone is not verified settlement or allocation.
  • Session: compare the entitled access state with the actual session or restriction. Check the result and timestamp of any restoration attempt, including failures or pending approval.
  • Signal: confirm there is no simultaneous access fault. A valid payment cannot repair a broken fiber or a failed radio link.
  • CPE status: check power, WAN link, last contact, and uptime before requesting a session refresh or restart.

Usual fix and confirmation

Finance should resolve an unmatched reference or incorrect allocation through its approved reconciliation process. Do not ask the customer to pay again solely because service remains offline. Once entitlement is confirmed, an authorized operator can apply the supported restoration workflow, inspect its result, and refresh only the affected session when required.

Close the loop with a fresh session observation and a customer traffic test. Record separate timestamps for payment verification, allocation, access change, and successful service. The paid-but-offline explainer expands this evidence chain; automatic suspension scope describes guided-pilot boundaries.

When to send a technician

Dispatch only if a premises fault remains after entitlement and access policy are correct. Keep unresolved allocation with finance and access-policy failures with the access team; do not hide either in a generic field visit.

5. “It keeps dropping”: correlate timestamps before rebooting

Three questions to ask

  1. What were the last two or three drop times, how long did each last, and did service recover by itself?
  2. Do all devices disconnect, does Wi-Fi disappear, or does Wi-Fi remain connected without internet?
  3. Do drops coincide with power changes, weather, busy hours, or a particular device or activity?

What to check

  • Balance: look for expiry, renewal, or restriction changes at the reported times rather than assuming every disconnect is a physical fault.
  • Session: compare reconnects, termination reasons, lease renewals, and authentication results with the customer's times. Distinguish stale records from repeated live reconnects.
  • Signal: review signal and error trends at those times, including other subscribers on the same access path. One good reading between drops does not exclude an intermittent fault.
  • CPE status: correlate uptime resets, power events, WAN-link changes, and recent firmware or configuration changes. Preserve available logs before a reboot clears useful evidence.

Usual fix and confirmation

Change one evidenced cause at a time: a loose accessible Ethernet cable, an approved power supply replacement, a reviewed configuration correction, or a network-side incident fix. Do not ask a customer to climb to an outdoor radio, open electrical equipment, or handle exposed fiber.

Agree an observation window that includes the usual failure period and a callback time. A few successful pings are not enough to close a recurring evening fault. Record what changed and whether the same symptom recurred.

When to send a technician

Dispatch for recurring power loss, damaged cable, weather-related ingress, optical loss, or alignment issues requiring physical work. Send timestamped evidence and the equipment identity. If drops align across a sector or site, the NOC should own the shared incident.

Worked example: a payment call with a separate fault

A fictional subscriber reports paying at 10:05 and still being offline at 10:20. Finance confirms the reference and allocation; the service is entitled to access. The router has power, but the ONT reports lost optical signal and no current session is present. Repeating payment or restoration attempts cannot resolve the optical fault. Check for a shared outage, then dispatch a technician if the evidence isolates the premises. After repair, verify the session and customer traffic before closing the call.

Close or hand off with evidence

Record the service identifier, verified caller authority, incident times and timezone, all four checks with source timestamps, the action and authorizing role, the customer traffic test, and the next owner. Leave passwords, payment PINs, and full payment credentials out of notes. For a visit, add site access arrangements, suspected physical fault, equipment model, measurements, and the parts or skills needed.

Close only after the original symptom is resolved and the customer confirms it. For intermittent faults, keep the agreed observation window open. If a check was unavailable, state the uncertainty and the next person responsible for resolving it.

Where ISPAgents fits

ISPAgents supports subscriber search and device assignment, plus TR-069/CWMP informs, parameter reads/writes, and diagnostics; device coverage depends on model and firmware.

The ISPAgents CRM service and Customer 360 foundation are guided pilot scope. The complete one-screen playbook described here is planned: do not assume that Customer 360 can reset Wi-Fi, refresh a live session, prove restoration, or dispatch a technician from one screen today.

ISPAgents service-policy preview, approvals, and result recording are guided pilot foundations; live router/RADIUS mutation remains gated. Customer billing and payment-adapter foundations are also guided pilot scope, not a promise of automatic payment-to-access restoration.

Evaluate the CRM guided-pilot scope, TR-069 CPE compatibility, and subscriber/device identity foundation against the tools your team uses today. Bring an anonymized support case to a guided-pilot discussion and agree which evidence and actions can actually be demonstrated.

Frequently asked questions

Should an unpaid balance end an outage investigation?

No. Check which service and period the balance covers, the suspension reason, grace rules, and any manual hold. A billing issue and a physical fault can coexist.

Does an online session prove the customer's internet works?

No. Session evidence needs a timestamp and can be stale. Confirm usable traffic from the customer's device and distinguish Wi-Fi association, IP addressing, DNS, and upstream reachability.

When should support send a technician?

Dispatch when the evidence points to a premises fault that cannot be safely resolved remotely, such as repeated power loss, damaged cabling, lost optical signal, or an access radio requiring alignment. Escalate shared access or upstream incidents to the NOC.

Can ISPAgents resolve all five calls from Customer 360 today?

That complete one-screen workflow is planned. ISPAgents Customer 360 and service-policy foundations are guided pilot scope; live router/RADIUS changes remain gated. Qualify each action and device path before relying on it.

Next step

Discuss the scope for your network.