A reused password, an unpatched remote-access service, or a convincing payment request can create a serious problem without a complicated attack. Before buying another security product, check the protections your business already depends on.
These five areas are a useful starting point. Assign an owner to each and test what you’ve set up. Some changes are quick; account recovery, backups, and network changes may take more planning than a weekend.
1. Turn on MFA, then fix your passwords
Start with email, banking and payroll, your domain registrar, and cloud administrator accounts. Protect the recovery methods too. An attacker who can reset the account may be able to get around its normal sign-in controls.
Prefer phishing-resistant methods such as FIDO security keys or passkeys where supported. CISA’s MFA guidance explains the distinction. Authenticator codes and SMS can add protection, but they aren’t phishing-resistant. Test enrollment and recovery with a normal employee account before rolling the change out.
Use a password manager to generate and store long, unique passwords. For services you operate, NIST’s password guidance rejects arbitrary composition rules and routine forced changes, while requiring changes when there is evidence of compromise and checks against common or compromised passwords. Those checks belong in the authentication service; a password manager alone doesn’t enforce them everywhere.
When someone leaves, revoke their accounts, sessions, tokens, and shared access as appropriate. Removing their password-manager account does not recall a credential they already copied.
2. Patch what faces the internet first
List the software and devices you rely on, their owners, and their update paths. Include appliances, remote-access tools, website plugins, and bundled software. You can’t make a useful patch decision about a system nobody knows exists.
Prioritize exposed services and vulnerabilities with evidence of exploitation, considering the access and data at risk. CISA’s Known Exploited Vulnerabilities catalog is one input to that decision. A vulnerability absent from the catalog still needs assessment.
Enable automatic updates where they fit the workload and test recovery for changes that could interrupt a critical service. After updating, check the version actually running. A successful installer doesn’t tell you whether an old process or container is still serving traffic.
3. Make phishing easy to report
Give people one obvious reporting route and make sure someone responds. Thank them for reporting a suspicious message, even when it turns out to be legitimate. If someone clicked or submitted credentials, you need that information quickly.
For payment or bank-detail changes, use a contact number from an established directory or supplier record. Check the specific recipient, amount, and destination account. Don’t use the number supplied in the suspicious request.
A callback helps only if the contact record is trustworthy and the employee is allowed to pause the transaction. Protect changes to that record, define escalation when the requester can’t be reached, and practice with a fictional payment. This reduces reliance on the incoming message; it doesn’t eliminate every form of payment fraud.
4. Back up like ransomware is coming
Keep recovery copies separate enough that compromise of an ordinary account doesn’t also let an attacker delete the backups. Offline copies or appropriately configured immutable storage can help. Check retention, administrative permissions, encryption keys, and the restore path before relying on the label “immutable.”
Choose a restore test around something the business needs: a mailbox, an application database, or a complete service. Recover it into an isolated environment and check the contents and access controls. Record how long it took and the age of the recovered data, then compare both with the business’s requirements.
Include cloud services in that review. Availability, deleted-item retention, and a recovery copy you control answer different questions. Find out what the provider can restore and what you must arrange separately.
5. Stop putting everything on one network
Decide which devices need to talk to which services. Guest Wi-Fi and devices such as printers and cameras are useful places to start separating access. Finance systems and customer data need rules based on who actually uses them.
A separate network name or VLAN isn’t enough on its own. Review the routing and firewall rules between segments. From a test device, check one connection that should work and one that should fail. Keep an administrative recovery path so a mistake doesn’t lock you out of the equipment you need to fix.
A realistic first month
Choose the order around your biggest exposure and the people available to do the work. A small team might begin with account protection and recovery, then inventory and urgent updates, followed by payment verification and a backup restore test.
Track each item with an owner, a result, and the next check. Leave unfinished work visible. “Backup enabled” and “the application restored successfully” are different milestones.
When it’s worth calling someone
Bring in help when a suspected incident needs investigation, a change could interrupt an important service, or customer and regulatory requirements need specialist review. You can still own the plan and ask for evidence that the work achieved its purpose.
If you’d like a second pair of eyes, contact us with the systems you’re concerned about and what you’ve already checked.
