This article sets out best practice for keeping a business website secure, from patching vulnerabilities quickly to limiting the damage if a site is hacked or breached. It draws on how we look after client platforms through our website platform management service at Bright Labs.
AI is helping attackers find and exploit website vulnerabilities faster than most businesses can patch them. Whatever your industry, a hacked website can expose client data, damage hard won trust and stall your pipeline, which makes website security a commercial risk as much as a technical one.
AI and website security
Microsoft's 2026 Digital Defense Report describes AI as changing the pace of cybersecurity, helping defenders and attackers alike find weaknesses in software faster than before. Three of its findings stand out for anyone responsible for a business website:
- Nearly 40,000 vulnerabilities were published in the first half of 2026, putting the year on track to roughly double
- The median time from a vulnerability being discovered in the wild to being weaponised has fallen well below 24 hours
- Enterprises can take 30 to 60 days to fix critical external vulnerabilities
That shift became very real in October 2026, when researchers used AI to uncover two critical vulnerabilities in PHP, the programming language behind an estimated 70% of websites, including those built on WordPress, Craft CMS and Laravel. The flaws had reportedly gone undiscovered for more than a decade, and several Australian government websites were taken offline while patches were applied (Information Age).
Attackers can now act on a new vulnerability within a day, while many organisations still take a month or more to patch it. Good website maintenance exists to close that gap.
For an established organisation the website is usually connected to lead generation, ecommerce, CRM, recruitment and customer service, so the cost of a breach or outage lands on revenue, reputation and sometimes regulatory obligations. The practices below apply whether you manage your website in house, through an agency or with a mix of both.
Recent website vulnerabilities
Over the past few months a steady run of serious vulnerabilities has affected the platforms that established businesses rely on, from popular WordPress plugins through to other CMS platforms. A few recent examples show how quickly a routine maintenance gap can turn into a business problem.
In most of the cases below a fix was available weeks or months before the attacks peaked. The sites at risk were running popular, reputable software that simply had not been updated, which makes this a maintenance and ownership problem far more than a technology one.
| When | Platform | What happened |
|---|---|---|
| October 2026 | PHP (PHP2Shell) | Two critical flaws in PHP itself, found with the help of AI, could let attackers run code on web servers without logging in or read sensitive server memory such as passwords and API keys. The Australian Signals Directorate issued a critical alert on 21 October, and Wikipedia, the Internet Archive and several Australian government sites were among those affected. (Information Age, ASD alert) |
| October 2026 | Bricksforge for WordPress | A flaw rated the maximum severity score of 10.0 allowed attackers to upload and run code without logging in. Patchstack observed exploitation attempts from 7 October. (Cyber Security News) |
| September 2026 | All-in-One WP Migration and Backup | A vulnerability in one of the most widely used WordPress backup plugins could lead to complete site takeover. It was patched on 20 August, yet two weeks later only 35% of installations had updated, leaving roughly 3.2 million sites exposed. (SecurityWeek) |
| September 2026 | WooCommerce Wholesale Lead Capture | Attackers used a file upload flaw to plant backdoors on WordPress eCommerce sites. The fix had been available since 20 February, but Wordfence still blocked more than 100,000 attacks, with activity spiking between June and August. (BleepingComputer) |
| June 2026 | Gravity SMTP for WordPress | A plugin used to send website email exposed its stored mail server credentials to anyone who asked, opening the door to phishing sent from the organisation's own domain. It was patched on 17 March, but Wordfence reported blocking over 17 million exploitation attempts by June. (Threat-Modeling.com) |
Security risks of an unmanaged website
Website maintenance tends to sit somewhere between IT and Marketing and nobody notices it until something breaks. When it does, the problem is rarely small:
- An untested update takes forms or checkout offline in the middle of a campaign
- A missed domain renewal takes the website and company email down together
- An outdated plugin gives attackers a way in to customer data your organisation is legally responsible for
- A silent integration failure sends leads to a CRM that never receives them
The financial impact of getting this wrong is significant and rising. IBM's annual study measures the full cost of serious data breaches, while the Australian Signals Directorate tracks the average loss reported for each cybercrime incident:
| Measure | Cost | Source |
|---|---|---|
| Average cost of a data breach in Australia | AUD $4.22 million, up 38% since 2019 | IBM Cost of a Data Breach Report 2026 |
| Average cost in Australia where attackers abused valid accounts | AUD $4.87 million, the most expensive way in | IBM Cost of a Data Breach Report 2026 |
| Global average cost of a data breach | US $4.99 million, a record high | IBM |
| Average loss per cybercrime report, large Australian businesses | $202,700, up 219% in a year | ASD Annual Cyber Threat Report 2024–25 |
| Average loss per cybercrime report, medium Australian businesses | $97,200, up 55% in a year | ASD Annual Cyber Threat Report 2024–25 |
Australian organisations that took more than 200 days to identify and contain a breach paid an average of AUD $5.17 million, compared with AUD $3.26 million for those that responded faster. How quickly a problem is spotted is one of the biggest levers on what it costs.
The organisations that avoid these problems treat their website as a platform with an owner, a budget and a maintenance plan, in the same way they would treat finance systems or a CRM. The sections below cover what that plan needs to include.
Secure hosting and environments
Cheap shared hosting is designed for sites that can tolerate the occasional slowdown or outage. A business platform needs infrastructure chosen for its traffic, performance requirements, data sensitivity and expected growth, with someone accountable for keeping it secure.
Two hosting decisions matter more than many teams realise:
- Where your data lives. Government, health and other regulated organisations often need data kept within Australia, which is achievable by hosting in Australian regions on AWS and GCP.
- Separate environments. Development, testing and live environments should be kept apart so changes can be checked properly before customers ever see them.
If your website handles personal or regulated data, you should be able to ask your provider where it is stored and get a clear answer straight away.
At Bright Labs our managed hosting and infrastructure runs on Google Cloud Platform and Amazon Web Services, using shared or dedicated environments depending on what each platform needs. Our system administrators take care of:
- Server configuration and provisioning
- Operating system updates and security patches
- Performance monitoring and tuning
- Incident response and remediation
Backups and breach recovery
A backup only has value if you can restore from it and many organisations only discover their backups were incomplete on the day they need them. Copying files to a laptop or shared drive is not a recovery plan for a business site. Best practice covers:
- The database, uploaded files and code, not just one of them
- Automated backups stored separately from the live server
- A defined retention period so older versions are available if a problem goes unnoticed
- Restores tested on a schedule, so the team knows how long recovery takes
Agree two numbers with whoever manages your site: how much data you can afford to lose and how quickly the site needs to be back online. Those answers should shape the backup frequency and the infrastructure behind it.
Patch management
Outdated software is one of the most common ways websites are compromised, yet clicking "update all" on a live business site is a risk of its own. An update to the CMS, a plugin or a server package can break forms, integrations or checkout without warning. Patching also has to reach below the CMS, because server software such as PHP needs the same attention and versions that are no longer supported stop receiving security fixes altogether. Best practice is a planned update cycle:
- Monitor security advisories for the CMS, plugins and server software you run
- Apply updates in a testing environment first
- Check the features the business relies on, such as forms, search, checkout and integrations
- Release to the live site at a sensible time, with urgent security patches fast tracked
- Keep a record of what changed so problems can be traced quickly
Under our platform management agreements, Bright Labs handles plugin updates, server upgrades, performance tuning and general upkeep as ongoing work across Craft CMS, WordPress and Laravel platforms.
In September 2026, only 35% of sites running a popular WordPress backup plugin had applied a critical security fix two weeks after it was released, leaving roughly 3.2 million sites exposed (SecurityWeek).
Licensing matters here too. Every plugin and theme should come from a reputable source with an active developer behind it, because pirated or abandoned plugins save a little money and create a large, unmanaged risk.
WordPress and Craft CMS security
WordPress and Craft CMS are the two content management systems we work with most and while the principles in this article apply to both, the risks tend to come from different places.
| WordPress | Craft CMS | |
|---|---|---|
| Source of risk | Third party plugins and themes, often from many different developers | The CMS core, a smaller plugin ecosystem and control panel access |
| How updates are delivered | Each plugin and theme is updated separately by its own developer | Core and plugin updates are managed through Composer as part of a controlled release |
| Where advisories are published | Plugin developers and security firms such as Wordfence and Patchstack | Craft's own security advisories on GitHub |
| Biggest day to day risk | Outdated or abandoned plugins | Delayed security releases and loosely managed user accounts |
WordPress security
The recent vulnerabilities listed earlier in this article affected WordPress plugins rather than WordPress itself, which reflects where most of the platform's risk sits. Good WordPress security starts with keeping the plugin footprint as small and well maintained as possible:
- Remove plugins that are inactive, duplicated or no longer supported by their developer
- Use licensed plugins from reputable developers with a track record of fixing issues quickly
- Apply security releases quickly rather than waiting for the next scheduled update cycle
- Put a web application firewall in front of the site to block known attacks while patches are tested
Every plugin you install is another developer's code running on your website. Fewer, better supported plugins mean fewer ways in.
Craft CMS security
Craft's smaller ecosystem and controlled update process give it a strong security footing, but it is not immune. In March 2026 a Craft CMS advisory showed that any logged in control panel user, including authors and editors, could run code on an unpatched Craft 4 or 5 site. As a Craft CMS enterprise certified partner, we recommend that every Craft CMS site follows these principles:
- Stay on a supported release branch and apply security releases as soon as they are tested
- Treat every control panel account as privileged, not just administrators
- Review whether public GraphQL access is needed and restrict it if not
- Subscribe to Craft's security advisories so new issues are seen on the day they are published
Security and access
Putting the site behind a content delivery network and web application firewall reduces exposure to bots and automated attacks, but some of the larger risks sit in who can log in and what they can do. Good access hygiene looks like this:
- Every CMS user has their own account, with no shared logins
- Multi factor authentication is switched on for every account
- Each person has only the permissions their role requires
- Access is removed the same week staff, contractors or agencies leave
- Hosting, domain and third party accounts are reviewed as carefully as the CMS
Hosting, domain and DNS accounts are often set up by one person years ago and forgotten. They deserve the same scrutiny as your CMS logins.
On the platforms we look after, Bright Labs keeps systems up to date, monitors for suspicious activity, manages access and acts quickly when new risks emerge. If an incident occurs, the priority is to contain it, which can mean restricting access, applying patches or taking systems offline to protect data, while keeping the client informed at every step.
Website security monitoring
If the first you hear about an outage is a customer complaint or a drop in leads, your site is not being monitored properly. Good monitoring runs around the clock and tracks:
- Availability, so downtime is caught within minutes
- Response times and page speed, so slowdowns are spotted before users notice
- Errors and failed requests across the site and its integrations
- Unusual behaviour that can signal an attack in progress
Our monitoring, uptime and security service watches platform availability, performance and the signals that point to trouble ahead. Issues are acknowledged promptly, handled by severity and followed up with an incident report explaining what happened and what was done about it.
Performance deserves the same attention as uptime. A site that slows down after a content update or traffic spike affects conversions and search visibility, so Core Web Vitals should be tracked over time rather than checked once at launch, ideally alongside your SEO campaign.
Email, forms and integrations
Some of the most damaging website failures are invisible. A contact form that stops delivering enquiries, an order confirmation that lands in spam or a CRM integration that silently drops records can run for weeks before anyone notices.
Email plugins hold credentials too. In 2026 a WordPress plugin used to send website email exposed its stored mail server login details to anyone who asked, opening the door to phishing sent from the organisation's own domain (Threat-Modeling.com).
To keep these channels secure and working, best practice includes:
- Sending transactional email through a dedicated email delivery service rather than the server's default mail function
- Setting up SPF, DKIM and DMARC records correctly on the domain
- Testing forms and integrations after every significant update
- Alerting the team when a form or integration fails, rather than letting it fail silently
- Giving every CRM or ERP integration a clear owner who knows what normal looks like
Domain and DNS governance
Your domain name controls your website and usually your email, yet it is often registered to a former employee, an old agency or a personal credit card. When that account cannot be accessed or a renewal notice is missed, the business can lose its site and inbox in one go.
Domain and DNS ownership is most easily lost during agency changes, mergers and acquisitions. Make it part of every handover checklist.
To keep control of both, follow these steps:
- Register the domain to the organisation, not an individual
- Keep login details in a shared password manager
- Turn on auto renewal and keep billing details current
- Document every DNS record so anyone managing the site knows what it does and can restore it if it is changed or deleted
Website management and support
Most of the risks above come back to one question: who is responsible? When hosting sits with one provider, development with another and monitoring with nobody, problems fall between the gaps and each party assumes someone else is handling them.
The simplest fix is one accountable team for hosting, platform care, monitoring and support, working to agreed service levels. That is how we structure client services and delivery at Bright Labs. Clients work with a dedicated producer during projects and between projects our support team manages and prioritises requests without needing a formal project for every change, so the people who know the platform stay with it.
Whatever arrangement you choose, make sure these essentials are written down:
- Who to contact, and how, when something goes wrong
- Expected response and resolution times by severity
- What the support agreement includes and excludes
- Where the site's documentation, credentials and architecture notes live
Website security checklist
If you are responsible for an organisation's website, these questions will show you quickly where the gaps are.
| Area | Question to ask |
|---|---|
| Hosting | Is the site on managed infrastructure sized for its traffic, with data stored where your obligations require? |
| Environments | Are changes tested somewhere other than the live site before release? |
| Backups | When was a restore last tested and how long did it take? |
| Updates | Is there a regular update schedule? Are all plugins licensed and supported? |
| Access | Does every user have their own login with multi factor authentication and is access removed when people leave? |
| Monitoring | Will someone be alerted to an outage, error spike or slowdown before customers notice? |
| Email and forms | Are website emails sent through a proper delivery service and are form submissions monitored? |
| Domain and DNS | Is the domain owned by the organisation, set to auto renew and are the DNS records documented? |
| Ownership | Is one team accountable for the platform, working to written service levels? |
If several of those answers are unclear, an independent website security audit is a good place to start. Get in touch with the Bright Labs team to talk through how your platform is managed today.




