Conor O’Neill, Co-Founder and CEO of penetration testing provider OnSecurity, explains why SMEs are at risk and how to tighten cybersecurity
AI has increased the speed and complexity of cyberattacks1, by allowing hackers to automate their activity and operate at scale. Cybercriminals are now using AI to write exploit code quickly, identify vulnerabilities in systems, adjust tactics by analysing defences and most commonly, generate phishing and deepfake content.
Phishing attacks were the most prevalent and disruptive type of cyberattack last year2, and with 155 firms, including Google, Microsoft, Anthropic and OpenAI, signing an open letter3 warning that external cyberattacks using AI will become both more widespread and sophisticated over the coming months, it is now even more important for businesses to prioritise cybersecurity.
‘The idea that small businesses aren’t at risk is a dangerous misconception’
Although cybercriminals often target organisations that handle high-value transactions and sensitive data, where the potential return is highest, smaller businesses are still at great risk. They typically have fewer resources, smaller budgets and fewer regulatory requirements. As such, so less is invested in cybersecurity, making small businesses easier to target, with attackers able to mass scan, find weaknesses and gain access.
The signs that businesses should look out for
Hacks don’t always happen on the first attempt. Cybercriminals need to gather information about the company and its systems before attacking.
AI has sped up the rate at which attackers can scan systems for vulnerabilities, so any sudden increases in network traffic should be investigated. Establishing a baseline of normal traffic activity across systems allows for easier detection of any unusual patterns.
Receiving unusual emails internally and externally, unusual login attempts, and slow devices can also be signs of a cyberattack4 .
Social engineering attacks are extremely common, with cybercriminals using AI-generated or deepfake communications to impersonate employees to gather information. A report from Verizon, analysing 31,000 security incidents, found that 16% of breaches were social engineering, making it the third most common breach5.
AI can make extremely convincing messages, videos and audio, making detection much more difficult. The signs of a social engineering attack include unexpected requests for sensitive information, urgent language, suspicious links and spelling or grammar mistakes in communications.
6 hacking prevention methods
Encrypting customer and payment data
If encrypted data is stolen in a cyberattack, it will be unreadable without the decryption key, preventing attackers from accessing or exploiting the information.
Using multi-factor authentication
This adds an extra layer of security by requiring users to provide more than one form of verification when logging in, leaving stolen passwords useless.
Running regular security audits and penetration testing
Organisations need to ensure their most critical assets are secure through regular penetration testing, to assess for vulnerabilities before they are exploited.
Training staff to spot threats
Staff must be properly trained to recognise the signs of cyberattacks to easily identify attempts and report them immediately. When employees consistently recognise suspicious behaviour and report it quickly, organisations gain early warning of attacks, reducing the number of successful breaches and creating a culture where security is everyone’s responsibility.
Keeping systems and software up to date
Outdated software can contain weaknesses that attackers may exploit to gain access to systems.
Creating an incident response plan
Having a response plan in place means businesses can act quickly if a cyberattack occurs, and should include how to identify, contain and respond to an incident, alongside who is responsible for each stage. This helps to limit the disruption caused and protect systems.
As AI makes cyberattacks faster and increasingly more difficult to detect, businesses need to proactively prevent and prepare for potential hacks. A combination of technology, processes and training can help companies to mitigate the risk of attacks and enhance their resilience.
Article Sources:
-
https://www.ncsc.gov.uk/news/the-ai-shift-in-cyber-risk-why-leaders-must-act-now
-
https://www.verizon.com/business/resources/executivebriefs/2026-dbir-executive-summary.pdf
About Conor O’Neill:
As Co-Founder and CEO of OnSecurity, Conor brings 15+ years of experience in penetration testing and security consulting. Conor combines deep technical expertise with strategic leadership. He drives product innovation, develops next-generation security services, and builds teams that deliver impactful, cutting-edge cybersecurity solutions for modern enterprises. His expertise has been featured in publications such as TechRadar, Insight Magazine and Silicon Republic.
About OnSecurity:
OnSecurity, a CREST-approved AI-augmented penetration testing vendor based in the UK, is dedicated to delivering high-impact, high-intelligence penetration testing services to businesses of all sizes. Our mission is to simplify the management and delivery of pentesting for our customers, using our platform to help them improve their security posture through expert testing, actionable insights, and unparalleled customer service.
