Opens in a new tab
vmblog logo 2024 wht (updated)

Your incident response plan won’t survive the clock

Share: 

Karl Bagci | Published: November 18, 2026
vmblog predictions 2027

By Karl Bagci, Director of IT and Information Security, Exclaimer

Most organizations can produce an incident response plan on request. Far fewer have made one work against a deadline, and in 2027 plenty of US organizations will find out that theirs is really just a title page.

Two reporting rules matter here. US public companies have four business days to disclose a material cybersecurity incident, counted from the point they decide it is material. CISA’s final rule under the Cyber Incident Reporting for Critical Infrastructure Act (CIRCIA) is with the White House for review. If it lands as expected, many more organizations in critical infrastructure sectors will be working to a 72-hour reporting clock. Neither deadline waits for the facts to be complete: you report what you know and update it later. In the first hours of an incident, a fair amount of what you think you know turns out to be wrong.

So, the plan on paper proves very little. The only useful test is an honest one. Start by naming the person who signs off on notification. That’s usually when it turns out nobody’s quite sure whose call it is. Once someone owns the decision, put them on the clock and have them make the call with the early picture incomplete and partly wrong.

Then run the exercise again without your best responder. Most teams have one: the person who knows where the logs live and which vendor to call at 2 a.m. If the plan only works when they’re reachable, you have a dependency, not a process.

Your plan also depends on people you don’t employ

Asking a supplier whether they have an incident response plan is becoming a weak question. Everyone has a plan. Over the next year, I expect customers to ask what that plan has actually been through. When was it last tested? How quickly would the supplier tell you something had gone wrong? Who on their side makes that call?

You can’t make decisions about an incident nobody has told you about, so your response is only as fast as their notification. A supplier that spends two days deciding whether to call you has spent two days of your own assessment for you.

It matters most for the platforms your communications run through, because those are what you lean on once something has gone wrong. Ask for the date of the last exercise and what it found, and agree on a named contact and a notification window when you sign the contract. Better still, put your most important suppliers in the room the next time you run a tabletop. Ten minutes on “how would you reach us, and how would we know it was you?” will tell you more than any questionnaire.

The warning will probably arrive by email

That last question is getting harder to answer. People have been predicting the death of email for years. AI is making it more important, not less. As more work gets automated and spread across chat tools and platforms, email is still the channel people treat as official. It’s where a supplier’s breach notice is likely to land. That’s also why attackers like it so much.

Business email compromise accounted for more than $3 billion in losses reported to the FBI in 2025, and the FBI has warned that AI chat tools can quickly produce official-sounding emails that mimic a CEO. An incident is an ideal moment to send one. Everyone is expecting urgent messages from people they rarely hear from, and nobody wants to be the one who slowed things down.

Polished writing no longer tells you much. I once received a well-written email from the CEO of a security vendor and replied asking whether it was a phishing test. It wasn’t. I do this for a living, and I couldn’t tell from the message alone.

What can be checked is narrower and more useful. Authentication tells you a message really came from the domain it claims. It can’t tell you whether that account has been compromised or whether the domain is a lookalike, so an unusual request still needs confirming another way. During an incident, that second channel needs to exist already. If you and a key supplier haven’t agreed how you’d verify each other when email can’t be trusted, you will be improvising that part under the same clock as everything else.

The real question for 2027 is what people will expect an email to prove about whoever, or whatever, sent it.

##

ABOUT THE AUTHOR

karl bagci

Karl Bagci is the Director of IT and Information Security at Exclaimer, with extensive experience across cybersecurity, operations, and compliance. He has held leadership roles at companies like Dubber, Cronofy, and UNiDAYS, where he has implemented security programs, achieved ISO certifications, and enhanced operational efficiency. Known for his expertise in IT service management, DevOps, and security frameworks, Karl is a strong advocate for workplace culture and continuous improvement.