Why now

Attackers have hired AI. Most brands have not.

AI has made attacks faster, cheaper and harder to spot. Retail brands sit right in the path: lots of shoppers, lots of logins, card payments and big sale days. And in India, the rules on reporting and protecting data now have clocks attached.

Speed

Attacks now run at machine speed.

The time between a first foothold and real damage has shrunk from days to minutes, and sometimes seconds.

29 min

Average time for an attacker to move from first access to other systems in 2025. The fastest was 27 seconds.

Source: CrowdStrike, 2026
+89%

Rise in attacks by AI-enabled adversaries in a single year.

Source: CrowdStrike, 2026
80 to 90%

Share of a real espionage campaign run by an AI agent, with humans stepping in only at a few decision points.

Source: Anthropic, 13 Nov 2025
22 sec

Median time in 2025 for first access to be handed to a second criminal group. In 2022 it took more than eight hours.

Source: Google Cloud, Mandiant, 23 Mar 2026
$25M

Moved by one employee after a video call with deepfake versions of senior managers.

Source: World Economic Forum, 4 Feb 2025
₹22 crore

Average cost of a data breach in India in 2025. Breaches here took 263 days to find and contain.

Source: IBM, 7 Aug 2025
Retail

Every sale day is a working day for attackers.

These are the attacks we see most against commerce brands. Each one maps to a Fynd Guard or an agent in Fynd Secure.

Bots outnumber shoppers

51% of web traffic is now automated and 37% is bad bots. They scrape prices, hoard stock on drop days, test stolen cards and abuse coupons.

Source: Imperva, 14 Apr 2025
Accounts get taken over

Account takeover attacks rose 40% in a year. Attackers test breached passwords at scale and drain wallets, loyalty points and saved cards.

Source: Imperva, 14 Apr 2025
Checkout pages get skimmed

A single changed script on a payment page can copy card details. PCI DSS 4.0.1 now requires brands to control and watch those scripts.

Source: PCI Security Standards Council, 30 Jan 2025
Help desks get tricked

Social engineering of a third party led to ransomware at M&S, hit online and store operations and cut profit by about £300 million.

Source: BBC, 10 Jun 2025; CISA and partners, 29 Jul 2025
AI shopping agents get hijacked

Assistants that can look up orders or issue refunds can be steered by hidden instructions. OWASP now lists these as top agentic risks.

Source: OWASP, 9 Dec 2025
Leaders get impersonated

Voice and video deepfakes of founders and finance heads are now cheap to make and convincing enough to move money.

Source: World Economic Forum, 4 Feb 2025
India

The rules now come with clocks.

A short view of what applies to brands selling in India. Check the official text with your legal team.

  1. 28 Apr 2022
    CERT-In Directions

    Report covered cyber incidents to CERT-In within 6 hours of noticing them. Keep ICT system logs for 180 days, in India.

    Source: CERT-In, 28 Apr 2022
  2. 31 Mar 2025
    PCI DSS 4.0.1 checkout rules

    Requirements 6.4.3 and 11.6.1 apply: keep an inventory of payment page scripts, approve them and detect unauthorised changes.

    Source: PCI Security Standards Council, 30 Jan 2025
  3. 13 Nov 2025
    DPDP Rules notified

    The Digital Personal Data Protection Rules are published, with duties starting in phases.

    Source: MeitY, 13 Nov 2025
  4. 13 May 2027
    DPDP security and breach duties apply

    Reasonable security safeguards become mandatory. Tell the Data Protection Board and affected people without delay, and send a detailed report within 72 hours. Penalties go up to ₹250 crore.

    Source: MeitY, 13 Nov 2025; MeitY, 2023

This is a plain-language summary, not legal advice.

People

People alone cannot keep up.

The world is short of 4.8 million security people. Most brands cannot hire a 24x7 team, and even good teams cannot read every alert at 3 am on a sale night.

Source: ISC2, 23 Oct 2024

IBM found that heavy use of AI and automation in security cuts the cost of a breach sharply, yet 73% of Indian organisations it surveyed use little or none.

Source: IBM, 7 Aug 2025
What good looks like now
  • Detect in seconds, not days

    Guards sit inside Fynd and see attacks as they start. Agents read every signal, all day and all night.

  • Contain in minutes

    Safe, reversible actions run on their own. Bigger ones are prepared and sent to you with the blast radius.

  • See every system in one place

    Fynd properties, clouds, on-prem, SaaS, laptops and stores land in one timeline.

  • Keep people in charge

    You set what agents can do. Some actions are never automatic. Every action is explained and logged.

  • Report inside the clock

    CERT-In and DPDP clocks start on their own and drafts are ready for your team to review and submit.

Close the gap before your next sale.

Early access is open to Fynd customers now. We start in read-only shadow mode, so nothing changes until you say so.