Quote for the day:
"Listen with curiosity, speak with honesty act with integrity." -- Roy T Bennett
AI Agents Transform Enterprise Application Development
There's now discussion about the agent development life cycle and the need to
supervise or manage AI agent developers - calling for agent governance and
infrastructure changes. New products, services and partnerships announced in
the past few weeks support this trend. ... Enterprises were cautious about
entrusting public models and agents with intellectual property. But the
partnership with Anthropic could make models more trustworthy. "Enterprises
are looking for AI they can actually trust with their code, their data and
their day-to-day operations," said Mike Krieger, chief product officer at
Anthropic. ... Embedding agentic AI within the fabric of enterprise
architecture enables organizations to unlock transformative agility, reduce
cognitive load and accelerate innovation - without compromising trust,
compliance or control - says an IBM report titled "Architecting secure
enterprise AI agents with MCP." Developers adopted globally recognized models
such as Capability Maturity Model Integration, or CMMI, and CMMI-DEV as paths
to improve the software development and maintenance processes. ... Enterprises
must be prepared to implement radical process and infrastructure changes to
successfully adopt AI agents in software delivery. AI agents must be managed
by a central governance framework to enable complete visibility into agents,
agent performance monitoring and security.There’s no such thing as quantum incident response – and that changes everything
CISOs are directing attention to have quantum security risks added to the
corporate risk register. It belongs there. But the problem to be solved is not
a quick fix, despite what some snake oil salesmen might be pushing. There is
no simple configuration checkbox on AWS or Azure or GCP where you “turn on”
post-quantum cryptography (PQC) and then you’re good to go. ... Without
significant engagement from developers, QA teams and product owners, the
quantum decryption risk will remain in play. You cannot transfer this risk by
adding more cyber insurance policy coverage. The entire cyber insurance
industry itself is in a bit of an existential doubt situation regarding
whether cybersecurity can reasonably be insured against, given the systemic
impacts of supply chain attacks that cascade across entire industries. ...The
moment when a cryptographically relevant quantum computer comes into existence
won’t arrive with fanfare or bombast. Hence, the idea of the silent boom. But
by then, it will be too late for incident response. What you should do Monday
morning: Start that data classification exercise. Figure out what needs
protecting for the long term versus what has a shorter shelf life. In the
world of DNS, we have Time To Live (TTL) that declares how long a resolver can
cache a response. Think of a “PQC TTL” for your sensitive data, because not
everything needs 30-year protection.Hackers Use Blockchain to Hide Malware in Plain Sight
At least two hacking groups are using public blockchains to conceal and
control malware in ways that make their operations nearly impossible to
dismantle, shows research from Google's Threat Intelligence Group. ... The
technique, known as EtherHiding, embeds malicious instructions in blockchain
smart contracts rather than traditional servers. Since the blockchain is
decentralized and immutable, attackers gain what the researchers call a
"bulletproof" infrastructure. The development signals an "escalation in the
threat landscape," said Robert Wallace, consulting leader at Mandiant, which
is part of Google Cloud. Hackers have found a method "resistant to law
enforcement takedowns" that and can be "easily modified for new campaigns."
... The group over time expanded its architecture from a single smart contract
to a three-tier system mimicking a software "proxy pattern." This allows rapid
updates without touching the compromised sites. One contract acts as a router,
another fingerprints the victim's system and a third holds encrypted payload
data and decryption keys. A single blockchain transaction, costing as little
as a dollar in network fees, can change lure URLs or encryption keys across
thousands of infected sites. The researchers said the threat actor used social
engineering tricks like fake Cloudflare verification or Chrome update prompts
to persuade victims to run malicious commands.Everyone’s adopting AI, few are managing the risk
Across industries, many organizations are caught in what AuditBoard calls the
“middle maturity trap.” Teams are active, frameworks are updated, and risks
are logged, but progress fades after early success. When boards include risk
oversight as a standing agenda item and align on shared performance goals,
activity becomes consistent and forward-looking. When governance and ownership
are unclear, adoption slows and collaboration fades. ... Many enterprises are
adopting or updating risk frameworks, but implementation depth varies. The
typical organization maps its controls to several frameworks, while leading
firms embed thousands of requirements into daily operations. The report warns
that “surface compliance” is common. Breadth without depth leaves gaps that
only appear during audits or disruptions. Mature programs treat frameworks as
living systems that evolve with business and regulatory change. ... The
findings show that many organizations are investing heavily in risk management
and AI, but maturity depends less on technology and more on integration.
Advanced organizations use governance to connect teams and turn data into
foresight. AuditBoard’s research suggests that as AI becomes more embedded in
enterprise systems, risk leaders will need to move beyond activity and focus
on consistency. Those that do will be better positioned to anticipate change
and turn risk management into a strategic advantage.A mini-CrowdStrike moment? Windows 11 update cripples dev environments
The October 2025 cumulative update, (KB5066835), addressed security issues in
Windows operating systems (OSes), but also appears to have blocked Windows’
ability to talk within itself. Localhost allows apps and services to
communicate internally without using internet or external network access.
Developers use the function to develop, test, and debug websites and apps
locally on a Windows machine before releasing them to the public. ... When
localhost stops working, entire application development environments can be
impacted or “even grind to a halt,” causing internal processes and services to
fail and stop communicating, he pointed out. This means developers are unable
to test or run web applications locally. This issue is really about “denial of
service,” where tools and processes dependent on internal loopback services
break, he noted. Developers can’t debug locally, and automated testing
processes can fail. At the same time, IT departments are left to troubleshoot,
field an influx of service tickets, roll back patches, and look for
workarounds. “This bug is definitely disruptive enough to cause delays, lost
productivity, and frustration across teams,” said Avakian. ... This type of
issue underscores the importance of quality control and thorough testing by
third-party suppliers and vendors before releasing updates to commercial
markets, he said. Not doing so can have significant downstream impacts and
“erode trust” in the update process while making teams more cautious about
patching.How Banks of Every Size Can Put AI to Work, and Take Back Control
For smaller banks and credit unions, the AI conversation begins with math.
They want the same digital responsiveness as larger competitors but can’t
afford the infrastructure or staffing that traditionally make that possible.
The promise of AI, especially low-code and automated implementation, changes
that equation. What once required teams of engineers months of coding can now
be deployed out-of-the-box, configured and pushed live in a day. That shift
finally brings digital innovation within reach for smaller institutions that
had long been priced out of it. But even when self-service tools are
available, many institutions still rely on outside help for routine changes or
maintenance. For these players, the first question is whether they’re willing
or able to take product dev work inhouse, even with "AI inside"; the next
question is whether they can find partners that can meet them on their own
terms. ... For mid-sized players, the AI opportunity centers on reclaiming
control. These institutions typically have strong internal teams and clear
strategic ideas, yet they remain bound by vendor SLAs that slow innovation.
The gap between what they can envision and what they can deliver is wide.
AI-driven orchestration tools, especially those that let internal teams
configure and launch digital products directly, can help close that gap. By
removing layers of technical dependency, mid-sized institutions can move from
periodic rollouts to something closer to iterative improvement. Why your AI is failing — and how a smarter data architecture can fix it
Traditional enterprises operate four separate, incompatible technology stacks,
each optimized for different computing eras, not for AI reasoning
capabilities. ... When you try to deploy AI across these fragmented stacks,
chaos follows. The same business data gets replicated across systems with
different formats and validation rules. Semantic relationships between
business entities get lost during integration. Context critical for
intelligent decision-making gets stripped away to optimize for system
performance. AI systems receive technically clean datasets that are
semantically impoverished and contextually devoid of meaning. ... As
organizations begin shaping their enterprise general intelligence (EGI)
architecture, critical operational intelligence remains trapped in
disconnected silos. Engineering designs live in PLM systems, isolated from the
ERP bill of materials. Quality metrics sit locked in MES platforms with no
linkage to supplier performance data. Process parameters exist independently
of equipment maintenance records. ... Enterprises solving the data
architecture challenge gain sustainable competitive advantages. AI deployment
timelines are measured in weeks rather than months. Decision accuracy reaches
enterprise-grade reliability. Intelligence scales across all business domains.
Innovation accelerates as AI creates new capabilities rather than just
automating existing processes.Under the hood of AI agents: A technical guide to the next frontier of gen AI
With agents, authorization works in two directions. First, of course, users
require authorization to run the agents they’ve created. But as the agent is
acting on the user’s behalf, it will usually require its own authorization to
access networked resources. There are a few different ways to approach the
problem of authorization. One is with an access delegation algorithm like
OAuth, which essentially plumbs the authorization process through the agentic
system. ... Agents also need to remember their prior interactions with their
clients. If last week I told the restaurant booking agent what type of food I
like, I don’t want to have to tell it again this week. The same goes for my
price tolerance, the sort of ambiance I’m looking for, and so on. Long-term
memory allows the agent to look up what it needs to know about prior
conversations with the user. Agents don’t typically create long-term memories
themselves, however. Instead, after a session is complete, the whole
conversation passes to a separate AI model, which creates new long-term
memories or updates existing ones. ... Agents are a new kind of software
system, and they require new ways to think about observing, monitoring and
auditing their behavior. Some of the questions we ask will look familiar:
Whether the agents are running fast enough, how much they’re costing, how many
tool calls they’re making and whether users are happy. Data Is the New Advantage – If You Can Hold On To It
Proprietary data has emerged as one of the most valuable assets for
enterprises—and increasingly, the expectation is that data must be stored
indefinitely, ready to fuel future models, insights, and innovations as the
technology continues to evolve. ... Globally, data architects, managers, and
protectors are in uncharted territory. The arrival of generative AI has proven
just how unpredictable and fast-moving technological leaps can be – and if
there’s one thing the past few years have taught us, it’s that we can’t know
what comes next. The only way to prepare is to ensure proprietary data is not
just stored but preserved indefinitely. Tomorrow’s breakthroughs – whether in
AI, analytics, or some other yet-unimagined technology – will depend on the
depth and quality of the data you have today, and how well you can utilize the
storage technologies of your choice to serve your data usage and workflow needs.
... The lesson is clear: don’t get left behind, because your competitors are
learning these lessons as well. The enterprises that thrive in this next era of
digital innovation will be those that recognize the enduring value of their
data. That means keeping it all and planning to keep it forever. By embracing
hybrid storage strategies that combine the strengths of tape, cloud, and
on-premises systems, organizations can rise to the challenge of exponential
growth, protect themselves from evolving threats, and ensure they are ready for
whatever comes next. In the age of AI, your competitive advantage won’t just
come from your technology stack.
No comments:
Post a Comment