Agentic AI digest (Sept 8–12, 2025)

#ai #aiops

Agentic AI digest (Sept 8–12 2025)

Dataminr brings agentic AI to cybersecurity platforms

On Sept 8 Dataminr announced that its Pulse for Cyber Risk service now includes agentic AI capabilities. New features—Live Briefs, Intel Agents and Cyber Anomaly Alerts—use Dataminr’s multi-modal fusion AI and ReGenAI models to provide context-rich threat assessments, with anomaly detection that spots “weak signals” across sources and triggers alerts in real time. The company said embedding these agents into SIEM and SOAR tools such as Splunk eliminates blind spots and reduces investigation time; integration with Palo Alto Networks’ XSOAR is coming soon. Ash D’Souza of Dataminr noted that agentic AI allows analysts to stay within their existing workflows while receiving timely recommendations, and partner Blackwood emphasized that the new agents remove complexity for customers.

Read the Dataminr announcement

Automotive agents: Google Cloud and Qualcomm

A joint press release from Google Cloud and Qualcomm on Sept 8 highlighted a deepened collaboration to bring agentic AI experiences to automobiles. Google’s Automotive AI Agent, powered by Gemini models, will run on Qualcomm’s Snapdragon Digital Chassis to deliver multimodal, hybrid edge-to-cloud agents for in-car navigation, entertainment and control. Automakers will be able to build branded AI companions that combine on-device responsiveness with cloud-based reasoning; the companies said the reference architecture will reduce development time and provide a foundation for new services like personalized assistants and context-aware recommendations. Executives from both companies touted the potential of agentic AI to redefine how drivers interact with vehicles.

Read the Google/Qualcomm announcement

Adobe unveils enterprise AI agents

On Sept 10 Adobe announced the general availability of AI agents within its Experience Platform. The new Agent Orchestrator includes a reasoning engine that interprets natural-language prompts and activates multiple agents as part of an orchestrated plan. Adobe introduced out-of-the-box agents for audience creation, journey orchestration, experimentation, data insights, site optimization and product support, all of which integrate with Adobe’s enterprise applications and can operate with a human-in-the-loop. The company also previewed Agent Composer, a tool that lets businesses customize agents to their brand guidelines and policies, and announced partnerships with firms like Cognizant, Google Cloud and PwC. Anjul Bhambhri, Adobe’s engineering SVP, said the goal is to embed agentic AI into data, content and experience workflows to unlock productivity and personalization.

Read the Adobe press release

NTT DATA frames agentic AI as the next wave in business services

In an opinion piece for its corporate blog on Sept 1, NTT DATA described agentic AI as ushering in a new era of “services as software” for business process outsourcing. Basudeb Bhaumik argued that intelligent agents will run entire workflows—from invoice reconciliation to claims assessment—without human handoffs, enabling modular business process-as-a-service (BPaaS) delivered via APIs. He noted that agentic BPaaS separates the “engine” (AI agents executing tasks) from the “dashboard” (the modular interface for subscribing to capabilities), giving companies agility and faster time to value. NTT DATA cited examples such as AI agents handling contact-center calls, insurance claims, account reconciliation and HR onboarding; one client reportedly cut case-handling time by 40 % while improving accuracy by 30 %. The article recommended that organizations align on outcomes, prepare data, design governance and modularize services before adopting agentic AI.

Read the NTT DATA blog

Supply chain benefits and case studies

Supply Chain Connect published a feature on Sept 15 (outside but relevant to the week) explaining how agentic AI is transforming manufacturing and logistics. The article noted that autonomous agents can automate workflows, track shipments in real time, optimize loads and routes, and analyze operations to suggest productivity improvements. By learning from prior tasks, agentic systems can distinguish genuine demand trends from outliers, enabling better forecasting and decision-making. The story highlighted how Cummins uses RFID tags and AI to track reusable transport containers, cutting packaging waste by millions of pounds and reducing bottlenecks.

Read the Supply Chain Connect article

Leadership skills for the agentic era

A CIO opinion piece discussed the leadership skills needed as agentic AI enters the enterprise mainstream. The author argued that executives must become “agent architects”, mastering multi-modal prompt design and overseeing agentic workflows; “innovation orchestrators”, using AI agents to free employees for creative and strategic work; and “ethical stewards”, establishing governance and risk-mitigation frameworks to ensure autonomous agents operate responsibly. With 76 % of surveyed companies planning to adopt agentic AI yet only 56 % aware of the risks, the article warned that leaders must balance innovation with accountability.

Read the CIO article

Takeaways

The second week of September 2025 highlighted the rapid spread of agentic AI across industries. Dataminr embedded autonomous agents into cybersecurity tools, Google and Qualcomm partnered to bring multimodal agents into cars, Adobe rolled out customizable agents for marketing and CX, and NTT DATA framed agentic BPaaS as the next frontier in outsourcing. At the same time, analysts urged organizations to develop new leadership skills and governance frameworks to harness these agents responsibly. From supply chains to customer journeys, agentic AI is shifting from hype to deployment, promising efficiency gains while raising fresh questions about oversight.