AI Enablement Radar week 18: from pilots to governed everyday work
Adam Olofsson Hammare
The week 18 AI radar has one clear pattern: AI enablement is no longer about finding one more chatbot. It is about building workflows where people, data, permissions and follow-up fit together.
For Hammer Automation's best-fit readers — Nordic small businesses with 1–10 people, solo operators, school leaders and admin-heavy teams — the practical lesson is simple: do not copy the enterprise budget, but do copy the discipline. Start with one real workflow, name an owner and decide which data AI is allowed to use.
Definition: An agentic workflow is an AI-supported workflow where the system can plan several steps, use tools and return an output for human review. Governance means rules, ownership, logs and checkpoints that make AI use traceable and safe.
Top news this week
-
Microsoft and Accenture scale Copilot to 743,000 people. Accenture began rolling out Microsoft 365 Copilot to a workforce the size of a large city, described as Copilot's largest enterprise deployment so far. For small teams, the lesson is not the size; it is the pattern: begin with tasks such as meeting notes, email drafts and document work where value appears quickly.
Source: Microsoft Source: Accenture is rolling out Copilot to a workforce the size of Denver
-
Google Cloud introduces Gemini Enterprise Agent Platform. The platform brings agent development, governance and deployment into one enterprise environment. For non-technical teams, the signal is that the next phase is not only better chat answers, but safer connections between AI, documents, systems and identity.
Source: Google: Gemini Enterprise Agent Platform optimizes your agents
-
Accenture and Deloitte build services around Gemini Enterprise. Accenture describes support for organizations adopting agentic AI with Gemini Enterprise, while Deloitte launched a dedicated Google Cloud Agentic Transformation Practice. The market is moving from “which tool?” to “who helps us make the work pattern function?”.
Source: Accenture Helps Organizations Advance Agentic AI with Gemini Enterprise
Source: Deloitte Launches Google Cloud Agentic Transformation Practice
-
IBM Bob launches as an AI development partner with checkpoints. IBM describes Bob as an AI-first development partner for enterprise teams, with multi-model orchestration and controls across the software development lifecycle. Even if small businesses do not build large code platforms, the principle matters: use AI where there are human stop-points before anything goes live.
Source: IBM: Introducing IBM Bob
-
The EU AI Act is moving toward practical deadlines. Holland & Knight highlights that U.S. companies with AI systems used in the EU may need to watch August 2026, while IAPP follows the Omnibus debate. For Swedish small businesses and schools, this is a reminder: write down which AI tools you use, what they are used for and who approves sensitive cases.
Source: Holland & Knight: U.S. Companies Face EU AI Act's Possible August 2026 Compliance Deadline
Source: IAPP: AI Act Omnibus: What just happened and what comes next?
Enterprise Adoption
-
Merck and The Home Depot use Gemini Enterprise for agentic AI. Merck describes a multi-year Google Cloud collaboration valued at up to $1 billion for R&D, manufacturing, commercial and corporate functions. The Home Depot is equipping thousands of associates with Gemini Enterprise for agentic workflows. Small teams should translate this into one question: which recurring workflow would benefit most if AI could read the right context and suggest the next step?
Source: Merck and Google Cloud Partner to Accelerate Agentic AI Enterprise Transformation
Source: The Home Depot and Google Cloud Launch Agentic AI Tools
-
Novo Nordisk and Customers Bank use OpenAI in core processes. Novo Nordisk connects OpenAI to R&D, manufacturing and commercial functions. Customers Bank describes a multi-year collaboration for commercial banking. For small Nordic organizations, the lesson is that AI projects should map to a measurable process: quoting, onboarding, documentation, customer service or lesson planning.
Source: Pharmaceutical Technology: Novo Nordisk and OpenAI forge alliance
Source: Customers Bank Announces Strategic Collaboration with OpenAI
Tools & Platforms
-
Google Cloud, ServiceNow and Adobe move toward agent platforms. Google Cloud positions Gemini Enterprise Agent Platform as a place to build, scale, govern and optimize agents. ServiceNow and Google Cloud connect AI agents to autonomous enterprise operations. Adobe launches CX Enterprise for agentic customer experience flows. Workflow automation means connecting repetitive work steps so information can move between people and systems without unnecessary manual re-entry.
Source: Google: Gemini Enterprise Agent Platform optimizes your agents
Source: ServiceNow and Google Cloud Unite AI Agents for Autonomous Enterprise Operations
Source: Adobe News: Adobe Redefines Customer Experience Orchestration Vision
-
Runloop and Weights & Biases point to AI agent testing. Runloop launched a benchmark orchestration platform with a Weights & Biases integration. Benchmarking means testing an AI system against known tasks so quality, cost and errors can be compared over time. For small teams, that can be as simple as ten real customer requests or student questions tested before a new routine is adopted.
Source: Runloop Launches Industry-First Benchmark Orchestration Platform
Use Cases
-
Professional services: Accenture shows that Copilot can become an everyday tool in consulting work. Small agencies can start with meeting minutes, summaries, proposals and internal knowledge bases.
Source: Microsoft Source: Accenture is rolling out Copilot to a workforce the size of Denver
-
Health and public sector: WHO/Europe published a first snapshot of AI in health care across the 27 EU Member States. Schools, municipal-adjacent operations and smaller health providers should read this as a governance signal: AI can help, but it must be documented and followed up.
Source: WHO/Europe: First-ever snapshot of AI in health care across EU Member States
-
Customer service: Netomi raised $110 million with Accenture Ventures and Adobe Ventures involved for agentic customer experiences. For small businesses, this does not mean buying enterprise CX immediately. It means customer service is one of the clearest areas for testing AI with human review.
Source: VentureBeat: Netomi raises $110 million as Accenture and Adobe bet on AI for customer service
Governance & Regulation
-
NIST prepares an AI RMF profile for critical infrastructure. AI RMF means AI Risk Management Framework: a framework for identifying, measuring and managing risks in AI systems. Even small businesses can use the same thinking in miniature: list risks, set boundaries and document who is allowed to approve what.
Source: NIST: Concept Note: AI RMF Profile on Trustworthy AI in Critical Infrastructure
-
ProcureAbility shows a governance gap between procurement and IT. Its 2026 CPO-CIO report says weak collaboration between procurement and IT limits AI deployments. For a 1–10 person company, the equivalent is simple: the person buying the tool and the person owning the workflow must assess risk together.
Key signals for small teams
- Large companies buy platforms; small teams need methods. The key question is not whether you choose Microsoft Copilot, Gemini, ChatGPT or Claude. The key question is whether you know which tasks the tool may support, which data is approved and when a human must say yes.
- A sandbox is needed. A sandbox is a limited test environment where AI can be tried without affecting real customers, students, invoices or business-critical systems.
- Measure before you scale. Choose three everyday metrics: time saved, errors caught and how often the team actually uses the routine.
Companies to watch
-
Netomi: agentic customer experience with new funding and strong enterprise interest.
-
Runloop: infrastructure for testing and traceability of AI agents.
-
AIQA Global: an independent AI governance rating system, signaling that boards and insurers want evidence, not only AI policies.
-
Version 1 and CreateFuture: the consulting market is consolidating around AI-first digital transformation.
Source: AIQA Global Launches First Independent Rating System
What you can test next week
- Choose one workflow: for example incoming customer questions, quote preparation, lesson planning, meeting follow-up or an internal FAQ.
- Write one simple AI rule: what may be entered, what may not be entered, and who reviews the output?
- Build a mini-benchmark: collect ten real examples and test the same prompt or routine every week.
- Decide the next step: if the workflow saves time without creating extra risk, it can become a first candidate for Tool Forge or Skill Forge with Hammer Automation.
