{
  "id": 34174,
  "title": "6 Questions Every Enterprise Has to Answer About AI",
  "url": "https://urgent.news/2026/08/02/6-questions-every-enterprise-has-to-answer-about-ai",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-02T07:42:30.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/hunter_g_50e2ec233acd07b5/6-questions-every-enterprise-has-to-answer-about-ai-2f9k"
  },
  "original_language": "en",
  "account": "There are six crucial questions enterprises need to address in the age of AI, as outlined at KPMG's annual Tech and Innovation Symposium. These questions have shifted from the previous concerns about adopting AI and proving ROI to more foundational issues. Last year, businesses were focused on integrating AI into their systems, but this year, the conversation has transformed into understanding how to navigate the new problems that agentic AI brings.\n\nFirstly, the shift from assisted AI to agentic AI has necessitated a rethinking of how organizations approach AI. This transition is no longer just about using AI to assist in work but as a tool that can perform work autonomously. As such, enterprises need to consider whether they should be redesigning their processes or merely adding AI functionality on top of existing systems. Steve Chase, from KPMG, warned against the former, stating that bolting an AI strategy onto existing processes and systems could lead to significant problems due to under-utilization and increased costs.\n\nSecondly, organizations must shift from selecting vendors to designing architectures. In the assisted-AI era, the approach was to choose the best vendor for each problem. However, in the agentic AI era, enterprises need to think in terms of architectures that include multi-model systems with different intelligence tiers, routing layers to direct requests to the appropriate model, and design mechanisms for integrating people and functions with the relevant context, data, and system integrations, all underpinned by guardrails.\n\nThirdly, the question of how to provision costs across different groups in an organization is becoming increasingly critical. This involves monitoring and measuring AI usage, which has been likened to the prevalence of the term \"token\" during the peak of the crypto era. Visibility into AI cost and its relationship to output is essential for deciding which individuals, teams, or projects should have access to specific models and at what level, ensuring effective cost management.\n\nFourthly, the messy work of enablement is highlighted as more than just producing training videos. It's about real work, pushing teams to use new tools for new tasks and then figuring out how to share this knowledge across different parts of the organization. This involves pairing AI-redesigned engineering teams and early adopters with various business units to ensure that the essential skills and mindsets become part of the essential toolkit for all departments, including marketing, sales, and back-office operations.\n\nFifthly, the external dimension of AI transformation is gaining attention. Enterprises are moving away from input-based pricing to outcomes-based pricing, exploring new categories of products and services, and re-evaluating the very nature of their offerings. If AI-enabled agents can run audits persistently, the traditional notion of an audit might change. Most organizations are treating themselves as \"patient zero\" in this transformation, addressing internal transformation first before considering radical changes to their external offerings. However, this needs to be done in real-time, balancing the need to service legacy customers with the drive to innovate.\n\nLastly, enterprises must design for obsolescence. As they build new systems, they need to anticipate that these systems will likely need to be rebuilt a few months after deployment due to continuous changes in harnesses, interaction patterns, customer expectations, markets, policies, and other factors. This requires building flexibility into their systems to adapt to the dynamic nature of the AI landscape, highlighting that AI isn't just about building and installing tools but redesigning the entire organization around agentic work.",
  "summary": "You've probably sat in this meeting. Someone asks, \"so where are we with AI, actually?\" — and six people give six different answers. Last year, the questions were \"should we do this?\" and \"how do we prove ROI?\" This year, those questions have mostly disappeared. What replaced them are harder, more foundational questions. If nobody at your company is asking them yet, that's the thing to worry…",
  "key_points": [
    "Enterprises must rethink processes vs. adding AI functionality to agentic AI.",
    "Shift from vendor selection to designing architectures with multi-model systems.",
    "Cost provisioning across groups becomes critical with AI usage monitoring."
  ],
  "editors_take": null,
  "illustration": "https://urgent.news/ill/34174.png",
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}