What we see in the field.
What works, what fails, and how to think clearly about AI implementation. Our own writing, plus notable AI news.
All insights
How agents are transforming work
AI agents are moving from demos to deployments, but the ROI math is more complex than vendor pitches suggest. While demo environments show 10x productivity gains, production deployments typically deliver 1.5–3x returns after 18–30 months of integration work costing 4–6x the license fees. This guide helps operations leaders and IT executives at mid-market companies evaluate agent investments honestly, covering where ROI models break down, what predicts positive returns, and the critical questions to answer before greenlighting deployment.
AI Industry NewsWorkday Expands Enterprise AI Strategy with New Autonomous Agents – Cloud Wars
Workday's expansion into autonomous AI agents signals a major shift in enterprise HR and finance workflows. Unlike traditional automation that follows predefined rules, AI agents pursue goals and decide how to achieve them—creating new governance challenges most mid-market organizations aren't prepared to handle. This analysis examines what autonomous agents actually change, where operational risks concentrate, and the critical questions buyers should ask before committing to agent-enabled enterprise software.
AI Industry NewsHP Inc. launches Frontier strategic partnership with OpenAI
Enterprise AI partnerships like HP's OpenAI deployment make headlines, but most AI initiatives fail due to readiness gaps, not technology limitations. This article examines the three critical dimensions of AI readiness—data governance, process ownership, and absorption capacity—that vendor assessments ignore. For mid-market leaders evaluating AI strategy, we outline the essential questions to answer before moving past pilot, the hidden costs that run 40–70% above visible budgets, and the organizational foundations that determine whether AI deployments deliver value or become another underused tool.
AI Industry NewsBACA Systems Runs Its Entire Business on Headless 360
A 50-person robotics manufacturer eliminated $200,000 in annual ERP fees by consolidating onto a headless CRM platform. This case reflects a broader pattern: mid-market companies are abandoning traditional enterprise software stacks as integration costs become visible, ERP talent grows scarce, and workflow expectations shift. Learn the tradeoffs nobody mentions and how to assess whether platform consolidation applies to your business before your next vendor renewal.
AI Industry NewsSamsung Electronics brings ChatGPT and Codex to employees
Samsung's ChatGPT deployment to 270,000 employees makes headlines, but enterprise AI ROI tells a different story. Most deployments achieve measurable productivity gains in only 15–25% of use cases, with payback windows of 14–22 months. This analysis breaks down where generative AI actually delivers returns, the hidden costs beyond licensing, and why mid-market companies with 50 well-deployed seats can outperform massive rollouts.
AI Industry NewsAccess OpenAI models and Codex through your Oracle cloud commitment
Oracle Cloud customers can now access OpenAI models through existing cloud commitments, but treating procurement shortcuts as architecture decisions is where 70-80% of enterprise AI projects fail. Learn why cloud credits are a financing mechanism, not an implementation strategy, and what separates successful deployments from costly stalls.
AI Industry NewsFrom data to decisions: how LSEG is scaling trusted AI
Enterprise AI projects fail at the production stage, not the prototype stage. Scaling AI is fundamentally a governance problem—success depends on defining data ownership, approval workflows, and accountability structures before expanding infrastructure. Learn why pilots succeed in controlled conditions while production introduces data quality challenges, organizational friction, and accountability gaps that determine whether your AI investment delivers value or becomes an expensive rebuild within eighteen months.
AI Industry NewsBBVA puts AI at the core of banking with OpenAI
Enterprise AI deployments like BBVA's 100,000-seat rollout make headlines, but the ROI reality is more complex. Most organizations see measurable productivity gains from only 15–25% of users, while 40–50% quietly return to old workflows. This analysis breaks down the true cost structure—where license fees represent just 25–35% of first-year investment—and identifies the specific conditions where enterprise AI math actually works: high-volume knowledge work, bottlenecked expertise, and committed leadership. For mid-market companies considering similar investments, expect Year 1 to be net negative, breakeven in Year 2, and ROI by Year 3 only if adoption holds.
AI Industry NewsHow Endava is redesigning software delivery around AI agents
AI agent pilots succeed technically but fail organizationally when companies lack decision frameworks for scaling. Learn the three deployment paths, realistic cost models including the productivity dip most vendors ignore, and the five questions that determine whether your AI initiative captures value or stalls between pilot and production.
AI Industry NewsStill Running Agentic Pilots? Here’s What 5 Companies Did to Ship AI Agents to Production
Five companies ran agentic AI pilots last year, but only two shipped to production. The difference wasn't model accuracy—it was building exception handling, human oversight systems, and failure scaffolding before optimizing the happy path. Learn why teams that accept 85% accuracy with robust fallbacks outperform those waiting for perfect models, and get the production checklist that separates shippers from eternal refiners.
AI Industry NewsTravelers deploys AI-powered claims countrywide with OpenAI
Travelers' nationwide AI claims deployment reveals what enterprise rollouts actually require. For mid-market insurers, the model isn't the hard part—it's the 18 months of process redesign, compliance review, and change management that determines success. Learn why 85% of AI deployments stall and the three readiness signals that separate measurable ROI from expensive pilots.
AI Industry NewsBoston Children’s uses AI to unlock new diagnoses
Healthcare AI pilots consistently fail to reach production—not because the models don't work, but because organizations underestimate the integration, workflow redesign, and clinician trust-building required. Learn why Boston Children's Hospital succeeded where most health systems stall, and what the successful 20% of AI deployments do differently.
AI Industry NewsHow Endava builds an agentic organization with Codex
Enterprise software vendors promise agentic AI will automate workflows and accelerate delivery, but most organizations will fail—not because the technology doesn't work, but because they're building agents into structures designed for human bottlenecks. This guide examines what actually changes when you deploy AI that acts rather than suggests, the hidden costs vendors omit from proposals, and why organizations taking the redesign path see 50–70% reductions in cycle time while those on the automation path struggle to move the needle.
AI Industry NewsCisco and OpenAI redefine enterprise engineering with Codex
Enterprise AI projects fail not at the proof-of-concept stage, but when organizations try to scale into production. This analysis examines why Cisco and OpenAI's engineering partnership highlights a common pattern: early wins create excitement, but the integration valley—data debt, workflow disruption, and maintenance overhead—derails most initiatives. Learn what the 20% of successful implementations share and the critical questions to ask before scaling any AI pilot.
AI Industry NewsMUFG aims to become AI-native with OpenAI
Enterprise AI adoption stories follow a predictable arc in vendor case studies: executive sponsor, pilot success, organization-wide rollout, transformed culture. What these narratives leave out is the invoice…
AI Industry NewsAdventHealth advances whole-person care with OpenAI
AdventHealth's deployment of ChatGPT for clinical documentation signals that healthcare AI has moved from experimental to operational. But for mid-market health systems, the real challenges aren't the AI models—they're EHR integration costs, physician adoption curves that stall at 50%, and compliance overhead that stretches ROI timelines to 18 months. This analysis breaks down what the headline numbers miss and offers a decision framework for healthcare IT leaders weighing AI investments against tighter margins.
AI Industry NewsDatabricks brings GPT-5.5 to enterprise agent workflows
Databricks integrates GPT-5.5 into enterprise agent workflows, but benchmark performance tells only part of the story. For mid-market IT leaders evaluating agentic AI, the real costs hide in integration engineering, guardrail development, monitoring infrastructure, and months of prompt iteration—typically 3-5x the quoted licensing fees. Learn which use cases deliver ROI, what questions to ask before committing, and why high-volume, low-variance processes offer the fastest payback on agent automation investments.
AI Industry NewsOpenAI and Dell partner to bring Codex to hybrid and on-premise enterprise environments
Enterprise AI coding tools fail in deployment more often than in demos. The gap between proof-of-concept and production stems from infrastructure, governance, and workflow challenges—not model capabilities. Learn why hybrid and on-premise deployments cost 2-4x more than cloud alternatives, why productivity gains typically run 10-20% rather than the 30-55% vendors promise, and what the successful 20% of deployments do differently to achieve real ROI from AI coding assistants.
AI Industry NewsServiceNow CEO Bill McDermott: Silicon Valley is getting enterprise AI wrong – Fast Company
Enterprise AI has a delivery problem, not a technology problem. The gap between vendor demos and production reality is measured in months, millions, and executive patience. This analysis explores why Silicon Valley's consumer-tech thinking breaks in enterprise environments, what the vendor pitch leaves out, and how mid-market companies can avoid the common pitfalls that kill AI projects before they deliver value.
AI Industry NewsHow BCU Is Transforming Banking Service with Agentforce
BCU's Agentforce deployment highlights what vendor case studies leave out: realistic ROI timelines for AI agents in financial services. Most mid-market credit unions reach payback in 14–18 months, not 90 days. This analysis breaks down the three-phase value model, where integration costs actually live, and what separates organizations that hit positive ROI faster from those that stall.
AI Industry NewsOpenAI launches DeployCo to help businesses build around intelligence
Most enterprise AI projects fail not because the model underperforms, but because organizations cannot absorb what the technology produces. Learn why 70-80% of AI initiatives stall during deployment and what the successful minority does differently—from integration budgeting to governance frameworks and adoption strategies that actually drive business value.
AI Industry NewsParloa builds service agents customers want to talk to
Voice AI for customer service fails not because the technology doesn't work, but because organizations underestimate the gap between prototype and production. This framework helps operations leaders and IT executives evaluate three approaches—platform-first, build-on-foundation, or agent-assist—based on actual costs, implementation timelines, and organizational readiness rather than vendor capabilities alone.
AI Industry NewsSingular Bank helps bankers move fast with ChatGPT and Codex
Singular Bank built an internal AI assistant using ChatGPT and Codex that saves bankers 60 to 90 minutes daily on meeting prep, portfolio analysis, and follow-up tasks. The key to their success: targeting three specific workflows instead of launching an enterprise-wide transformation. Learn why most AI banking projects fail, what the real costs look like, and how mid-market financial services firms can assess whether building an internal AI assistant makes sense for their organization.
AI Industry NewsOpenAI and PwC collaborate to reimagine the office of the CFO
Finance leaders evaluating AI partnerships like OpenAI-PwC should understand what these collaborations actually deliver versus what they promise. This analysis breaks down where finance AI projects fail, the true cost structure beyond vendor proposals, and what successful organizations do differently—from mapping exception-handling workflows to negotiating exit rights before signing contracts.
We also republish notable AI news from elsewhere in the industry. Browse AI industry news
Reading is not the same as starting.
The 90-day plan is the version of all this you can hand to your leadership team.
Bring us the pilot that stalled.
Or the one you have not started because nobody can tell you what it would cost.
You'll leave with 2 to 3 scored use cases, an effort estimate, and an honest cost range, whether or not we work together.