Latest Articles

Problem Management Delivers 10x ROI — Here's How to Calculate It The Business Case Problem Management reduces : Volume and impact of incidents Rework and repetitive fixes Wasted Subject Matter Expert (SME) time Friction between teams, vendors, and customers Problem Management increases : Service stability Operational confidence Quality of data for leadership decision-making Customer satisfaction and trust Documented ROI Examples Real case example : Benefit Area Impact Number of incidents decreased 4-6% Time for incident solution reduced 95% < 5 days Solver team capacity savings < 5% Reduced severity of incident impact on users < 3% Increase in end-user satisfaction 2-3% ROI Calculation Framework Direct Savings : Savings on solver teams capacity 5% of solver time saved 2,885 man-days saved annually Cost savings: significant Savings on incident solving process Fewer incidents to process Faster resolution Less rework Savings on SLA breach penalties Higher SLA compliance Fewer penalties Indirect Savings : Improved service quality leading to higher employee satisfaction Reduced risk of incidents impacting the business Knowledge reuse Sample ROI Calculation From a ServiceNow ROI model : Item Value Agents on the desk 10 Agent working hours per week 40 Weekly hours available 400 Weekly incident capacity 800 Avoidable incidents through Problem Management 2% Weekly incident saving 16 Yearly cost saving $8,320 Agent saving (FTE) 0.2 Real-World Investment Analysis From a documented case study : Item Value Investment €85,000 Benefits Year 1 €1,205,064 Net Annual Benefits Year 1 €1,188,064 NPV €995,058 IRR 1298% Payback Period 0.08 years Self-Funding Problem Management The FTE saving can be used to either provide a higher capacity on the desk or, and potentially more productively, contribute to the problem management process and drive further savings. In this way, there is a level of self-funding from implementing the process. How to Build Your Business Case Collect current data: Incident volume Incident resolution time SLA performance Support costs Estimate potential savings: What percentage of incidents are recurring? What is the cost of each incident? How much time could be saved? Project investment costs: Tooling Training Ongoing costs Present the business case: ROI calculation Payback period Non-financial benefits Conclusion The economic case for Problem Management is compelling. Organizations that invest in problem management see significant ROI through reduced incidents, improved efficiency, and better service quality. Action Items for Your Organization Collect data on incident costs Estimate potential savings from Problem Management Build a business case for investment Track actual savings after implementation Demonstrate value to leadership

Your AI Can Find the Root Cause in Seconds. Why Does It Still Take Hours to Fix the Problem? The Paradox Within the Paradox In Topic 1, we established the broader AI incident management paradox: AI is supposed to reduce work, but 44% of IT teams are spending more time on incident response. But there's a deeper paradox within that data. 61% of IT professionals say AI has accelerated root cause analysis —a genuine, measurable win. Yet 71% still manually double-check AI outputs , and 62% report difficulty trusting AI recommendations . This creates an absurd situation: your AI can identify the root cause in seconds, but your teams take hours to act because they don't trust what the AI is telling them. The "trust tax" is eroding the efficiency gains AI promises. Why Trust Is Different in Incident Management Trust in AI for incident management is fundamentally different from trust in AI for other applications. In a service desk context, trust isn't a nice-to-have—it's an operational necessity. The Cost of Being Wrong In incident management, the cost of AI error can be catastrophic. An incorrect root cause identification can send teams down the wrong path, wasting critical time during an outage. An AI-proposed fix that makes things worse can extend downtime and damage customer trust. The Speed of Decision-Making Incident management requires rapid decisions with incomplete information. When AI provides a recommendation, teams must decide whether to trust it in seconds, not hours. The pressure of the moment makes trust assessment harder. The Blame Factor If a human makes a mistake during an incident, there's a postmortem and a learning opportunity. If a human trusts an AI that makes a mistake, accountability becomes unclear. Who's responsible when AI gets it wrong? The Trust Tax in Action Consider how the trust gap manifests in real incident scenarios: Scenario AI Capability Trust Gap Impact Incident categorization AI assigns priority level Team verifies every categorization before routing Root cause identification AI identifies potential cause Team investigates independently, wasting time Resolution recommendation AI suggests fix Team researches whether the fix is safe Automated remediation AI can execute fix automatically Team disables automation, reverts to manual Predictive alerting AI detects anomaly Team treats as noise until manually verified Each verification step adds time, cognitive load, and frustration. The AI becomes a source of extra work rather than a productivity tool. Why Trust Is Low: The Data Quality Problem The root cause of low trust isn't just human psychology—it's data quality. 83% of IT professionals agree that AI is only as effective as the breadth and quality of data it can access . When AI models are trained on fragmented, inconsistent, or incomplete data, they produce unreliable outputs. And when teams see unreliable outputs, they stop trusting the AI. It's a vicious cycle: poor data leads to poor AI recommendations, which leads to low trust, which leads to manual verification, which reduces efficiency. The reality is that most IT organizations have spent years building operational processes on top of data that's incomplete, outdated, or inaccurate. They're now deploying AI on top of that foundation and wondering why it's not working. Building Trust Through Explainable AI The most effective approach to building trust is explainable AI (XAI) —AI that can show its reasoning in language humans can understand. What Explainable AI Looks Like in Incident Management AI Output Unexplainable AI Explainable AI Incident priority "Priority 1" "Priority 1 because: this service supports 5,000+ users, it's a revenue-generating application, and we've seen similar patterns lead to widespread outages" Root cause "Database connection issue" "Database connection issue affecting 3 of 12 replicas in us-east-1 region. This matches the pattern from the October 12 incident. Resolution from that incident was restarting replica 3 and 7." Resolution recommendation "Restart service" "Restart service X because: memory leak detected in logs, restart typically resolves within 2 minutes, and we've run this successfully 14 times in the last 30 days" Automation decision Auto-execute "I've identified a fix that I'm 94% confident will resolve. I'm requesting approval to execute. The fix is: restart service X. I've verified this works in 14 of 15 previous cases." Explainable AI builds trust by showing its work. Teams can follow the reasoning, verify the logic, and make informed decisions about whether to trust the recommendation. Strategies for Reducing the Trust Gap 1. Implement AI "Confidence Scoring" Every AI recommendation should include a confidence score. "I'm 94% confident this is the root cause" is more useful than "Here's the root cause." Teams can calibrate their verification effort based on confidence: Confidence Level Response 90-100% Review, typically execute 70-89% Review carefully, likely execute 50-69% Review thoroughly, escalate if uncertain Below 50% Treat as suggestion, investigate independently 2. Build AI Performance Dashboards Transparency about AI performance builds trust. Dashboards should show: Accuracy rates by incident type False positive/negative rates Time saved by AI Areas where AI struggles 3. Start with Low-Stakes Incidents Build trust with low-risk incidents before expanding to critical systems. Let AI prove itself on Tier 2 and Tier 3 incidents before moving to Tier 1. 4. Create AI Validation Workflows Design workflows where AI is used for initial triage and humans review outputs. Over time, as trust builds, expand AI autonomy. 5. Develop AI Training for Teams Teams need to understand how AI works, what it can and can't do, and how to interpret its outputs. Training reduces uncertainty and builds confidence. Measuring the Trust Tax To manage the trust gap, you need to measure it. Key metrics include: Metric What It Measures Target Time spent verifying AI outputs Trust tax in minutes Trend downward over time AI adoption rate Percentage of teams using AI recommendations 80%+ for non-critical incidents AI override rate Percentage of AI recommendations overridden by humans Decreasing over time AI accuracy Percentage of correct AI recommendations 90%+ for well-established use cases Trust sentiment Team confidence in AI Increasing over time Conclusion: Trust Is the Real AI Bottleneck The technical challenges of AI incident management are solvable. The harder challenge is organizational: building the trust that makes AI useful. Organizations that invest in explainable AI, data quality, and team training will see the trust gap shrink. Organizations that treat AI as a magic solution without investing in trust will continue to pay the trust tax. In incident management, AI's biggest bottleneck isn't technology. It's trust. Action Items for Your Organization Implement confidence scoring: Every AI recommendation should include a confidence score Build AI transparency: Show teams how AI reaches its conclusions Measure verification time: Understand the trust tax in your organization Start with low-stakes incidents: Build trust before expanding to critical systems Train teams on AI: Help them understand AI capabilities and limitations Track override rates: Understand when and why teams override AI recommendations  

The Cost of Late Detection Is Greater Than the Cost of Recovery — Why MTTD Is Your New Most Important Metric The Shift in Operational Priorities For decades, Mean Time to Resolution (MTTR) has been the gold standard metric for incident management. How quickly could you fix what broke? The faster, the better. But in 2026, a different metric is taking center stage: Mean Time to Detect (MTTD). Why Detection Matters More Than Recovery The logic is simple: By the time an incident reaches a support queue, damage has already occurred. Customers may already be impacted, revenue may already be lost, and operational resilience thresholds may already be breached. The organizations that outperform operationally are not necessarily those with the fastest recovery teams. They are the ones that: Detect anomalies earlier Understand service dependencies faster Identify blast radius immediately Correlate operational signals intelligently Escalate risk before users report issues The cost of late detection is often greater than the cost of recovery. The MTTD Paradox Focusing on MTTR creates a counterproductive dynamic: Focus Outcome Focus on faster recovery Fix symptoms, not causes Focus on faster detection Fix causes, prevent recurrence Focus on MTTR Celebrate heroes Focus on MTTD Prevent incidents When you focus on MTTR, you incentivize quick fixes that may not address root causes. The result is recurring incidents. When you focus on MTTD, you incentivize early detection of issues. The result is prevention and resilience. The Blast Radius Problem The importance of detection grows as the "blast radius" of incidents grows. Incident Type Blast Radius Importance of Detection Single-user issue Low Low Team-level issue Medium Medium System-level issue High High Business-critical issue Very high Very high AI-caused incident Potentially massive Critical When an AI agent can impact thousands of users in seconds, early detection is not just important—it's essential. How to Measure and Improve MTTD What MTTD Measures MTTD measures the time between: Incident occurrence (when the issue first appears) Incident detection (when your team becomes aware) Improving MTTD Strategy Impact Implement observability See issues that are invisible to monitoring Use AI-powered anomaly detection Replace static thresholds with dynamic models Implement real-time monitoring Detect issues in real-time, not in post-mortems Correlate alerts Reduce noise and surface real issues Set up automated detection Detect issues before humans would notice Build service maps Understand which services are affected The Interplay Between MTTD and MTTR MTTD and MTTR are not in competition. They work together: Scenario MTTD MTTR Outcome Late detection, slow recovery High High Extended impact Late detection, fast recovery High Low Still impacts users Early detection, slow recovery Low High Limited user impact Early detection, fast recovery Low Low Minimal user impact The best outcome is low MTTD (early detection) AND low MTTR (fast recovery). But if you have to choose, low MTTD (early detection) is more important—it prevents widespread impact. Real-World Impact Manufacturing Client A global manufacturing client rebuilt their observability stack with AI-powered capabilities. Results: Alert noise reduced by 80% MTTR reduced by 50% 15% decrease in support tickets 10% increase in successful order completions Financial Services Client A major bank implemented AI-powered detection and remediation: MTTR reduced from 30 hours to 1 hour (96% faster) Major incidents reduced by 70% Conclusion: Detection Is the New Competitive Advantage The organizations that lead in incident management will be those that detect issues before customers do. They'll invest in observability, AI-powered anomaly detection, and real-time monitoring. In the 2026 incident management landscape, early detection is not just a metric. It's a competitive advantage. Action Items for Your Organization Measure MTTD: Understand your current detection speed Invest in observability: Gain visibility into your entire stack Implement AI-powered anomaly detection: Replace static thresholds with dynamic models Correlate alerts: Reduce noise and surface real issues Set up automated detection: Detect issues before humans would notice Build service maps: Understand which services are affected  

Headline: Generate Stakeholder Analyses, Communication Timelines, and Target Messages with AI The Communication Challenge Change managers need to communicate effectively with diverse stakeholders across the organization. But crafting the right message, for the right audience, at the right time, through the right channel is complex and time-consuming. AI is changing this. AI-Generated Communication Plans Change managers can use AI to craft entire communication plans. Sergi suggests a prompt like this: "I want you to act as a change management lead. Use this data or project brief to create a simple stakeholder analysis table. Then, I'd like you to give me the right groups I should communicate this change to, the right impact levels, the right messages, and a three-phase communication plan" . What AI can generate: Stakeholder analysis tables Audience segmentation Impact levels for different groups Key messages by audience Three-phase communication plans (pre-launch, launch, post-launch) Two to three key activities or messages for each phase Targeted Communications AI can also help by identifying how best to communicate with a specific team. Carlos Martinez, formerly at Salesforce, used AI to analyze survey data to understand the biggest pain points for a specific function and craft a narrative that resonates with their experiences . How it works: Collect survey data or feedback Use AI to analyze pain points by function Identify what's working and what's not Craft messages that address specific concerns Channel and Timing Optimization After crafting the message, AI can help determine: What channels to use: Based on your company's usage patterns When to communicate: Based on engagement metrics What format to use: Text, video, interactive, or a combination  The Human Oversight Imperative Karunakaran cautions that humans need to oversee the communication plan to make sure everything stays human-centered. Employees get flooded with messages and emails—which they'll ignore if there are too many . AI generates content; humans provide judgment. Practical Implementation Step 1: Upload Your Data Project briefs Stakeholder information Historical communication data Step 2: Generate the Plan Use AI to generate draft plans Review and refine Step 3: Execute and Optimize Implement the plan Use AI to track engagement Adjust based on data Conclusion AI-powered communication plans enable change managers to create targeted, timely, and effective communications at scale. By automating the drafting process, AI frees change managers to focus on strategy and human connection. Action Items for Your Organization Create AI prompt templates for communication planning Use AI to generate stakeholder analyses Segment audiences and tailor messages Optimize channel and timing based on data Maintain human oversight for authenticity and empathy

You Can't Improve What You Don't Track — Why Service Request Documentation Matters Why Documentation Matters Documenting all service requests — current and closed — prevents work from falling through the cracks and enables continuous improvement . Key information to document includes: Type of request Completion timeline Assignee Requester Action taken SLAs  What to Document Field Why It Matters Request type Identifies patterns and trends Requestor Enables follow-up and satisfaction tracking Assignee Shows who handled the request Action taken Documents what was done Timeline Tracks completion time and SLA compliance Status Shows current and historical states Using Documentation for Improvement Documentation is valuable when improving processes, as teams can easily see data like how long requests typically take to complete or how many stakeholders are involved . Key analyses: Cycle time: How long does each request type take? Volume trends: Which request types are increasing? SLA compliance: Which requests miss targets? Satisfaction patterns: Which request types have low satisfaction? Documentation Tools Tool Type Benefits ITSM platform Structured data, built-in reporting Templates Consistent documentation Automation Automatic capture of timelines and actions Conclusion Documentation isn't just about record-keeping — it's about improvement. Organizations that document well can analyze, optimize, and continuously improve their service request management. Action Items for Your Organization Review what data you currently capture Identify gaps in documentation Standardize documentation fields Use documentation for reporting and analysis Create dashboards for key metrics  

Headline: A Control with No Owner Is Just a Good Intention The Ownership Imperative A control with no owner is just a good intention. Without clear accountability, controls: Fall out of date Stop operating effectively Are not tested or monitored Fail during audits Create gaps in risk coverage Effective controls management requires clear ownership and accountability. Defining Control Ownership Role Responsibility Control Owner Accountable for the control's design, operation, and effectiveness Control Operator Executes the control activity Control Tester Tests control effectiveness Control Approver Approves control changes or exceptions Control Ownership in Practice Control Owner Responsibilities: Ensure the control is designed effectively Monitor control operation Address control failures Coordinate testing Maintain control documentation Review and update controls regularly Approve exceptions Control Operator Responsibilities: Execute the control activity Document evidence of execution Report issues to the control owner Follow defined procedures Control Tester Responsibilities: Test control effectiveness Document test results Report findings Recommend improvements The Business Context Layer By integrating identity context into the CMDB, organizations can link identity and risk signals to controls and services : Add a risk-aware business lens: Business services can be enriched with identity-derived risk signals such as user sensitivity, segregation-of-duties violations, and access sprawl  Drive smarter ITSM decisions: Incident prioritization, change approvals, and request fulfillment can factor in identity risk  Improve visibility for governance: Linking identity to CIs and services creates a complete picture for access reviews, policy enforcement, and exception handling  The Problem with Fragmented Ownership The consequence of fragmented ownership: Controls are duplicated across teams Controls are inconsistent Accountability is unclear Gaps emerge between what documentation says and what actually happens The solution: Clear accountability structures Documented ownership in risk and control matrices Regular reviews of ownership assignments The Regulatory Driver Regulators increasingly expect clear accountability for controls. Frameworks like COSO and COBIT emphasize the importance of control activities and their ownership. Boards and regulators are demanding to know: Who is responsible for each control? How do we know controls are operating effectively? What happens when controls fail? Effective controls management requires clear ownership and accountability. Conclusion Control ownership and accountability are the foundation of effective controls management. Organizations that assign clear ownership, define responsibilities, and maintain accountability structures will have controls that operate effectively and withstand audit scrutiny. Action Items for Your Organization Assign clear ownership for all controls Document owner responsibilities Identify control operators and testers Review ownership assignments regularly Ensure owners have the resources and authority they need