Category: Feature

Featured on the front page

Use AI to fix Failure Demand

Failure demand is, in essence, the additional (and unnecessary) workload created when an organisation fails to provide a product or service accurately or completely at the first point of contact. In large citizen-facing organisations—government agencies, healthcare systems, or large federated enterprises—failure demand often arises from structural and procedural issues that, if left unmanaged, create spirals of repeated contacts, rework, complaints, and escalations.

Below are common causes of failure demand in large federated organisations, along with ways in which AI can help alleviate or prevent these issues.

1. Fragmented Information and Siloed Systems

Cause:

• Multiple disconnected databases or information systems mean that staff can’t easily access the correct, up-to-date information about a citizen or case.

• Different departments or agencies have their own processes, making it difficult to get a single, integrated view.

How AI Helps:

1. Data Integration & Master Data Management

• AI-driven data integration or entity resolution can match and merge records across siloed systems, providing a single source of truth.

2. Knowledge Graphs

• These can unify information from various internal and external systems, surfacing the relevant data to the front line or self-service portals in real time.

2. Repeated or Escalated Inquiries

Cause:

• Citizens have to call multiple times or contact different departments because they never receive the correct answer or a complete resolution on the first attempt.

• Instructions or next steps are unclear, requiring additional clarifications.

How AI Helps:

1. Natural Language Processing (NLP) for Triage

• AI-based chatbots and virtual assistants can quickly assess the request and route it to the correct team, reducing misrouted calls.

2. Automated Knowledge Bases

• AI can suggest the next best action or provide consistent answers to common questions, reducing inaccurate or incomplete information.

3. Lack of Process Visibility (for Both Staff and Citizens)

Cause:

• Citizens have little visibility into the status of their application, request, or case.

• Staff themselves may struggle to track cases as they move through different departments, leading to delays and confusion.

How AI Helps:

1. Predictive Tracking and Alerts

• AI can monitor case progress and send automatic notifications to both citizens and staff about status changes, required documents, or impending deadlines.

2. Process Mining and Workflow Optimisation

• AI-driven process mining tools analyse workflow logs to identify bottlenecks or high-friction steps, prompting proactive solutions.

4. Overly Complex or Confusing Service Design

Cause:

• Citizens are forced to navigate confusing online portals, physical forms, and long instructions, which leads to errors or incomplete submissions.

• Lack of standardisation across departments can create additional steps and inconsistencies.

How AI Helps:

1. Personalised Digital Assistants

• Virtual agents that guide citizens step-by-step, ensuring forms and data are filled correctly and explaining next steps in simple language.

2. Adaptive User Interfaces

• AI can tailor the user experience based on the user’s profile, automatically simplifying the path or adjusting the language for clarity.

5. Inconsistent Communication or Messaging

Cause:

• Different channels (phone, email, web chat, social media) give conflicting information or instructions.

• Citizens receive either no response or delayed responses, leading to additional follow-ups.

How AI Helps:

1. Omni-channel Response Orchestration

• AI models can be trained on policy guidelines and knowledge bases to ensure consistent, channel-agnostic responses.

2. Sentiment Analysis and Real-time Alerts

• Monitoring digital communications can quickly highlight negative or confused user sentiments, prompting staff to intervene before citizens need to escalate.

6. Manual, Repetitive Tasks Leading to Errors

Cause:

• Staff spend time on repetitive data entry and manual verification processes, which are prone to human error.

• A single mistake can lead to multiple follow-up calls and corrective work.

How AI Helps:

1. Optical Character Recognition (OCR) and Automated Data Entry

• AI tools can accurately parse large volumes of forms, extracting data and populating systems automatically.

2. Robotic Process Automation (RPA)

• Combining RPA with AI (“Intelligent Automation”) can handle repetitive workflows, flags issues automatically, and hand off only exceptions to human staff.

7. Limited Staff Training or High Staff Turnover

Cause:

• In large federated organisations, staff turnover can be high, or training may be inconsistent.

• Knowledge retention is poor, meaning new or rotating staff do not always have the expertise to handle calls correctly.

How AI Helps:

1. Real-time Call Guidance

• AI-driven recommendations can guide agents during phone or chat interactions, suggesting answers based on historical successful interactions.

2. Machine Learning for Training Gaps

• Analysis of interactions can highlight patterns of agent errors or knowledge gaps, guiding targeted staff training efforts.

8. Reactive Instead of Proactive Approach

Cause:

• Processes are often designed to react to incoming inquiries rather than preventing confusion or mistakes in the first place.

• Citizens only discover requirements (e.g., missing documents, extra steps) after they have already submitted something incorrectly.

How AI Helps:

1. Predictive Analytics

• By analysing historical data, AI can forecast which cases might lead to repeated follow-ups or escalate, prompting proactive outreach.

2. Proactive Communication

• Automated notifications (e.g., reminders, deadline notices) reduce the likelihood of citizens missing requirements and calling back to ask for clarifications.

9. Inability to Identify Root Causes

Cause:

• Without an organised way to analyse large volumes of calls, emails, and visits, it is difficult to understand why so many follow-ups or escalations happen.

• Root-cause analysis often requires manual effort, which is time-consuming and prone to oversight.

How AI Helps:

1. Text and Speech Analytics

• AI can analyse phone transcripts, chat logs, and emails to uncover themes, common queries, or shared blockers driving repeat contacts.

2. Topic Clustering

• AI clustering techniques group citizen complaints or issues, helping leadership see broader trends and attack the underlying causes.

10. Poor Feedback Loops Between Front-Line and Policy/Process Owners

Cause:

• Front-line staff and citizens encounter the same problems repeatedly, but those issues are not effectively communicated upstream to the departments that design the processes.

• This results in short-term fixes (workarounds) rather than systemic changes (resolutions to root causes).

How AI Helps:

1. Closed-Loop Feedback Systems

• AI-driven dashboards can aggregate real-time data on contact types, resolutions, and user satisfaction, automatically flagging major process issues.

2. Continuous Improvement Recommendations

• Machine Learning (ML) algorithms can recommend policy or process changes based on patterns and outcomes, pushing insights directly to policy owners.

Key Takeaways

1. Integration and Data Sharing

• Breaking down organisational silos is essential to reducing failure demand. AI can help by unifying and analysing disparate data.

2. Personalisation and Proactivity

• AI can provide personalised guidance and proactively alert citizens (and staff) to potential issues, cutting down on repeated contacts.

3. Automation of Low-Level Tasks

• Robotic Process Automation (RPA) and intelligent document processing reduce human error and free staff for more complex, value-adding activities.

4. Insight Generation

• Text analytics, speech analytics, and clustering methods can reveal hidden causes of frequent failures and drive continuous improvement.

By applying AI methods to target these root causes—fragmented data, repeated inquiries, manual errors, and slow feedback loops—large citizen-facing and federated organisations can decrease failure demand, improve citizen experiences, and allow staff to focus on more valuable, mission-critical tasks.

Feature Problem examples

Slow Software Deployment

Background: A tech company experiences slow software deployment, causing frequent delays in launching updates. This has led to customer dissatisfaction and a decline in product reliability perception.

Workaround:

The development team decides to increase manual testing and patching before each release to catch and fix issues quickly. This helps minimise the delays and ensures the software works as expected, but it’s not a perfect solution. It still consumes a lot of time and resources, adding to costs.

Symptom: Slow software deployment and frequent delays.
Workaround Applied: Manually patching and increasing testing time to catch last-minute issues.

Deeper Analysis:

Upon investigation, it is found that the cause of frequent delays is frequent bugs and integration issues appearing late in the development cycle. The manual patching helps to catch some of these issues, but it doesn’t address why they happen in the first place.

Cause: Frequent bugs and integration issues late in the development process.

Root Cause:

Looking further, the root cause was discovered to be a lack of proper code review and integration testing throughout the development process. Developers worked in silos, leading to a buildup of conflicts that were only noticed during final integration.

Root Cause: Lack of continuous integration and code reviews during development.

Solution:

The company decides to implement a Continuous Integration/Continuous Deployment (CI/CD) pipeline with automated testing and regular code reviews. This allows bugs to be detected earlier and fixed immediately, avoiding the last-minute rush to patch things up. Additionally, it encourages collaboration among developers, ensuring that code conflicts are resolved quickly and cleanly.

Solution: Implementing a CI/CD pipeline with automated testing and regular code reviews.

Outcome: With the new solution in place, the team can deploy software more reliably and quickly. The need for manual patches is reduced, and customers are happier with the timely, high-quality updates.

Summary:

Workaround: Manual patching and extended testing time.
Symptom Addressed: Slow software deployment.
Cause: Bugs and integration issues detected late.
Root Cause: Lack of continuous integration and code reviews.
Solution: Implementing a CI/CD pipeline with automated testing.

This scenario illustrates how a workaround can temporarily relieve symptoms but doesn’t solve the underlying issue. Giving developers greater access to testing tools and avoiding manual steps is a winner.

Feature Problem examples

The case of the Late Homework

Background:
A school teacher notices that many students consistently turn in their work late, the late homework disrupts the class schedule and affects the students’ marks.

Workaround:

To address this, the teacher starts offering extra marks to students who turn in their homework on time. This motivates some students to be more punctual, but it doesn’t solve the underlying problem, and many students still struggle to meet the deadlines.

Symptom: Students turning in homework late.
Workaround Applied: Offering extra marks for on-time submissions.

Deeper Analysis:

After talking to students, the teacher discovers that the cause is that students often forget about the homework or feel overwhelmed by the amount of work they have. Many students do their homework at the last minute or not at all because they can’t organise their time effectively.

Cause: Students forget about homework or feel overwhelmed, leading to procrastination.

Root Cause:

Going deeper, the root cause is found to be that students lack time management skills and do not use planners effectively. They aren’t taught how to break down assignments into smaller, manageable tasks, and as a result, they get overwhelmed and miss deadlines.

Root Cause: Poor time management skills and lack of guidance on using planners to organise assignments.

Solution:

The teacher implements a two-part solution. First, they teach a mini-lesson on time management skills, showing students how to use a planner and break down big assignments into smaller steps. Second, they introduce a classroom routine where students spend the last few minutes of each class updating their planners and setting goals for when they will complete their homework.

Solution: Teaching time management skills and incorporating planner use into the daily routine.

Outcome: With these changes, students learn to organise their time better and keep track of their assignments, leading to a significant improvement in on-time homework submissions. The extra marks incentive becomes less necessary, as students have developed the skills to manage their workload more effectively.

Summary:

Workaround: Offering extra marks for on-time homework submissions.
Symptom Addressed: Students consistently turning in homework late.
Cause: Students forget about homework or feel overwhelmed by their workload.
Root Cause: Poor time management skills and lack of guidance on using planners.
Solution: Teaching time management skills and incorporating planner use into the daily routine.

This example shows how applying a workaround can temporarily address a problem, but teaching students better habits and skills (the root cause) leads to a long-term solution.

Feature Problem examples

The case of the Delayed Programme milestones

Background: A large enterprise is undergoing a multi-year IT transformation program, which includes several projects aimed at modernising its core systems, migrating data to the cloud, and improving cybersecurity. However, the program is experiencing repeated delays, with milestones being missed across multiple projects. This leads to budget overruns, delayed programme milestones, frustrated stakeholders, and concerns about the feasibility of completing the program on time.

Workaround:

To deal with the delays, the program management office (PMO) starts reallocating resources between projects on an ad-hoc basis. Whenever a project falls behind, resources are borrowed from other projects to catch up. This helps address immediate issues, but it causes disruptions across other projects, leading to further delays and inefficiencies.

Symptom: Repeated delays and missed milestones across multiple projects in the program.
Workaround Applied: Reallocating resources between projects to address delays.

Deeper Analysis:

A deeper investigation reveals that the cause of the delays is poor coordination between projects and unclear dependencies. Many projects are interdependent, but they operate in silos, with little communication or alignment on shared milestones. For example, a data migration project might need to wait for a core system upgrade, but delays in one project cascade into others, creating a domino effect of missed deadlines.

Cause: Lack of coordination and communication between projects, leading to delays in shared milestones.

Root Cause:

The root cause of the issue is identified as the absence of a robust program governance framework that can effectively oversee and align multiple projects. The program lacks a centralised system for tracking dependencies and managing risks across projects. Additionally, there are no clear escalation procedures for when issues arise, leading to delays being addressed too late.

Root Cause: Lack of a centralised governance framework, ineffective tracking of dependencies, and unclear escalation procedures.

Solution:

The program management office (PMO) decides to implement a centralised program management framework to improve oversight and coordination. This includes the use of program management software that allows for comprehensive tracking of project dependencies, timelines, and risks across the program. The PMO also introduces program-level governance meetings, where project managers can report on progress, identify potential delays, and align on shared milestones. Clear escalation procedures are established, so issues can be addressed quickly and efficiently.

Solution: Centralised program management framework, comprehensive tracking software, regular governance meetings, and clear escalation procedures.

Outcome: With the new framework in place, projects within the program are better coordinated, and shared milestones are clearly defined and tracked. Dependencies between projects are managed proactively, reducing the risk of delays cascading across the program. Regular governance meetings help ensure that all projects are aligned, and issues are escalated and resolved in a timely manner. This leads to improved program performance, reduced budget overruns, and greater confidence among stakeholders.

Summary:

Workaround: Reallocating resources between projects to address delays.
Symptom Addressed: Repeated delays and missed milestones across multiple projects.
Cause: Lack of coordination and communication between projects, leading to delays in shared milestones.
Root Cause: Lack of a centralised governance framework, ineffective tracking of dependencies, and unclear escalation procedures.
Solution: Implementing a centralised program management framework, using tracking software, holding regular governance meetings, and establishing clear escalation procedures.

This example illustrates how addressing the root cause of poor program governance can lead to better coordination across projects, more efficient resource use, and successful program delivery, rather than relying on short-term fixes like ad-hoc resource reallocation.

Feature Problem examples Symptom

The case of the Coffee Shop Customer Complaints

Background: A popular coffee shop has been receiving frequent customer complaints about long wait times during peak hours. Many customers express frustration, and some even leave without ordering, affecting sales.

Workaround:

The shop manager decides to offer free coffee vouchers to customers who have to wait too long. This helps reduce complaints because customers feel compensated for the inconvenience, but it doesn’t solve the core issue of long wait times.

Symptom: Long wait times for customers during peak hours.
Workaround Applied: Offering free coffee vouchers to appease customers who wait too long.

Deeper Analysis:

On further investigation, the cause of the long wait times is found to be bottlenecks at the order-taking counter. The baristas are quick at making drinks, but there is only one register, so customers have to queue up to place their orders.

Cause: Bottlenecks at the order counter, leading to slow order processing.

Root Cause:

Digging even deeper, the root cause is discovered to be the lack of staff training on how to efficiently take orders and use the register. Additionally, the shop’s layout has the counter placed in a way that creates congestion, making it difficult for staff to move around freely during busy periods.

Root Cause: Inefficient staff training and poorly planned shop layout.

Solution:

The coffee shop implements two key changes. First, they retrain staff to handle orders more quickly, using clear scripts to minimise confusion. Second, they redesign the shop layout to add a second register and create a more streamlined space, so the staff can move efficiently even during rush hours.

Solution: Improved staff training and a redesigned layout with an additional register.

Outcome: With the new measures, customers are served faster, and wait times are reduced significantly. The shop no longer needs to rely on free vouchers to keep customers happy, as the main problem of long waits has been effectively addressed.

Summary:

Workaround: Offering free coffee vouchers to compensate for long wait times.
Symptom Addressed: Customer dissatisfaction due to long wait times.
Cause: Bottlenecks at the order counter.
Root Cause: Inefficient staff training and poor shop layout.
Solution: Retrain staff and redesign shop layout to improve workflow.

This example demonstrates how addressing the root cause, rather than just applying a quick fix, can lead to a more sustainable and effective solution.