The Tool-First Improvement Plan at Northwind Support
Northwind Support employs 42 agents handling 18,000 tickets a month for an insurance client. First-contact resolution (FCR) is 58% against a target of 75%, average handle time is 11 minutes, and a spreadsheet holds 300 improvement ideas that have never been triaged. The service manager plans to spend six months selecting and implementing a new ITSM tool with AI-assisted knowledge search, arguing that the tool will 'fix FCR'. Separately, two agents prototyped a simple knowledge article template in the existing tool and cut repeat contacts on one product line by 12% over three weeks. Approval for the tool requires a business case, and the CIO wants evidence of value within the current quarter. Staff are sceptical after a failed CRM rollout two years ago. The service manager has asked you, as continual improvement lead, to endorse the tool-first plan in next week's steering meeting.
1. Using the ITIL 4 guiding principles, evaluate the service manager's proposal and recommend a sequence of actions for the next quarter.
Read the worked answer
The proposal treats a tool as the improvement rather than as a means to one, which conflicts with focus on value: the value at stake is first-contact resolution and the customer effort it represents, not the acquisition of AI search. It also conflicts with start where you are - a working prototype already exists and has produced a 12% reduction in repeat contacts on one product line - and with progress iteratively with feedback, since a six-month selection cycle delays learning and repeats the pattern of the failed CRM rollout. Keep it simple and practical favours scaling the template before buying a platform, and optimize and automate says automate only after the knowledge workflow itself is improved. Collaborate and promote visibility is the missing ingredient with sceptical agents.
A defensible next-quarter sequence: (1) weeks 1–2, baseline FCR and repeat contacts by product line and triage the 300 ideas against that baseline; (2) weeks 2–8, scale the article template to two or three product lines with weekly measurement and agent involvement; (3) in parallel, derive tool requirements from what the prototype cannot do, and run a limited proof of concept; (4) at quarter end, decide on the tool using measured evidence. Endorse the plan conditionally, not as tool-first.
2. The CIO asks how the team should decide whether the ITSM tool is a worthwhile improvement. Describe the reasoning you would use.
Read the worked answer
Treat the purchase as a hypothesis to be tested, not a decision to be justified. State the predicted effect: 'AI-assisted search will raise FCR from 58% to at least 68% within two quarters without increasing average handle time or decreasing CSAT.' Then choose measures: lagging outcomes (FCR, repeat contact rate, CSAT, cost per contact), leading indicators (article creation and reuse, search success rate, agent adoption), and guardrails (AHT, backlog, agent workload). Establish the baseline before the pilot and use a comparison group - product lines without the tool - so seasonal and volume effects, which are material at 18,000 tickets a month, do not masquerade as benefit. Cost the whole thing: licences, implementation, integration, content migration, training, administration, and the opportunity cost of not scaling the template. Pre-agree decision thresholds and a kill criterion. Review at 8 and 12 weeks, and require the vendor to support a measurable pilot rather than a demo. This gives the CIO quarterly evidence and protects the team from a repeat of the CRM experience.
3. What should happen to the 300-item idea backlog before any tool decision is taken?
Read the worked answer
Before any tool decision, the 300 ideas need to become a usable improvement backlog. Triage them in a short, facilitated session with service owners and agents: remove duplicates, link each item to a service or product line, and record the symptom it addresses, an expected outcome and a rough value/effort score. Expect most to be fixes rather than improvements; those can go through normal incident and problem channels, leaving a smaller set of genuine improvement opportunities. Identify quick wins that need no tooling - the knowledge template is one. Assign an owner to every item that survives and set a review date. This is the 'where are we now' and 'where do we want to be' work of the continual improvement model, and it produces the business case evidence the tool purchase currently lacks. It also makes the register credible to staff, which matters after the CRM failure.