WFM guideRouting & skills

Contact centre skills management

Q: How many skills should a contact centre agent have?

The optimal number of skills for a contact centre agent depends on operational design. Specialists with 1–2 skills deliver the highest quality and efficiency on those contact types but create staffing risk (if they are absent, their queue has reduced capacity). Generalists with 5+ skills provide excellent staffing flexibility but typically have lower proficiency on each individual skill. Most well-managed contact centres use a tiered model: agents have a primary skill (the contact type they handle 70–80% of the time, where they have their deepest proficiency) and 1–2 secondary skills (where they can handle contacts competently during peaks or when their primary queue is in surplus). Beyond two secondary skills, proficiency typically declines and routing complexity increases. The exception is small operations — with fewer than 20 agents, breadth is operationally necessary and the proficiency trade-off is acceptable.

Skills-based routing is only as effective as the skills database is accurate. An agent assigned a skill they are not proficient in generates poor-quality contacts. An agent whose skill has decayed but whose record still shows it active gets routed contacts they struggle with. Skills management is the ongoing governance that keeps the database aligned with operational reality.

Skill proficiency levels: what they mean and how to use them

Most ACD systems support skill proficiency levels (typically 1–10 or 1–5). A lower-proficiency agent can handle the contact type, but the routing engine will prefer a higher-proficiency agent when one is available. Proficiency levels allow the operation to handle overflow with a capable but less experienced agent rather than either routing to a specialist or putting the customer in queue.

Level	Agent status	When they receive this skill's contacts	Typical QA/oversight requirement
1 (Trainee)	Newly trained, not yet signed off for independent handling	Training and nesting calls only — with supervisor monitoring. Not in live production routing.	100% call monitoring during nesting period
2 (Developing)	Completed initial training; handling live contacts with higher support availability	Last resort — only when all higher-proficiency agents are unavailable and the queue is near threshold. Consider a queue threshold delay before routing to Level 2.	30–40% monitoring; coaching focus on the skill type
3 (Competent)	Handles the skill type independently and consistently to standard	Active routing — will receive contacts for this skill type during normal operations. Will not be selected over higher-proficiency agents if they are available.	Standard QA monitoring rate (10–15%)
4 (Proficient)	Above-standard performance on this skill type; preferred handling agent	Primary routing destination for this skill type. First choice when available.	Standard QA monitoring rate; eligible for coaching peers
5 (Expert)	Highest proficiency; handles complex or escalated contacts in this skill type	Preferred for high-complexity or escalated contacts in this skill type. May be reserved for escalations only.	Reduced QA; used as calibration benchmark for the skill type

Specialist vs. generalist: the operational trade-off

Specialist model (1–2 skills per agent)

+Highest quality per contact type — deep expertise

+Faster AHT — specialists know the contact type well

+Simpler routing logic — fewer routing rules to maintain

−Staffing risk — absence in a specialist team reduces queue capacity proportionally more than in a generalist team

−Reduced intraday flexibility — cannot redeploy agents from a low-volume queue to a high-volume one

−Higher headcount requirement to achieve the same SL as a blended model

Generalist model (3+ skills per agent)

+Maximum intraday flexibility — agents can be moved to the highest-priority queue

+Lower headcount to achieve the same SL as a specialist model — the efficiency of pooling

+Resilience to absence — impact of individual absence is spread across all queues

−Lower proficiency per contact type — agents less deeply specialised

−Higher AHT on secondary skills — less familiar with the contact type

−Complex routing logic — more routing rules, more dependencies to manage

Recommended hybrid: Most contact centres with 50+ agents benefit from a tiered model: each agent has one primary skill (handled 70–80% of the time; highest proficiency) and 1–2 secondary skills (activated during peaks or when the primary queue is in surplus). This delivers most of the pooling efficiency benefit while maintaining specialist depth on primary skills and limiting routing complexity.

Skill decay: when skills become stale

An agent who does not handle a particular contact type for an extended period loses proficiency. Script and product knowledge becomes outdated. Process steps are forgotten. When they are routed a contact of that type, their AHT is higher, their FCR is lower, and their CSAT on those contacts is worse. Skill decay is a predictable operational risk — but only if the skills database is monitored.

Contact type	Typical skill decay timeline without refresh	Refresh mechanism
Simple transactional (address update, account balance)	6–12 months — low process complexity, slow decay	Brief refresher (30min) on any process changes; re-exposure to live contacts under monitoring
Product knowledge (features, pricing, eligibility)	3–6 months — product changes frequently accelerate decay	Product update communication and acknowledgement; knowledge base review; brief supervised handling
Complex troubleshooting (technical support)	2–4 months — high process complexity; specific diagnostic steps are forgotten	Half-day refresher training; supervised handling of 10–15 contacts before independent activation
Complaint handling	4–8 months — script recall and de-escalation techniques fade without regular use	QA review of current complaint contacts; role-play session; TL sign-off before reactivation
Regulatory / compliance-critical (FCA, Ofcom, consumer credit)	1–3 months — regulatory rules change frequently and require precise compliance	Mandatory refresher before every reactivation regardless of last handling date; L&D sign-off

Skills governance: keeping the database accurate

Review the skills database monthly

Run a report of all active skill assignments. Flag agents who have not handled a contact of a given skill type in more than 60 days. Review whether the skill should be suspended pending a refresher, or whether there is a routing reason they have not received contacts.

Suspend skills before they fully decay, not after a quality incident

Skills should be reviewed for suspension when an agent transitions to a different primary queue, goes on long-term absence, or has not handled the skill type within the decay window. Waiting for a quality incident to trigger a skill review is reactive — the customer has already had a poor contact.

Log all skill changes with reason and date

Every skill addition, level change, and removal should be recorded in a skills change log. The log creates an audit trail for compliance contacts and allows investigation of quality issues that coincide with skill changes.

Align skill assignments with QA performance data

If an agent's QA score on a particular skill type is consistently below the centre standard, their proficiency level on that skill should be reviewed downward. High-proficiency assignment should be based on demonstrated performance, not tenure or training completion alone.

Plan skill additions as part of the capacity plan, not reactively

Cross-training should be planned based on the WFM team's skill gap analysis — which queues are at risk of insufficient coverage if agents are absent. Reactive cross-training (in response to a coverage crisis) produces inadequately trained agents handling contacts they are not ready for.

Skills management questions

How many skills should a contact centre agent have?

The optimal number depends on operational design. Specialists (1–2 skills) deliver the highest quality and efficiency on those contact types but create staffing risk when absent. Generalists (5+ skills) provide staffing flexibility but typically have lower proficiency on each skill individually. Most well-managed contact centres use a tiered model: agents have one primary skill (70–80% of their contact volume; highest proficiency) and 1–2 secondary skills for overflow and cross-cover. Beyond two secondary skills, proficiency typically declines and routing complexity increases without proportional benefit. The exception is small operations (fewer than 20 agents) where breadth is operationally necessary and the proficiency trade-off is acceptable.