FrontLine for Workforce — multi-client scheduling that scales with you, not against you.
One workforce. Multiple clients. Multiple LOBs. The scheduling layer was the part you were hand-rolling in spreadsheets — because every generic WFM tool assumed one employer, one funnel. FrontLine treats multi-client as the data model, not a configuration toggle.
Workforce dashboard
Acme BPO · 247 agents · 4 clients · 94% coverage
Active agents
247
Coverage rate
94%
Adherence
93%
OT cost (today)
$2.4K
Live agent mix
Right now
RetailCo EN
68
RetailCo FR
22
FinCo
54
TelcoPro
38
Training
12
Break
18
Recent WFM activity
Coverage gap detected · RetailCo FR · 14:00–15:00 · 2 agents short
5m ago
OT offer accepted · Maya P. · 18:00–22:00 · FinCo
12m ago
Schedule republished · TelcoPro · 3 swap-driven changes
1h ago
Compliance gate blocked publish · ESA daily-max for 2 agents · resolved
Yesterday
LOB coverage · 14:00 ET
Live
Card Services
47/50
-3Customer Care
82/82
✓Bilingual FR
12/16
-4Tech Support
34/30
+4What's hard about WFM at a multi-client BPO
Generic WFM tools (NICE WFM, Verint, Calabrio) were built for a single employer running a single workforce against a single demand curve. Multi-client BPO breaks every one of those assumptions.
Multi-client volume forecasting that shares no pattern
RetailCo peaks Fridays. FinCo peaks Mondays. TelcoPro peaks Tuesday afternoons. Forecasting one workforce against four uncorrelated curves isn't harder than single-client — it's a different problem. Generic WFM tools choke when the curves combine.
Cross-LOB agent assignment is multidimensional
Mike works RetailCo EN 8–12, FinCo FR 12–4 — same agent. Generic WFM forces you to model him as two FTEs (over-counting) or one single-skilled FTE (under-utilizing). Either way: inflated headcount the CFO can't justify, or an SLA miss.
Real-time adherence needs multi-state context
Single-state adherence tells you "on call" but not "on FinCo's call." Your client-facing SLA reports become guesses dressed up as data — and renewals get hard when the numbers don't match the client's experience.
New client launches break your existing forecast
A 50-agent client launch breaks the existing forecast model. Most WFM systems need a fresh model and lose 30 days of history. The team forecasts blind through launch — and SLA misses concentrate in week 2.
Schedule publishing is a multi-stakeholder negotiation
WFM wants coverage, supervisor wants Mike off Friday, Mike wants Sunday morning, finance wants no OT. Negotiating all four takes 3 days; you have 6 hours. The schedule everyone signs off on still has compliance violations nobody caught.
Skills-based routing is a 2-D matrix, not a flat tag list
Mike is certified on FinCo banking but not RetailCo refunds. Generic WFM treats skills as a flat tag list; you need a 2-D matrix — skills × client. Miss either dimension and the wrong-skilled assignment breaks an SLA.
What FrontLine does for your WFM team
Seven WFM-specific surfaces designed for multi-client operations from the ground up. Schedule, forecast, adherence, self-service — every shift carries its client and line of business end-to-end.
Feature 01
Multi-client schedules where every shift knows its client and LOB
One agent can work multiple clients on the same day, on a single roster. Each shift carries its client, line of business, and activity, so per-client coverage, adherence, and SLA reports roll up exactly.
- Scoped shifts. Each shift carries its client, line of business, and activity. Routing eligibility is enforced when you build the schedule — agents only appear on shifts for clients they're certified for.
- One agent, every client. The same employee can work RetailCo, FinCo, and TelcoPro shifts on the same day, on one roster. No duplicate agent records and no parallel schedules to reconcile.
- Per-client coverage. Coverage per line of business is compared against the live per-client forecast. Under- and over-staffed intervals surface before publish, with exact rollups.
- Compliance at publish. Schedules that violate LNT rest periods, ESA daily-max hours, weekly OT caps, or predictive-scheduling notice rules are blocked at publish, with a rule citation.
- Auditable changes. Swap and exception workflows handle mid-cycle changes. Every edit is logged with reason, requester, and approver. Patterns and holiday templates apply across the roster.
Schedule grid · May 19 · Acme BPO
Schedule
scoped shifts
Activity legend
Compliance check passed · 0 LNT / ESA violations · published at 14:22
Feature 02
Multi-client forecasting engine — separate models, combined coverage
Each client's demand pattern modelled independently. Combined into a unified coverage requirement that respects skills, LOB eligibility, and jurisdictional constraints. Add a new client and the engine adapts; the existing models stay accurate.
- Per-client demand models. Each client gets its own seasonality, trend, and exogenous-event handling (promo launches, billing cycles, regulatory deadlines). The models are isolated so a RetailCo spike doesn't contaminate the FinCo forecast.
- Combined coverage. Requirements aggregate respecting skills × LOB eligibility, not just headcount. Erlang C is the math; the engine layers BPO-specific shrinkage and the multi-client occupancy mix on top.
- What-if scenarios. Model "new client at 50 agents launching June 1" before signing the contract. See hiring lead times, capacity gaps, and shrinkage impact in advance.
- Forecast accuracy tracking. Actuals vs. forecasts per client. The system identifies which models are drifting before they affect publishing.
- Standard WFM math. Erlang C, shrinkage, occupancy, ASA targets — tunable per client requirement. No black box; the math is visible and editable.
Forecast · Week of May 19
Multi-client volume forecast
Per-client demand modelled independently; combined coverage requirement below
Combined coverage requirement (Erlang C · BPO shrinkage layered) · 247 agents at peak
What-if · upcoming
New client launch · 50 agents · Jun 15
+45 hires needed · start May 10RetailCo summer promo · 30% volume lift
+12 hires · OT 60 h/wk for 4 wksFinCo Q4 billing cycle peak
Pre-staffed · no action neededFeature 03
Real-time adherence — agent state by client, not just on/off
Live view of every agent's state across every client they serve. On a call? Which client. On a break? Counted against today's allowance. Adherence broken down by client for SLA reporting that holds up under scrutiny.
- Multi-state model. Agent state = (logged-in × client × activity). Generic WFM treats agent state as a flat enum ("on call"); we treat it as a structured tuple so per-client reporting is exact.
- Client-disaggregated adherence. SLA reports per client roll up to the right contract. Your RetailCo monthly report has the right numbers because adherence was tracked per client all along, not approximated after the fact.
- Real-time triggers. When adherence drops below threshold for a specific client, the right supervisor gets paged — not a generic ops escalation.
- Auto-reconciliation. Late returns from break auto-flag for review rather than blind discipline. Agents who consistently come back on time but had one exception get a chance to explain.
- Live floor view. See who's where in real time, by client. Not "X agents logged in" but "X on RetailCo, Y on FinCo, Z on training, W on break."
Real-time adherence · floor view
Live floor view
247 agents · last refresh 8s ago
Mike Y.
On callFinCo · Inbound
4 min
Aisha O.
On callRetailCo · Inbound
2 min
Tyler B.
Wrap-upFinCo · ACW
45 sec
Maya P.
On callTelcoPro · Outbound
6 min
Noor F.
AvailableRetailCo · Idle
12 sec
Devon R.
Break— · Lunch
23 min
Priya K.
On callRetailCo · Inbound
1 min
Carlos D.
Late returnFinCo · Flagged
+3 min
Per-client adherence
RetailCo
96%
90 ag.
FinCo
91%
54 ag.
TelcoPro
94%
38 ag.
Real-time trigger: FinCo adherence < 92% for 10 min → supervisor paged · Carlos D. wrap-up flagged
Feature 04
Agent self-scheduling — picks, swaps, trades, bidding
Available shifts go on a pick board; agents claim what works. Mid-cycle adjustments happen through swap and trade workflows. Bidding for the next cycle's preferred shifts. All compliance-gated so agents can only do what the rules allow.
- Pick board. Open shifts visible to eligible agents only — skill-gated, LOB-gated, jurisdiction-gated. Mike sees the FinCo shifts he can claim, not the RetailCo refund shifts he isn't certified for.
- Shift swap. Agent A trades a Tuesday shift for agent B's Friday shift. The system validates both: rest periods, skill eligibility, coverage impact. The swap completes only if both sides pass.
- Trade board. Post a shift you want to give up; eligible agents claim it. Same compliance gating as direct swaps; the trade board is just a wider audience.
- Bidding for next cycle. Rank-ordered preferences for the next scheduling cycle. Seniority, performance tier, skills coverage all factor in. The bidding logic is configurable per client requirement.
- Compliance gating. Trades can't violate rest periods, OT caps, jurisdictional notice rules, or skill mismatches. Agents see only the actions that would pass all gates; ineligible options are filtered out, not rejected after the fact.
- Manager review. HR / ops review the queue of pending trades from one screen. Approve, hold, or reject with reason.
Self-scheduling · pick + trade
Open shifts · eligible agents see this list
Skill-gated · LOB-gated · rest-period gated
Sat May 24 · 12 pm – 8 pm
CSR · RetailCo
Sun May 25 · 8 am – 4 pm
Sr. CSR · FinCo
Mon May 26 · 2 pm – 10 pm
CSR · TelcoPro
Pending trades · ops review
Maya P. · Tue 8-4
Noor F. · Wed 9-5
Carlos D. · Sat 6-2
Every action passes 4 gates: skill match · LOB eligibility · rest periods · weekly OT cap.
Feature 05
Skills + competency matrix — the 2-D map of who-can-do-what-for-whom
Skills aren't a flat tag list. They're a matrix of capabilities × clients × expiry. Routing, scheduling, hiring, and training all reference the same matrix — no parallel skill databases drifting out of sync.
- Per-client certification. An agent is "billing-certified for RetailCo" — not just "billing-certified." The same skill can be certified for one client and not another.
- Expiry tracking. Certifications expire. The system flags renewals 30/60/90 days in advance so retraining slots can be scheduled before the cert lapses.
- LOB eligibility computation. Roll up skills × certifications into "who can serve this LOB right now" rosters. Hiring asks WFM "who else could we move to this LOB" and gets a real answer.
- Skill gap analysis. Find the N agents closest to filling a gap — existing skills, plus 1–2 training paths. Training spend gets targeted instead of broadcast.
- Routing integration. Voice routing sees the same eligibility matrix. No parallel tag lists in your CCaaS that drift out of sync with WFM.
Skills + competency matrix
Skill × client certification
Competencies are per-client, not flat. Expiry tracked.
Billing
Refunds
Tech I
Tech II
FR support
Aisha O. · Tech II / RetailCo expires Jun 12 · re-training scheduled May 30
Feature 06
Capacity planning — long-horizon staffing across clients
6-month and 12-month workforce planning. New client launches, seasonality, attrition modelling. The numbers Recruiting needs to plan their funnel and Finance needs to plan the budget.
- Per-client horizon. Model demand for each client 12 months out. The model respects each client's growth curve and contract milestones (e.g., "Client X is contractually adding 30 agents in September").
- Attrition modelling. Historical attrition by tenure / role / client feeds back into the staffing plan. The 90-day washout curve is baked in; you plan for it instead of being surprised by it.
- Hiring lead time. WFM tells Recruiting "you need to start hiring 12 agents for client X in May to meet July go-live." The number isn't a guess; it's derived from the forecast + ramp time + expected attrition.
- What-if scenarios. Model 3 client expansion scenarios — see hiring + training + budget impact of each. Finance leans on these for the annual plan.
- Monthly reconciliation. Actual vs. plan; identify drift early. When the plan said you needed 12 hires by May 1 and you have 8, the gap surfaces in the same place as the plan itself.
Capacity planning · 12-month horizon
Headcount plan · per-client · 12 months
Planned vs. actual · attrition modelled per tenure tier
Recruiting lead time: start hiring for new client X by May 10 (5 weeks pre-go-live).
Expansion scenarios — model before signing
RetailCo +30 agents
+30 over 8 wks+$1.2M / yrFinCo +50 agents
+50 over 10 wks+$2.0M / yrTelcoPro +25 agents
+25 over 6 wks+$1.0M / yrFeature 07
Intraday OT / VTO — same-day capacity adjustment
Volume spikes? Push OT offers to eligible agents in seconds. Demand crashes? Push VTO offers. Both compliance-gated, both audit-logged, both bounded by acceptance windows. Same-day capacity adjustment within the rules.
- OT offers. Push to agents with available capacity, the right skills, and no rest-period violation. Eligibility filtering happens before the offer goes out — agents only see what they could legally accept.
- VTO offers. Push to agents on shift with the right voluntary-time-off agreement. Some clients prohibit VTO during critical windows; the system respects per-client rules.
- Eligibility filtering. Skills match, jurisdictional rest rules, weekly OT caps, predictive-scheduling notice rules — all enforced before the offer reaches an agent.
- Acceptance window. Time-bound (typically 15–30 minutes); auto-expires. Agents who don't respond don't block coverage; the system pushes to the next eligible cohort.
- Audit chain. Who was offered, who accepted, who declined, why. Labour-relations defence baked into the workflow — when the union asks "why didn't Mike get the OT," the answer is one screen.
Intraday OT offer · OT-2891
Coverage gap detected: RetailCo EN · 14:00–18:00 · 4 agents short. OT offer auto-staged.
OT offer
RetailCo EN · 14:00–18:00 · 4 agents needed · +$32/hr · expires in 22 min
Eligible agents (skill match · no rest violation · OT cap available)
8 eligible · pushing in priority orderMaya P.
TelcoPro · ending 1 pm
Devon R.
Off shift
Aisha O.
RetailCo · ending 1 pm
Priya K.
Off shift
Carlos D.
FinCo · ending 1 pm
Acceptance window: 30 min · auto-expires · then pushes to next eligible cohort
All offers / responses logged · labour-relations defence audit chain
Schedule that knows about HR events — leave, probation, coaching, training
Schedules don't live in WFM alone. Leave approved in HR, probation review days, coaching blocks from QA, training enrollments — all surface on the schedule grid as activities. WFM doesn't reconcile across systems; the activities arrive as data. Same data model HR sees on the employee record — different role-tuned views.
- Leave blocks. HR-approved leave shows up on the agent's row as a non-productive block. The coverage shortage flags immediately so WFM can plan replacement before the leave starts, not after.
- Probation milestones. 30 / 60 / 90 review days highlight on the affected agent's row with a popover linking directly to the review. The manager doesn't forget; WFM doesn't double-book over the review window.
- Coaching blocks. When QA assigns a coaching session, WFM auto-protects the block from being scheduled over. The coaching team's calendar and the WFM grid stay in sync.
- Training enrollment. Training time blocks appear on the grid with the course attached. Capacity planning knows training reduces productive headcount this week; the forecast accounts for it.
- Mid-leave check-ins. Scheduled HR check-ins surface as tiny event markers on the agent's row, even while the agent is on leave. The HR person doesn't have to consult a separate calendar.
- Return-from-leave handoff. The day an agent returns, the grid shows the re-onboarding checklist for the manager. Schedule sync, training updates, role changes during the absence — all surface on the same row.
Schedule grid · May 19 · Acme BPO
Schedule
scoped shifts
Activity legend
Compliance check passed · 0 LNT / ESA violations · published at 14:22
Business outcomes for the people running the BPO
What this looks like in margin terms for the COO or owner.
Schedule cycle time compresses dramatically
The compression comes from drag-to-assign + automated compliance checks + multi-stakeholder negotiation in one tool. WFM team size doesn't have to grow with client count; the cycle scales linearly with seats added, not exponentially.
Adherence visibility per client improves SLA delivery
When you can see which client an agent is currently serving, your SLA reports become accurate. Clients trust them. QBR meetings stop being defensive; renewal conversations get easier.
Schedule violations caught at publish, not after
LNT, ESA, OT cap rules — all flag before publish, not after a complaint. Configurable rule packs handle predictive-scheduling jurisdictions (Oregon, Washington, NYC, San Francisco) per client requirement. Labour-relations defence baked into the workflow with full audit chain.
New client launch absorbed without breaking forecast
Bringing on a 50-agent client used to mean 30 days of chaos. The forecasting engine handles the transition; the schedule grid scales linearly; existing client SLAs stay intact during the ramp.
What WFM actually gets back
Directional outcomes — real magnitude depends on your prior schedule cadence, client mix, and forecast maturity.
Hours, not days
Per cycle. The compression comes from drag-to-assign + integrated compliance checks + multi-stakeholder negotiation in one tool.
Per-client SLA visibility
Multi-state adherence with the client context attached — your SLA reports finally reflect which client the agent was actually serving.
Zero LNT / ESA / OT violations slipping through
Compliance checks block publish; nothing slips. Jurisdiction-specific predictive-scheduling rule packs are configurable per client requirement.
Days, not weeks to absorb
Into the forecast model. The transition layer handles workforce mix changes automatically; existing client SLAs stay intact.
Regulatory posture
FrontLine Workforce vs. a generic WFM
Six things BPO WFM teams notice in their first month away from NICE WFM / Verint / Calabrio.
| FrontLine Workforce | Generic WFM | |
|---|---|---|
Multi-LOB agent assignment at the data layer (not as a tag) Generic WFM models multi-client as workforce tags; you need it as a structured (client × LOB × skill) tuple. | ||
Multi-client forecasting that adapts to new client launches Most WFM systems require a fresh forecast model when adding a client; you lose 30 days of pattern data. | ||
Real-time adherence with per-client state context Single-state adherence ("on call") doesn't tell you which client the call is for. Multi-state does. | ||
Skills as 2-D matrix (skills × client) — not a flat tag list Generic WFM treats skills as a flat list; you need cert-per-client to route correctly across LOBs. | ||
Compliance rule checks at publish (LNT · ESA · OT · notice) Some WFM systems support OT caps; few enforce jurisdictional notice rules or LNT-specific constraints. | ||
Native integration with HR (leave + probation + coaching on grid) HR events live in a separate system; WFM teams reconcile manually. We surface them on the grid. |
In production today
Workforce Management (Module 23) is shipped and running on customer tenants; Skills & Competency (Module 25) is shipped. See the technical status, full feature list, and weekly progress on the Atlas.
Questions WFM analysts actually ask before signing
Pulled from real fit calls. Short, direct answers.
How does multi-LOB agent assignment actually work in the data model?+
What forecasting engine do you use? Is it just Erlang C?+
Can we keep our current voice routing (Genesys, NICE CXone) and use FrontLine for scheduling only?+
How do you handle predictive-scheduling laws (Oregon, Seattle, NYC, San Francisco) — schedule changes have notice requirements?+
Can agents self-schedule without breaking compliance (rest periods, OT, skill mismatches)?+
Ready to see it in your environment?
Two ways to take the next step.
Other teams using FrontLine
WFM doesn't live alone. Here's what the rest of FrontLine looks like from the team next door.