E Erlangly

What Erlangly does

Everything a planner, a supervisor and an agent need, and nothing they don't.

Forecasts you can trust

  • Import interval history by CSV or send it through the API: contacts, handle time and abandons, per line of business or ACD queue.
  • A weighted forecast by weekday and interval, with growth, warnings when history is short, and your own overrides kept through recalculation.
  • Holidays in your jurisdiction are left out of the history.
  • Staffing per interval: Erlang C for voice, concurrency for chat, workload for email, with service targets, occupancy caps and shrinkage by date.
  • Forecast accuracy against what really came in.
A four-week forecast for one line of business: contacts and agents needed per interval.

Schedules in minutes

  • Shift patterns say what shifts can look like: days, start windows, lengths, and where breaks and lunches go.
  • Generation builds weeks at a time against the published forecast, for 500 agents in seconds, using contracts, binding availability, approved time off and multi-skilled agents across lines of business.
  • Compare drafts, hand-edit shifts and breaks, then publish. Every change after publishing needs a reason and is in the audit trail.
  • Several clients, one pool of agents: nobody is double-booked across clients.
A day of generated shifts with breaks and lunches, and the coverage strip.

Labour law built in

  • Signed rule packs for Canada (10 provinces and federal), Jamaica and 19 US states: rest between shifts, days in a row, daily and weekly hours, overtime, minimum shifts and break law.
  • Each person follows the law where they work: their own jurisdiction, their site's, or the organisation's.
  • Checked when schedules are generated, edited, swapped and published. Warnings need a reason; violations stop the change.
  • Tighten any rule yourself; loosening one needs an admin and a reason.
Labour rules for Prince Edward Island: each rule, whether it warns or blocks publishing, and the year's holidays.

The floor, live

  • Your ACD sends agent states to a simple HTTPS API, with its own token; if it can't call a web API itself, a short script in front of it can, and we'll help you write it. Its agent IDs and state codes are mapped in settings.
  • Adherence is scored every 15 minutes against the published schedule, and an adherence report shows it by person, team and line of business.
  • A wallboard for supervisors and wall TVs: who's out of adherence, time in state, alarms, saved views.
  • Exceptions (late, sick, meetings, overtime) are reported by agents and approved by supervisors, and adherence follows.
The live wallboard.

For agents, on their phones

  • Their week, with breaks, days off and time off, in their own time zone.
  • Ask for time off, swap a shift with a colleague who works the same line of business, tell their supervisor they're running late.
  • "Got it" tells the planner they've seen a new schedule; any later change asks again.
  • Emails when a schedule is published or a request is decided, once you connect your mail server.

Built for the way BPOs work

Several clients and lines of business on one install, with the same agents across them. Single-client centres simply don't see the client layer.

Self-hosted, and yours to automate

  • One Docker container on your server, installed with one command; upgrades snapshot your data and roll back by themselves if something's wrong.
  • Nightly backups, an update check you can turn off, no telemetry.
  • A JSON API with personal tokens for your own scripts and AI agents, and an erlangly command on the server.
  • Four roles (admin, planner, supervisor, agent) and a full audit trail.
$ erlangly schedule 2026-10-06 --lob billing
Tuesday 6 October 2026: 142 shifts
  06:00–14:30  BILLING    Andre Brown    break 07:30, lunch 09:30, break 12:00
  …

$ curl -H "Authorization: Bearer erl_…" \
    https://wfm.example.com/api/v1/shifts?from=2026-10-06