Tell us where you are
Two minutes, in your own words. No forms to print, no appointment to book. A person reviews anything the system cannot decide on its own.
Plan
Skills the standard intake would miss
Target roles, ranked by fit and demand
| Role | Why you | Pay band | Demand | Competition |
|---|
Open roles, ranked against your plan
Pulled from connected boards and scored against your plan. Each one shows why it fits and what to emphasise. Logging an application creates the file-trail record your caseworker sees.
Close the biggest gap first
Does a person need to step in?
Today at Port Coquitlam
Every client, sorted by who needs you
| Client | Path | Readiness | Status | Next action |
|---|
Human decisions
Select a client
WorkBC Pathways: one system, many paths
A six-month, 50-client pilot testing whether individually tailored, AI-assisted employment plans can help British Columbians reach better work faster while reducing administrative workload for caseworkers.
The problem
WorkBC serves over 100,000 people a year through 102 community centres. Every one of them gets broadly the same thing: a standard intake interview, generic workshops, and weekly activity reporting that measures effort rather than fit. A senior professional and a first-time job seeker receive the same tools. Caseworkers spend most of their week on data entry and compliance, and their discretion is reserved for paperwork instead of for the people who need it.
The result is slow, expensive, and uniform. Clients with a specific, fixable gap wait weeks for a referral that a well-briefed system could have scoped in minutes. Clients in real difficulty get the same fifteen-minute check-in as everyone else.
The model
Pathways inverts the default. The system does the parts that are pattern work: reading a work history for transferable skills, mapping current BC demand, building a short learning plan, tailoring applications, and writing the case note. A caseworker is pulled in only when something requires judgement or authority: a funding decision, a hardship, a credential question, or a client who is not progressing.
| Step | Today | In the pilot |
|---|---|---|
| Intake | Booked appointment, standard questionnaire | Self-serve in minutes; a person reviews only flagged files |
| Action plan | Generic, built in a meeting | Individual, built from the client's actual history and live demand |
| Upskilling | Referral to a course, weeks of approval | Free micro-learning starts same week; paid training goes to a human with a written rationale already drafted |
| Job search | Volume quota, logged by hand | Targeted applications, logged automatically with a file trail |
| Reporting | Caseworker types it | System drafts it; caseworker approves |
| Check-ins | Same cadence for everyone | Set by readiness score and escalation flags |
What stays with people
- Identity verification and eligibility
- Approving any public money: training, equipment, wage subsidies
- Hardship, housing, legal and immigration questions
- Clients who stall for two review cycles
- Signing declarations. The client signs their own, every time
- No impersonating a client to staff, by voice, video or text
- No fabricated experience, applications or attendance
- No evading portal security; integration is by agreement or by export
- No decisions without a named reviewer on the file
- Client data stays in Canada and is deleted on exit
Three commitments the pilot is built on
Every decision is transparent. A referral, a funding decision, or a refusal cannot be recorded without a decision record: the criteria that apply, quoted from the program rules; which were met and which were not; the evidence used; who decided; and how to ask for a review. The client and the caseworker see the same record. A refusal without reasons is not possible in the system, because the form will not save.
Every decision is fair and based on facts. The same intake produces the same plan at every centre, for every caseworker. Scores and targets show their inputs. A client can dispute a fact, the plan recalculates, and the change is logged. Discretion still exists, but it is written down and reviewable.
Every client gets sector expertise. No caseworker can be an expert in trades, health care, logistics, early childhood education and digital production at once. The system can. Each plan opens with a sector brief for that field in BC: how hiring really works, which credentials matter, where roles are posted, realistic pay, and what a strong application looks like. The caseworker reads the same brief, so a generalist can hold an informed conversation with a specialist client.
- Eligible for WorkBC case-managed services Met
- Viable business idea with a target market Met
- Not already operating the proposed business Met
- Gap in self-employment knowledge or skills that the program addresses Disputed
Strengths and weaknesses, stated plainly
A pilot only earns trust if it is honest about both sides. This is our assessment of the current model and of Pathways, and the pilot is designed to test the claims in both columns.
- Physical centres in 102 communities; reaches people with no device or connection
- Staff who know local employers and the local community
- Established eligibility, funding and privacy rules that already carry legal weight
- Human judgement for hardship and complex lives
- Accountable under FOIPPA and an existing complaints process
- One generic service for every client, regardless of field or level
- Decisions are often verbal, unrecorded, or given without criteria, so they are hard to review
- Outcomes depend on which caseworker and which centre you get
- No caseworker can be an expert across every sector; specialist clients are advised by generalists
- Staff time goes to data entry and compliance, not to the people who need it
- Weeks from intake to action; volume quotas measure effort, not fit
- Individual plan in hours, built from the client's real history and live BC demand
- Sector expertise for every field, available to client and caseworker alike
- Every decision recorded with criteria, evidence, decision-maker and review path
- Consistent: same intake, same rules, same output at any centre
- Caseworker time moves to funding decisions, hardship and clients who stall
- Full file trail for every application and activity; audit is a query, not a phone call
- AI can be wrong, and confidently so: a bad pay band or a wrong credential path misleads a client. Mitigation: sources shown, facts disputable, caseworker sign-off on anything that commits money
- Bias in training data can steer clients by name, gender or origin. Mitigation: evaluator audits plans across demographic groups; intake hides name and origin from the planning step
- Clients without a device, connection or digital confidence are left behind. Mitigation: assisted intake at the centre; the system serves staff as well as clients
- Over-reliance: staff may stop reading and simply approve. Mitigation: approval requires editing or annotating the draft; sampled approvals are reviewed
- Privacy: a rich profile is a rich target. Mitigation: Canadian hosting, data minimisation, deletion on exit, privacy impact assessment before day one
- Accountability is unclear when the system contributes to a decision. Mitigation: a named person owns every decision; the system advises, it never decides
- Unproven at scale. Mitigation: this pilot, with a control group and an independent evaluator
Pilot design
Fifty clients at one WorkBC centre, recruited at intake and randomly assigned: 25 on Pathways, 25 on standard service, tracked for six months. Same eligibility rules, same funding rules, same staff. An independent evaluator holds the data.
| Measure | How | What we expect to see |
|---|---|---|
| Time from intake to action plan | Timestamps | Hours, not weeks |
| Staff hours per client per month | Time logs | Materially lower, with hours shifted to flagged clients |
| Employment at 90 and 180 days | Client and employer confirmation | At least equal; the pilot tests whether it is higher |
| Wage at placement | Offer letters | Higher median, because targeting replaces volume |
| Client satisfaction | Survey at 30, 90, 180 days | Higher; fewer redundant appointments |
| Audit accuracy | Sample check of logged activity | Every entry has a file trail |
| Decision transparency | Share of decisions with a complete record: criteria, evidence, decision-maker, review path | 100% in Pathways; measured, not assumed, in standard service |
| Consistency | Same anonymised intake run at three centres; evaluator compares plans | Same targets and readiness within a narrow band |
| Fairness | Evaluator compares outcomes and plan quality across gender, age, origin and sector | No gap larger than in standard service; any gap found is published |
| Sector accuracy | Industry reviewers rate sector briefs and pay bands for 10 fields | Rated accurate and current by people who hire in that field |
| Reviews and complaints | Count of review requests and outcomes | Fewer complaints; faster, written resolutions |
Targets are hypotheses. The pilot exists to measure them, not to assert them.
What we need from the Province
A sandbox agreement with one centre operator; permission to run intake and reporting in parallel with the existing portal for the pilot cohort; a data-sharing and privacy agreement under FOIPPA; and an evaluator named by the Ministry. No new legislation. No change to eligibility or funding rules.
Why now
The Province's own Future of WorkBC engagement asked what the next service model should look like. The tools to build individually tailored plans at scale exist today and are cheap. The only open question is whether they produce better outcomes for British Columbians than the standard offer. That is a question a small, honest pilot can answer in six months.