A job post is the canonical way to brief the platform on a piece of work and start receiving applications from matched experts. The composer lives at /client/jobs/new and combines an AI-assisted import, a structured form, and (on Pro) live previews of how the brief is matching the expert base. This article walks through what each part does.
Who can post a job
Any member of a client organization with posting access can use the composer. This includes the org owner and any invited member assigned the right role. The composer resolves access through org membership, not the original signup user, so invited admins can post on behalf of the organization the same way the owner does.
Each post consumes one job token from the organization's balance. If the balance is zero, the composer redirects to /client/credits/purchase before the form opens.
Quick Import AI
The composer opens with a Quick Import option that pre-fills most of the form from an existing ToR document. There are two import modes: pasting the ToR text directly into a panel, or uploading the file (PDF or document).
The AI extracts the title, description, categories, services, and other fields that can be inferred from the source. Fields filled by the AI are visually marked with an amber border so the client knows which values came from automatic extraction and which were entered manually. Suggested services and categories appear as accept-or-skip chips above the relevant inputs.
Quick Import is available on Starter, Basic, and Pro, every tier that can post a job. The file or pasted text is sent to Google's Gemini API for parsing and is not stored by ConsultEarth in its own storage or database. The structured output is what reaches the database, and only after the client reviews and submits the brief. See /help/ai-features for the full account of what AI features do with submitted data.
The AI output is a draft. Every field should be reviewed before submitting, since the extraction occasionally mis-classifies budget format, categories, or required services.
Writing the brief
The form collects the following fields:
| Field | Purpose |
|---|---|
| Title | Short headline for the brief |
| Description | The body of the ToR. Paragraph-length is the floor; a few hundred words is typical for useful applications. Description quality drives applicant quality. |
| Budget | Optional. Stored as a number; not exposed to applicants until they apply |
| Country | Where the work is based or where the deliverables are needed |
| Duration | Free-text estimate (for example "3 months" or "ongoing") |
| Project type | Remote, On-site, or Hybrid |
| Minimum years of experience | A floor used by the matching algorithm |
| Expertise categories | Drives matching; subject to taxonomy. See below. |
| Services | Drives matching; mix of taxonomy items and free-text. See below. |
| Languages | Languages the expert must be able to work in |
| Proposal deadline (days) | How many days applicants have to submit a proposal. Default 10, and capped to the listing window (it cannot run longer than the brief stays live). |
| Required documents | Optional. A list of files the applicant must attach to a proposal, each with an instruction string |
| Deliverables | Structured, numbered list (3 to 7 items recommended). Each deliverable has a title and an optional description. Carried into the engagement workspace if the client later hires an applicant. |
The structured deliverables list replaces what used to be a single free-text scope field. Applicants read the brief as a numbered list of outcomes, not a wall of text. If the client later hires one of the applicants, those same deliverables seed the engagement workspace, so the brief and the work record share one source of truth.
Categories and services
Categories and services are the two axes the matching algorithm uses to rank applicants. They are not interchangeable.
Expertise categories are coarse-grained and drawn from a fixed taxonomy. They function as exact matches: an expert with the category "Climate Risk" matches a brief with that category, and an expert without it does not. Pick the smallest set that genuinely describes the work; over-tagging dilutes the match signal.
Services are finer-grained verbs (or short noun phrases) describing concrete activities. They are matched fuzzily using token similarity (Jaccard with leading-verb stripping), so "Develop M&E framework" and "M&E framework development" score the same. The composer offers existing taxonomy items as a search dropdown and accepts custom entries up to 100 characters when no existing item fits.
Both axes feed a matchscore that ranks applicants in the client's pipeline and ranks briefs in each expert's opportunity feed.
What posting costs
Every post consumes one job token, and the token price depends on the client's tier. Starter is $50 per post, Basic is $200 per post, and Pro includes four tokens a month with discounted $150 top-ups after that. See /help/basic-vs-pro for the full comparison.
How long a brief stays live and how many applicants it accepts are both capped by tier. The client can choose a shorter listing window at post time, but never a longer one.
| Tier | Maximum listing window | Maximum applicants per brief |
|---|---|---|
| Starter | 14 days | 15 |
| Basic | 30 days | 25 |
| Pro | 60 days | Unlimited |
When a brief reaches its applicant cap, it hard-closes to new applications, so the client is not left sorting a wall of marginal interest. When the listing window ends, the brief auto-expires and stops accepting applications; the expired record stays in the client dashboard for reference and can be relisted if the work is still active. On Starter, the listing clock starts when an admin approves the brief, not at the moment of posting.
What happens after posting
Starter briefs are held for manual admin approval before any expert can see them. Status starts as pending_vetting, and the brief stays hidden until an admin approves it; there is no automatic publish. Admins check the brief is real, the categories and services are plausible, and that no obvious policy issues are present. Approval typically takes 24 to 48 hours, though it can run longer.
Basic and Pro briefs publish immediately with status open and appear in the expert feed as soon as the form is submitted. Neither tier is held for vetting.
Once a brief is live, matched experts see it in their opportunity feed sorted by matchscore. Experts express interest, request to apply, and submit proposals through the standard flow. Each interaction adds the expert to the per-job pipeline, where the client moves applicants through stages: Applied, Shortlisted, Interviewing, Offered, Hired, Declined, or Withdrawn.
Starter clients run the pipeline up to interviewing but finalize the hire off-platform. A Starter client can shortlist, interview, review proposals, and message applicants, but cannot extend an on-platform offer, the step that opens an engagement workspace. To hire and manage the work on ConsultEarth, a Starter client upgrades to Basic or Pro.
Pro features visible while composing
Pro clients see two extra panels appear once the categories and services are filled in:
- Indicative match count. A live estimate of how many experts on the platform match the current draft. The count only appears when there are 20 or more matches. Below that threshold, the panel stays empty to avoid over-narrowing the brief based on a small-sample signal.
- Past experts panel. For organizations that have hired before, a ranked list of past-engaged experts whose profiles match the current brief. Each row has an invite checkbox; selected experts receive an "invite to apply" notification when the brief is submitted, riding the standard application flow rather than a separate back-channel rehire.
Both panels recompute as the client edits the brief. Neither blocks submission; they are advisory.
Related
- Basic vs Pro: which tier fits: /help/basic-vs-pro
- What our AI features do with your data: /help/ai-features
- How expert names are masked: /help/name-masking