What Is an AI Proposals Column? How Edera AI Surfaces Top Candidates

The candidate discovery problem

When applications arrive for an open role, they typically land in a single "New Applications" or "Applied" column on the recruiter's Kanban board. A role with 120 applicants means 120 candidate cards in that column, all looking roughly the same. The recruiter's task is to work through them sequentially, one by one, to identify who deserves to advance.

This is fundamentally a discovery problem. The information the recruiter needs — which candidates are the strongest matches — is buried inside 120 individual CVs. There is no way to see at a glance which candidates are worth reviewing first without reading every application. The strong candidates are mixed in with the weak ones, and the only way to separate them is manual effort.

Traditional approaches to this problem include sorting by date (which tells you nothing about quality), filtering by a single keyword (which is too crude to be useful), or reviewing a random subset (which risks missing the best candidates). None of these approaches actually solve the discovery problem. They just manage it slightly differently.

How the AI Proposals column works

The AI Proposals column in Treegarden solves the discovery problem by creating a dedicated Kanban stage that is populated automatically with top-scoring candidates. Here is exactly how it works:

Step 1: Scoring. When the recruiter initiates AI scoring for a job, Edera AI analyses every candidate's CV against the job requirements. Each candidate receives a score from 0 to 100 based on configurable weights across skills, experience, education and keywords.

Step 2: Selection. Candidates who score above the threshold are identified as strong matches. These are the applicants whose profiles align most closely with what the role requires, based on the weights the recruiter has configured.

Step 3: Placement. The selected candidates are placed in the AI Proposals column on the Kanban board. This is a separate stage that sits alongside the recruiter's existing pipeline stages (Applied, Phone Screen, Interview, Offer, etc.).

Step 4: Review. The recruiter opens the job and immediately sees their strongest candidates in the AI Proposals column. Instead of starting with 120 unsorted applications, they start with a curated shortlist of the best matches. They review these candidates first, make decisions, and then optionally review the remaining applicant pool.

Visual prioritisation with score badges

Candidates in the AI Proposals column (and throughout the Kanban board) display colour-coded score badges that provide instant visual information about match strength:

Green badges (70-100%) appear on candidate cards with strong matches. When you see a cluster of green badges in your AI Proposals column, you know these candidates closely match the role requirements across the weighted dimensions.

Blue badges (40-69%) appear on moderate matches. These candidates have relevant elements in their profile but also have areas where their background diverges from the ideal specification.

The badges eliminate the need to open each candidate's profile to understand their relative strength. A recruiter can scan the AI Proposals column and immediately identify which candidates to review first, second and third — all without clicking into a single profile.

How it integrates with existing Kanban workflows

The AI Proposals column does not replace or disrupt existing recruitment workflows. It adds to them. Here is how it fits into a typical Kanban-based hiring process:

Existing stages remain unchanged. Your Applied, Phone Screen, Interview, Offer and any custom stages work exactly as before. The AI Proposals column appears as an additional stage, typically positioned before the first manual review stage.

Drag-and-drop works normally. Candidates in the AI Proposals column can be dragged to any other stage, just like candidates in any other column. When a recruiter reviews a proposed candidate and decides to advance them, they drag the card to Phone Screen or Interview. If they decide the candidate is not right despite the AI's recommendation, they can drag them to Rejected or simply leave them.

Manual additions still work. Recruiters can still manually advance any candidate from any stage, including candidates the AI did not propose. The AI Proposals column is a recommendation, not a constraint. The recruiter always has final authority over who advances.

Why a dedicated column matters more than a sort order

Some ATS platforms offer AI scoring but present it only as a sort order or a filter on the existing application list. You can sort candidates by score, highest first, and review from the top. Why is a dedicated column better?

Visibility without action. A sort order requires the recruiter to actively choose to sort by AI score. A dedicated column surfaces top candidates automatically, every time the recruiter opens the job. The AI recommendation is visible by default, not hidden behind a dropdown.

Separation of concerns. When AI-proposed candidates are in their own column, the recruiter can clearly see: here are the candidates the AI recommends, and here is the full applicant pool. This separation makes it easy to compare the AI's recommendations against the recruiter's own assessment of the pool.

Progress tracking. A Kanban column has a natural "empty when done" workflow. The recruiter works through the AI Proposals column, advancing or declining each candidate, until the column is empty. This gives a clear sense of progress that a sorted list does not provide.

Team collaboration. When multiple team members access the same job, the AI Proposals column provides a shared understanding of which candidates the AI identified as strongest. A sort order is a personal view; a column is a shared workspace element visible to the entire hiring team.

A practical workflow with AI Proposals

Here is how a typical recruiter's workflow changes with the AI Proposals column:

Monday morning. You open a job that received 85 applications over the weekend. Instead of facing an unsorted stack of 85 candidates, you see 12 candidates in the AI Proposals column, each with a green or blue score badge.

First 30 minutes. You review the 12 proposed candidates. You open each profile, check their CV and score breakdown, and make a decision. Eight look strong — you drag them to Phone Screen. Three are reasonable but not quite right — you leave them for now. One is a clear mismatch despite a moderate score — the AI over-weighted a skill that is less important than the weights suggested.

Next 15 minutes. You skim the remaining 73 candidates in the Applied column. With the green and blue badges visible on all candidate cards, you quickly spot two more candidates the AI did not propose but who caught your eye. You advance them to Phone Screen as well.

Total time: 45 minutes. Without the AI Proposals column, reviewing all 85 candidates manually would have taken 4-7 hours. You have identified 10 strong candidates in under an hour, with confidence that the AI's systematic analysis did not miss obvious matches in the remaining pool.

What the AI Proposals column does not do

Setting clear expectations about what this feature is and is not helps teams use it effectively:

It does not make hiring decisions. The column surfaces candidates for review. The recruiter decides who advances. No candidate is hired or rejected because of their presence or absence in the Proposals column.

It does not run automatically. Scoring is user-initiated in Treegarden. The recruiter decides when to run AI scoring. The Proposals column is populated when scoring runs, not when applications arrive.

It does not replace recruiter judgement. The AI evaluates match against configurable criteria. It does not assess motivation, cultural fit, communication quality, or the dozens of intangible factors that experienced recruiters consider. It narrows the field; the recruiter evaluates who is in it.

It does not hide candidates. Every applicant remains visible on the Kanban board. The Proposals column highlights the strongest matches; it does not suppress the others.

Getting accurate proposals through weight configuration

The quality of the AI Proposals column is directly determined by how well the scoring weights are configured for each job. Because Treegarden allows per-job weight configuration, the recruiter controls what the AI prioritises.

For a senior backend engineering role, you might configure: skills 45%, experience 30%, education 10%, keywords 15%. The AI Proposals column will then surface candidates with the strongest technical skills and deepest relevant experience, which is exactly what a senior engineering hire requires.

For a graduate marketing analyst role, you might configure: skills 20%, experience 10%, education 45%, keywords 25%. The Proposals column will prioritise candidates with relevant degrees and industry vocabulary, which is appropriate when experience is expected to be limited.

The weight configuration is the mechanism through which the recruiter's hiring knowledge translates into the AI's selection criteria. Taking two minutes to configure weights before running scoring is the single most impactful step for ensuring the Proposals column surfaces the right candidates.

Frequently asked questions

What is the AI Proposals column in Treegarden?

The AI Proposals column is an automatically created stage on your Kanban hiring board. When Edera AI scores candidates for a job, it identifies top-scoring applicants and places them in this dedicated column, giving recruiters a prioritised view of the strongest candidates.

Does the AI Proposals column replace the recruiter's decision?

No. The column surfaces candidates for the recruiter's review — it does not make hiring decisions. Candidates in the Proposals column still require human evaluation before advancing. Recruiters can move candidates, ignore the AI's recommendation, or review candidates from other stages.

What happens to candidates not placed in the AI Proposals column?

Nothing negative. They remain in their original stage on the Kanban board. They are not rejected, hidden or removed. The recruiter can still review and advance any candidate regardless of AI selection.

Can I customise which candidates appear in the AI Proposals column?

You control the scoring weights that determine which candidates score highest. By adjusting per-job weights for skills, experience, education and keywords, you influence who appears in the Proposals column.