Understand the work.
Choose three live roles and one recruiter. Inspect the CVs, notes and previous conversations, and capture the current workflow, outcomes and time spent. Agree the scope and proposed success measures before building.
A private proposal · AI SprintSECOND ORDER / SVA — 01
Your next placement
could already be
in your database.
Let’s find out. A focused, two-week experiment to turn candidate data and past conversations into people worth contacting now.
Open the proposal02 weeks03 proposed live roles£1,500 fixed investmentA small test. A clearer next move.
01 / The opportunityIntelligence, applied.
Not another match score.
A reason to reconnect.
Can we turn overlooked candidate records and past conversations into interested, qualified candidates for live roles—with less manual work?
Then · Previous conversation
“The role looks good.
But I can’t relocate.”
Now · A new opportunity
A relevant role.
This time, remote.
The connection
Surface the evidence. Explain what changed. Reconfirm what we don’t know.
02 / What we’ll doOne proposed experiment.
Find the people.
Start the conversation.
A focused experiment across three live roles, from reviewing your existing database to seeing what happens when your team reconnects.
Find people worth revisiting.
Use CVs, recruiter notes and previous conversations to identify relevant candidates your usual process may overlook.
Bring the picture up to date.
Refresh promising profiles with external information where available, adding technical evidence when useful.
Support the conversation.
Prepare personalised outreach and follow-ups, and capture current interest, availability and preferences.
See what moves forward.
Track responses, qualified candidates, progression and time spent against the current workflow.
This is our starting point. If the work reveals a more valuable opportunity, we’ll agree an adjusted focus within the same two-week sprint and budget.
03 / What it could look likeAn example of the candidate review.
A shortlist with
the story attached.
One role. 10–15 candidate cards. Enough evidence to make a better decision, without adding another complicated workflow.
Start with a job description and recruiter input: essentials, acceptable alternatives, location, salary and previous rejection reasons. Search the agreed pool, review the shortlist, then send approved messages through your existing workflow.
SVA / Candidate intelligenceIllustrative prototype · fictional data
Live role
Review shortlist Senior Backend Engineer
Python · AWS · Remote · £80–100k
AM
Alex Morgan
Backend Engineer · London
01 / Technical fit
Python and AWS in the CV. Production systems experience. Team leadership still to verify.
Source: candidate CV02 / Relationship context
Declined a previous role because relocation wasn’t possible. This role is remote.
Source: recruiter note · 14 Jan 202603 / Current availability
Unknown. Salary expectations and openness to a move need reconfirming.
Last contact: 14 Jan 2026 · Owner: SamFeedback captured in the prototypeWorth contacting · Rejection reason · Reply · Qualified conversation
Technical fit, relationship context and availability stay separate. A historic note—or recent GitHub activity—is not evidence that someone wants a new job.
04 / The proposed two-week plan10 working days. Evidence at every step.
Small enough to move.
Real enough to learn.
Baseline
Find people. Refresh the picture.
Build retrieval and evidence-backed candidate cards. Refresh promising profiles with available external information, adding technical evidence where useful. Prepare personalised outreach, follow-ups and feedback capture for the agreed candidate pool.
Review, then reconnect.
Compare candidates with the usual shortlist; where practical, review without revealing which method found them. The recruiter reviews and sends approved outreach through the existing workflow. Capture current interest, availability and preferences as replies arrive.
See what moves forward.
Support follow-ups and record responses, qualified candidates, progression, time spent and failure reasons. Deliver the prototype, findings and a continue / change / stop recommendation. Later responses can inform the evaluation beyond the sprint.
05 / What success could look likeProposed targets. Not promised outcomes.
Better than your baseline.
Or we haven’t proved it.
We’ll agree these proposed targets and what counts as an interested, qualified candidate before building. If we adjust the focus together, we’ll adjust the measures too. Outcomes are not guaranteed.
+3
Additional candidates worth contacting
On at least two of three roles, beyond the usual Ceipal shortlist. Track how many become interested, qualified candidates.
30%
Less manual work
To produce a comparable shortlist and prepare outreach, including review and corrections. Log our manual research time too.
Real
signals.
signals.
Interest, qualification & progression
Track responses, current interest, qualified candidates and next steps. Some outcomes may arrive after the two-week sprint.
The test is whether finding overlooked people, refreshing their profiles and supporting the conversation produces better outcomes with less manual work than the current workflow. Record rejection reasons and failed suggestions as carefully as the wins.
06 / What we need from youA little access. Minimal interruption.
Let us see the work.
Then let us get to it.
One point of contact, a short kickoff and a focused review. We’ll do the research and build independently, keeping day-to-day interruptions to your team low.
Ceipal access
Scoped access to the relevant roles, candidate records, CVs, recruiter notes and interaction history—so we can see how the team actually works.
Your communication tool
Slack, Teams or whichever business tool you use. Relevant channels help us understand handoffs, decisions and relationship ownership.
Internal knowledge
Team documentation, process guides, role briefs and your knowledge base. The context that helps us work without repeatedly asking the team.
No ATS access? A CSV export can work. Include CVs and dated interaction history where available. It is a less effective route for observing the workflow, and we’ll confirm what the export can genuinely support before building.
07 / The investmentPut the opportunity in your own numbers.
A small experiment.
A measurable upside.
£1,500Fixed fee / two weeks
A working prototype for the agreed hypothesis, an evidence-based evaluation and a clear recommendation. The proposed candidate experiment covers three live roles. AI token and experiment infrastructure costs are covered during the sprint.
Decide the next stepThe break-even calculator
Editable planning assumption, not a live exchange rate.
Break-even0.15of one placement
Illustrative ROI5.7x£8,500 after sprint fee
At £10,000 per placement, 15% of one placement fee covers the sprint.
Revenue-based illustration, not a forecast or profit calculation. ROI multiplier = (additional placement fees − £1,500) ÷ £1,500. Excludes delivery costs, taxes and future implementation fees.