Two Agencies, One Role, Very Different Results
Imagine two recruitment agencies both working a mandate for the same type of role. A technology company in Lagos has asked each agency to find a Senior Product Manager. The role has a competitive salary and strong growth trajectory. It is a good mandate from a good client, and both agencies are motivated to deliver.
Agency A has been in business for 11 years. They have an established team and a solid reputation. They post the role on three job boards on Monday morning and start receiving applications by Monday afternoon. By Tuesday they have 140 applications. By Wednesday they have 210.
A senior recruiter starts working through the CVs on Wednesday morning. She spends about 4 minutes on each one. By Thursday afternoon she has reviewed 180 CVs, identified 22 worth considering, and emailed all 22. By Friday she has heard back from 9. She schedules a preliminary call with each of them for the following week.
On Wednesday of the following week 9 days after the role was posted she sends the client a longlist of 7 candidates with brief write-ups. The client reviews it over Thursday and Friday and asks for a shortlist of 3 for interviews.
Total time from role posted to shortlist delivered: 14 days.
Agency B posted the same role type on Monday. By Tuesday afternoon their automated screening had reviewed all 180 applications received to that point and ranked them against the role criteria. The 23 who met the defined threshold received an automated, professional acknowledgement and an assessment invitation an AI-generated set of questions built from the actual job description.
By Wednesday morning, 15 candidates had completed the assessment. The recruiter spent 90 minutes reviewing the ranked results, watching the assessment responses for the top 10, and selecting 6 for a brief video call. Those calls were scheduled for Wednesday afternoon and Thursday morning.
By Friday morning 4 days after the role was posted the recruiter sent the client a shortlist of 4 candidates with structured assessment scores, video interview notes, and a written evaluation for each one.
Total time from role posted to shortlist delivered: 4 days.
The client gave the mandate to Agency B for the next search before the first one had even resulted in a placement.
The Before: What Manual Hiring Actually Looks Like Day by Day
The story above is not exceptional. It reflects the difference between a manual hiring workflow and a structured, partially automated one. To understand why the gap exists, it helps to trace what actually happens at each stage in a manual process.
Day 1 to 3: The Inbox Fills
A job is posted. Applications start arriving. In a manual workflow, they arrive in an inbox or a basic ATS that functions as an organised inbox. Nobody is reviewing them yet the recruiter is managing other active roles and cannot pause everything to process applications the moment they arrive. The applications accumulate.
During this period, every hour that passes is an hour in which candidates are also applying elsewhere, taking other calls, and mentally prioritising opportunities based on which processes feel active and responsive.
Day 3 to 7: The CV Review Block
The recruiter sets aside time to review applications. This is typically a dedicated block a morning, a full day rather than a rolling review, because reviewing CVs requires sustained attention and is disrupted by interruptions. The recruiter works through the stack, making quick decisions, flagging some for follow-up, rejecting others.
The problem with a review block is that it is inherently a bottleneck. The recruiter can only review at a human pace. The quality of decisions made at the start of the block differs from the quality at the end, as discussed in the previous blog post on decision fatigue. And the decisions are made in isolation the recruiter's mental model of what they are looking for drifts slightly as they work through the stack.
Day 7 to 10: The Outreach Lag
The recruiter emails or calls the candidates who made the cut. Some respond quickly. Some take days. Some do not respond at all. The recruiter sends follow-ups. Some of the strong candidates have already accepted interviews elsewhere and are less available than they were a week ago.
The assessment, if there is one, gets sent separately — often after the initial call, often days later. Candidates who complete it do so at varying speeds. Chasing non-completions takes recruiter time. Reviewing results takes more time.
Day 10 to 14: The Longlist
Eventually a longlist is compiled. Write-ups are done. The client receives a document that represents a week and a half of recruiter effort on this one role.
The After: What a Structured, Partially Automated Process Looks Like
The alternative is not a fundamentally different philosophy about hiring. It is the same process application, screening, assessment, interview, shortlist executed with different infrastructure.
Hour 0 to 4: Applications Arrive, AI Screens Immediately
As soon as a candidate submits an application through the role's unique link, the AI reads their CV and scores it against the criteria the recruiter set when building the job. This happens in seconds. The candidate receives a professional, specific acknowledgement immediately. If they meet the criteria, they receive the assessment invitation in the same message. If they do not, they receive a respectful notification.
The recruiter does not need to be involved in any of this. They can be on calls, in meetings, or working other roles. The pipeline is processing.
Hour 4 to 24: Assessment Completion Begins
Candidates start completing the assessment. The recruiter checks the dashboard once or twice but does not need to action anything yet. The AI grades each completed assessment automatically. Candidates who score above the defined pass mark are flagged for the recruiter's review.
Day 1 to 2: The Recruiter Reviews
The recruiter opens the dashboard and sees a ranked list of candidates who have completed the assessment and scored above the threshold. They do not read raw CVs. They read structured summaries with key data points highlighted. For each candidate, they see the CV score, the assessment score, the competency breakdown, and the key strengths and concerns identified by the AI.
They select which candidates to invite for a conversation. They schedule calls or video interviews directly from the platform. Invitations go out automatically with scheduling links.
Day 2 to 4: Conversations and Shortlist
The recruiter has brief, focused conversations with a small number of pre-screened, assessed candidates. Because the early work is done and documented, these conversations can be shorter and more substantive the recruiter already knows the candidate's background and can focus on the specific questions that the screening and assessment flagged.
The shortlist is assembled from a position of genuine insight rather than a first impression. Every candidate on it has been evaluated through a consistent, documented process. The client-facing report is generated from the data already captured not written from scratch.
The Four Changes That Create the Biggest Difference
The gap between Agency A's 14 days and Agency B's 4 days in the opening story is not the result of a single dramatic change. It is the cumulative effect of four specific process improvements, each of which contributes to the overall outcome.
Change 1: Screening on receipt rather than in batches
The most important single change is processing applications as they arrive rather than in a batch review session days later. This eliminates the accumulation lag, maintains candidate engagement at its peak, and means the recruiter is always working with current, interested candidates rather than candidates who have been waiting for a week.
AI makes this possible because it can screen at a pace no human team can match hundreds of CVs in the time it would take a human to review one. The recruiter does not need to choose between responsiveness and thoroughness. They get both.
Change 2: Sending assessments immediately to qualified candidates
The second change is collapsing the gap between screening and assessment. In a manual process, the assessment comes after a call, after scheduling, after multiple exchanges. In a structured automated process, a candidate who passes screening receives an assessment invitation within minutes of their application being reviewed.
This keeps the candidate engaged at the moment their interest is highest. It gives the recruiter structured data before the first human conversation rather than after. And it means that by the time the recruiter speaks to a candidate, they know significantly more about them than they would from a CV review alone.
Change 3: Using AI to generate assessments from the job description
Generic assessments competency tests that are not specific to the role have lower completion rates and produce less useful signal than assessments that are visibly relevant to the job. When candidates can see why each question is relevant to what they would actually do in the role, they engage more seriously and produce more authentic responses.
AI-generated assessments built from the specific job description eliminate the time a recruiter would otherwise spend building or selecting an assessment manually. They also ensure that the assessment is always current and relevant not a generic test that has been on the shelf for two years.
Change 4: Having structured data before the first conversation
The most significant quality improvement in the entire process is that the recruiter's first conversation with a shortlisted candidate is informed by structured data CV score, assessment results, competency breakdown rather than a quick scan of a CV in the 30 seconds before the call.
This changes the conversation. The recruiter can go deeper on the areas that the assessment flagged as strong. They can probe the areas that the screening identified as potential gaps. They can use the time to assess fit, cultural alignment, and the candidate's genuine motivation for the role the things that only emerge through conversation rather than covering basic factual ground that the assessment already established.
The result is a shorter, more productive conversation and a more accurate assessment of candidate quality.
The Numbers Behind the Change
The shift from manual to structured automated hiring is not just faster. It is measurably better across multiple dimensions.
Time to shortlist drops from an average of 12 to 16 days to 3 to 5 days for roles with consistent application volumes. This is not a marginal improvement it is a 70 to 80 percent reduction in the time between posting and delivering a qualified shortlist.
Candidate drop-off between application and assessment completion drops significantly when the assessment is sent immediately after screening rather than days later. As noted in our earlier post on pipeline drop-off, assessment completion rates at 24 hours are nearly three times higher than at one week.
Shortlist quality improves because the recruiter is selecting from a set of candidates who have been evaluated through a consistent process rather than from a set that reflects the variability of a multi-hour manual review session.
Client confidence in the shortlist increases because every candidate comes with documented evaluation data. The recruiter can tell the client exactly what criteria each candidate was evaluated against, what their score was, and why they were included. This is a qualitatively different conversation than being able to say the recruiter reviewed their CV and thought they seemed strong.
What This Requires to Implement
The shift from a manual to a structured automated process does not require a large team, a long implementation project, or a significant upfront investment. It requires three specific things.
First, a workflow that is designed around automation from the start rather than treating automation as an add-on to an existing manual process. The job needs to be created with screening criteria defined. The assessment needs to be generated before applications are received, not after. The pass marks need to be set. This takes 15 to 20 minutes when a role is posted time that is returned many times over in the efficiency of what follows.
Second, a platform that connects each stage of the process so that a candidate moving from screening to assessment to interview does so within one system, not across three disconnected tools. The efficiency gains from automation disappear if the recruiter still needs to manually export data from a screening tool and import it into an assessment tool and then manage invitations from a separate calendar system.
Third, the intention to use the data the process produces. A structured automated pipeline generates more candidate data than a manual process does. That data is only valuable if the recruiter uses it reviewing assessment results before calls, using competency breakdowns to structure interviews, and building from the documentation the system produces rather than starting from scratch at every stage.
How Tiaago Makes This Possible
Tiaago connects every stage of the pipeline in one workflow. When a recruiter creates a job on Tiaago they set their screening criteria and the assessment pass marks in the same process. When candidates apply through the unique job link, AI reads their CVs immediately. Qualified candidates receive an AI-generated assessment within minutes. Candidates who pass the assessment receive a scheduling link. The recruiter's dashboard shows them exactly where every candidate is in the pipeline, what their scores are, and what needs their attention.
The recruiter's involvement is concentrated at the points where human judgment and human relationship actually matter reviewing shortlists, conducting interviews, and making decisions. Everything between those moments is handled.
At NGN 40 per resume screened and NGN 100 per assessment taken, the cost of processing 200 applications and sending 40 assessments is NGN 12,000. The cost of a recruiter spending two full days manually processing the same volume is significantly higher, measured in salary cost, and produces a slower, less documented result.
The 24-hour shortlist is not a promise. It is the logical outcome of a process that moves immediately rather than waiting for human bandwidth to become available.