The candidate database nobody opens anymore
- Keyword search that misses people. You have to guess the right words, their variants and their synonyms. Anyone who described their work differently stays invisible.
- Known candidates, searched for elsewhere. When you can't find a past candidate, you start from scratch on job boards, with all the cost and delay that involves.
- Incomplete candidate profiles. A resume imported without being read is just a file. To know what the candidate can do, you have to open it again.
What happens to your resumes
- Every resume becomes a profile. The AI extracts work history, skills, education, languages and location, and stores them in the candidate profile.
- Search in plain language. You type "senior Java developer in Lyon, open to remote work", just as you'd say it to a colleague, and the database finds similar profiles.
- Suggestions for every new role. As soon as a job is posted, the people in your database who match it are suggested to you, with their score and its reasons.
Job board resume database or internal candidate database: what's the difference?
Both are useful, but they don't serve the same moment.
| Job board resume database | Internal candidate database | |
|---|---|---|
| Who the candidates are | People registered on the job site | People who have already applied to you |
| What you know about them | Their resume | Their resume, your conversations, your evaluations |
| Cost | Often pay-per-use or subscription | Already built |
| Competition | The same profiles are seen by other recruiters | Profiles only you know |
| What Delectus does | Nothing: Delectus doesn't give access to them | Search by meaning and suggestions for every role |
Job board resume databases vs. your own candidate database: what's the difference?
"Resume database" can mean two very different things. Job boards offer resume databases, free or paid, made up of the resumes of candidates registered on their site. A company or an agency also has its own internal candidate database, made up of every resume it has received over the course of its hiring.
Delectus works on the second kind. Your internal candidate database is often full of people you've already met, who know your company and who have sometimes grown since. It's the fastest source for a new role, as long as you can actually search it.
How does an AI candidate database work?
Every resume, whether it comes in through a job opening or an import, is read by the AI and turned into a structured profile. Each candidate's background is then represented in a way that lets it be compared to a role or a search. A traditional search compares words. This one compares jobs and experience, so a "digital project manager" comes up in a search for a "web product owner" when their background justifies it.
Search happens at two moments. You run it yourself, in everyday language, when a need comes up. And it runs on its own every time you post a job, giving you the closest past candidates, their match score and the reason for each suggestion.
Searching your candidate database in plain language
You describe the profile the way you'd describe it to a colleague, with the job, the experience and what matters for the role. Results are then narrowed with filters on experience, skills or location. Here are a few searches a recruiter could type exactly as written.
- "Bilingual English executive assistant, law firm experience"
- "Industrial maintenance technician, electrical certification, Lille area"
- "IT project manager who has already led an ERP migration"
- "B2B sales rep in healthcare, willing to travel"
Reading a profile without reopening the resume
Every result opens an easy-to-read candidate profile. A summary sums up the candidate's background with their strengths and gaps for the target role, and the match score is broken down criterion by criterion. Your notes, tags, expected salary, notice period and the history of your exchanges are all in one place, including those from a past search.
For a specific question, the AI chat queries the profile for you. You can ask whether the candidate has worked in a factory, managed a team or how long they've used a piece of software, and the answer is based on the resume.
Example: reviving your candidate database for a new role
A company opens a digital project manager role. When it posts the job, it gets around ten profiles it already knows, including a finalist from a previous search, two recent unsolicited applications and a former intern who has since taken on more responsibility. It adds them to the job's pipeline, alongside the new applications.
A candidate you met a year ago may have changed jobs, which is why it pays to reach out again before putting them forward. The SMS prescreening agent handles that. It introduces itself as an AI, asks for the candidate's consent, then checks their interest and availability for the new role. The voice prescreening agent can take over by phone. The answers update the candidate profile, and you're notified when a candidate replies.
Importing an existing candidate database
Resumes received for your job openings go straight into the candidate database. An existing database, whether it comes from an old tool or from your ATS, can be imported. Each imported resume is read, turned into a structured profile and immediately becomes viewable, comparable with open roles and searchable.
An old database gets back the value it had lost. Resumes nobody opened for lack of time become profiles you can find, read and contact again.
A candidate database for companies, agencies and IT consulting firms
In a company, it's mainly used to find finalists from past searches before posting a job. In a recruitment agency, the pool of candidates it has met is its main asset, and the candidate database becomes every recruiter's day-to-day tool. In an IT consulting firm, it brings available consultants and already qualified candidates together in one place when a client request comes in.
Candidate databases and GDPR: the rules to follow
Keeping resumes is processing personal data. The recruitment guide from the CNIL, France's data protection authority, recommends not keeping data on an unsuccessful candidate for more than two years after the last contact, unless they agree, and informing them that it's being kept.
In Delectus, resumes stay in your workspace, hosted in the European Union, and aren't used to train any model. The rest depends on your own practices.
- Tell candidates, when they apply, that their resume may be kept and reused for other roles
- Delete or anonymize resumes once the set retention period is over
- Respond to access and deletion requests
- Restrict access to the candidate database to the people who recruit
Frequently asked questions
What is a candidate database?
It's the collection of resumes a company or an agency has received or gathered. An AI candidate database analyzes them so you can search for people by skills and experience, not just by keywords.
Can I import an existing candidate database?
Yes. Resumes from your old database or your ATS can be imported, and each one is analyzed and turned into a searchable profile.
Does Delectus give access to outside candidates' resumes?
No. Delectus works on your own resumes, the ones you've received and the ones you import. It isn't a resume database you rent.
How do you search a candidate database effectively?
Describe the profile by its responsibilities and skills rather than a list of keywords, then narrow it down with the experience, skills or location filters. Searching by meaning also finds people who described themselves in other words.
How long can you keep a resume in a candidate database?
The CNIL, France's data protection authority, recommends not keeping an unsuccessful candidate's resume for more than two years after the last contact, unless the candidate agrees to a longer period.
What is an AI candidate database?
It's a resume database you search by meaning rather than by keywords, and that suggests on its own the people who fit a new role.
Can you search a candidate database without writing a Boolean query?
Yes. You write what you're looking for in everyday language, as you'd say it to a colleague, and the database finds similar profiles.
Can I reach out to past candidates from the candidate database?
Yes. Once the profiles are added to a job opening, the SMS prescreening agent texts them to check their interest and availability, and the voice prescreening agent can call them.
