Recruitment has a scheduling problem that nobody designed and everybody lives with.
Consultants work roughly nine to six. Candidates (the good ones, the ones in work) job-hunt at half past nine at night, on their phone, quietly, because they can't be seen doing it at their desk. Clients scope agencies between meetings and send enquiries at seven in the evening.
So both sides of your market are most active when your team isn't.
The candidate side
A candidate looking at a role on your site wants to know a short list of things before they'll register or apply:
- What does it actually pay? (Not "competitive".)
- Is it remote, hybrid, or on-site, and how many days?
- Is it permanent, contract, or temp?
- Where exactly is it?
- Who's the client, or at least what sector and size?
- What's the process and how long does it take?
If those are unanswered, most people don't ask. They move to the next listing. Candidate enquiries are low-effort to abandon.
A bot answering from your published role content converts a browse into a registration, at the hour the browsing actually happens.
Structured screening, gathered automatically
The first consultant call with a new candidate is mostly data collection: what are you looking for, where, for how much, when can you start, are you eligible to work here, what have you done.
That's a twenty-minute call, and a meaningful share of them end in "we don't have anything for you".
A bot can gather the same structure:
- Role type and level
- Location and commute tolerance, or remote preference
- Salary expectation
- Notice period
- Right to work status
- Key skills or sector experience
- Whether they're actively looking or open to hearing about things
That arrives as a record a consultant can act on. The calls they then make are with candidates who fit something live, which is a better use of consultant time and a better experience for the candidate, who isn't being interviewed for nothing.
Consultants routinely spend a large share of the week on first calls with candidates they cannot place. Gathering the basics beforehand lets them prioritise, and it means the candidates who are not a fit get a fast honest answer instead of a slow one.
The client side is worth more
An employer enquiry is worth many multiples of a candidate registration, and it arrives through the same website.
Configure the bot to identify it fast (anyone asking about fees, how you work, or saying they're hiring), and switch into a different conversation:
- What role, and how many?
- Permanent, contract, or temp?
- Salary or rate range
- Location and working pattern
- When do they need someone
- Are they working with other agencies
- Who's the contact and what's their number
Then route it to the consultant who covers that sector, immediately, with a flag. Speed matters as much on this side: an employer who's just decided to hire is talking to two or three agencies that week.
Fee questions
Agencies are traditionally cagey about fees, and employers increasingly won't engage without a number.
Publish your structure (percentage bands, whether there's a rebate period, how contract margins work), and let the bot answer. It filters out the employers who were never going to pay agency rates and accelerates the ones who will. Both outcomes save consultant time.
Compliance and candidate data
Two things to configure deliberately.
Right to work. The bot can ask whether someone has the right to work in the relevant country. It must not ask about nationality, visa specifics, or anything that strays into discriminatory territory. Keep the question simple and binary, and leave verification to your process.
Data. Be explicit in the bot's opening that details will be used to match them to roles, and keep collection proportionate. A chat window is not where a full work history should live. Gather enough to route and prioritise; the CV and detail belong in your ATS through a proper submission.
Configure it never to ask about age, health, family circumstances, or anything else that has no place in a screening conversation.
Candidate care during a process
There's a retention use as well. Candidates in an active process generate a steady stream of questions: when will I hear, what should I expect at the second stage, what's the dress code, who am I meeting.
Consultants answer these constantly and they're mostly generic. Handing them to a bot improves candidate experience (because they get an answer immediately instead of waiting for a consultant between meetings), and gives the consultant their day back.
Guardrails
Don't let it make placement promises or suggest a candidate is likely to get a role. Don't let it discuss another candidate's application. Don't let it negotiate salary.
And don't let it handle rejections. A candidate turned down by a chatbot will tell people, and reputation is most of what a recruitment brand has.
Configuring it properly
Point it at your website. Live roles, sector pages, fee structure, process explanations. Split routing at the first question: candidate or employer.
Set the candidate capture fields and the client capture fields separately, and flag employer enquiries as priority.
Measuring it honestly
Candidate registrations captured outside working hours, and time-to-first-contact on employer enquiries.
The second is the one that wins business. In a market where three agencies get the same brief, the one that responded in an hour rather than a day starts with an advantage nothing else compensates for.