Service
AI lead qualification and lead management
Most SMEs do not lack leads. They lack time to sort them. An AI front desk answers every enquiry, asks the questions a good salesperson would ask, scores the answers, follows up on a schedule, and hands a person a short brief when the lead is ready. Kyn Technology builds this for SMEs in Singapore and Southeast Asia, on top of the channels and CRM already in use.
A focused first release reaches production in about 14 days once scope and access are settled. Rights in the custom work are assigned to the client on full payment, as set out on /legal.
On this page
What the system does
Answers within minutes, at any hour
The front desk agent replies on WhatsApp and Instagram from approved information: services, price ranges, lead times, location. Questions outside that information go to a person.
Asks five qualifying questions
The questions are set per business. A typical set: what is needed, when, where, the budget range, and who decides. The agent asks them in conversation, not as a form, and records each answer against the lead.
Follows up on day 1, 3 and 5
Leads that go quiet get a follow-up on day 1, day 3 and day 5, each one referring to what the customer already said. After day 5 with no reply the lead is marked cold and leaves the active list; the schedule is set per business.
Checks availability before promising
Where the business sells dates or capacity, a date-clash rule checks the calendar. If the requested date is taken, the agent offers the nearest open date instead of a refusal.
Hands over with a brief
When a lead is ready, the assigned salesperson receives a copy-ready brief: who, what, when, budget, decision-maker, what was asked and what was promised. The first sales call starts at question six, not question one.
Records why deals are lost
Lost leads carry a reason from a fixed list: price, date, no reply, chose a competitor, out of scope. Each reason has an owner and a fix, so the list drives a change, not a report.
Two scores, not one
A single lead score hides the difference between a customer who has given every detail but is shopping around, and one who has given two details but asked to visit tomorrow. The build keeps two measures apart.
Readiness meter
Five details, five slots. Each answered question fills one slot. A lead at 4 of 5 has enough information for a quote.
Buying signals
Weighted marks for behaviour: asking for a price, requesting a site visit, requesting a sample, replying quickly, naming a date. Weights are set with the business and adjusted when results come in.
When a person is notified
A salesperson is notified when readiness reaches 4 of 5, or immediately on a price, visit or sample request, whatever the readiness. The rule is written down and visible on each lead, so nobody argues about why a lead was or was not passed on.
The method is explained step by step in the guide how SMEs qualify leads without a sales team.
A build in production
For an events and creative studio, KYN built the front desk described above: five qualifying questions on WhatsApp and Instagram, follow-ups on day 1, 3 and 5, a five-detail readiness meter kept separate from weighted buying signals, notification at 4 of 5 or on a price, visit or sample request, a date-clash rule, and a “why we lose” list with an owner and fix for each reason. A referral network tracks who sent each lead and queues a thank-you for the referrer.
For a materials and stone supplier, contacts are flagged VIP by rule after three confirmed projects, so repeat buyers reach a senior person first.
Where people stay in the loop
- Prices, discounts, quotes and contract terms are always handled by a person.
- Scoring weights and notification rules are set and changed by the business, not by the model.
- Any message the agent cannot answer from approved information goes to a person.
- A salesperson can take over any conversation at any point; the agent stops on that thread.
What it is not
Not a black-box score
Every point on a lead traces to an answer or an action the customer took. If a score cannot be explained, it is not used.
Not an autonomous closer
The agent qualifies and books. It does not negotiate. Closing stays with people.
Not cold outreach
Follow-ups go only to people who contacted the business first. Messaging bought lists is outside the service and outside WhatsApp’s rules, which require opt-in for business-initiated messages.
Data protection
Qualifying answers include personal data and sometimes budgets. Builds follow Singapore’s Personal Data Protection Act, Indonesia’s Personal Data Protection Law (Law No. 27 of 2022) and Malaysia’s Personal Data Protection Act 2010 where they apply. The agent collects only what the five questions need, retention is agreed in the scope, and customers can ask for their record to be corrected or deleted. See /pricing for fees and /solutions for how a build runs.
Frequently asked questions
What is AI lead qualification?
Software that answers an enquiry, asks a fixed set of qualifying questions in conversation, scores the answers and the customer’s behaviour, and passes ready leads to a salesperson with a summary.
Which five questions should we ask?
They are set per business. Most use some version of: what is needed, when, where, budget range, and who decides. The test is whether a salesperson could prepare a quote from the five answers.
Will customers know they are talking to an AI?
They should. KYN recommends stating that first replies come from an assistant and offering a person on request. Hiding it costs trust when the customer finds out.
Does it work with our existing CRM?
Yes, where the CRM has an API or reliable import. Scores, answers and the handover brief are written to the lead record. A lightweight CRM is built only when the current one cannot hold the data.
How long does it take and who owns it?
About 14 days for a focused first release once scope and access are settled. Rights in the custom work are assigned to the client on full payment.
What if the scoring is wrong?
Weights are reviewed against outcomes after the first weeks: which leads closed, which did not. The business changes the weights; every change is logged with a date.