A quote goes out on Tuesday. The customer asks one question on Thursday. By Monday, that reply is buried under supplier emails. Nobody has deliberately ignored the customer; the next step simply has no place to live.
This is a useful problem to bring to AI sales. You already have interest in your business. What you need is a reliable way to see who is waiting, what they need and who should respond.
What is AI sales, and where does it fit?
AI sales means using artificial intelligence to help with sales work, including reading enquiries, summarising conversations and drafting follow-up messages. Combined with customer records and reminders, it can support the journey from first contact to a decision.
Useful tasks include summarising meetings, preparing follow-up drafts and showing which opportunities need attention. The tools and connections available in your own setup determine what AI can actually do.
At Green Kauri, AI Sales focuses on organising leads, keeping follow-ups visible and preparing the next action. A lead is simply a person or business that may become a customer. Begin with existing enquiries so you can compare the proposed process with work you already understand.
1. How do you keep every enquiry in view?
Give each enquiry one record, one owner and one next action. AI can help extract details from messages, but someone still needs to check that the right customer and request are attached to the record.
Choose one place for your team to look each morning. That might be a shared list or a CRM, which means customer relationship management software. The important part is that an enquiry does not depend on someone remembering an unread email.
- Customer: name and the contact details needed to reply.
- Request: what they want, with a link to the original message.
- Status: new, waiting for details, quote sent, won or lost.
- Owner: the person responsible for moving it forward.
- Next action and date: a specific task and when it is due.
For a renovation enquiry, “Call Sarah on Tuesday to confirm the site visit” is useful. “Follow up soon” leaves the same uncertainty you started with. Check for duplicate enquiries before creating another record for the same job.
2. What should the first reply achieve?
The first reply should acknowledge the request, answer what you can and explain the next step. Use AI to prepare it from approved business information and the customer’s actual message.
Suppose Sarah asks for a kitchen renovation quote and gives her suburb and preferred start month. A useful reply asks about the missing detail needed for a site visit. It should not ask her to repeat everything she has already written.
Example draft: “Hi Sarah, thanks for your kitchen renovation enquiry. You mentioned October as your preferred start month. Could you send a couple of photos of the current kitchen? Alex will then contact you to arrange a suitable time to discuss the work.”
Only use that wording if Alex owns the next step and photos are part of your process. Keep prices, availability and promised response times tied to information your team has confirmed. When a fact is missing, ask a question or leave the draft for review.
3. How do you follow up without pestering people?
Base follow-up on the customer’s situation and stop scheduled messages when the conversation changes. A reminder should help someone make a decision or raise a question.
When you send a quote, ask when the customer expects to review it. Record that date. For a business customer waiting for a monthly approval meeting, the next useful contact may be after that meeting, rather than tomorrow morning.
If no date is agreed, you could test one short check-in after a few working days. This is a starting rule to review with your team, not a universal best interval. Give the message a clear purpose:
Example draft: “Hi James, I’m checking whether you have any questions about the quote we sent on Tuesday. If your timing has changed, let us know and we can update our notes.”
Before sending, check for a recent reply, an accepted quote or a request to stop. If your system cannot reliably detect these, keep messages as drafts. Set a limit on unanswered reminders and assign any further contact to a person.
4. Which sales decisions should stay with a person?
Keep decisions about discounts, unusual promises, complaints and unclear requirements with your team. AI can prepare the context so the person taking over does not have to start from the beginning.
A useful handover includes the customer’s request, the latest quote, what has already been promised and the question that needs a decision. Include the original messages so the reviewer can check the summary.
For example, a commercial cleaning customer may ask to add weekend work to a proposal. The system could flag the changed requirement and draft questions about access and frequency. A person should confirm staffing, scope and price before the customer receives a commitment.
Make the handover visible in the daily list. Name the owner and show why the item needs attention. Otherwise, an automated process can still leave a customer waiting at exactly the point where a thoughtful reply matters most.
5. How do you know whether the process helps?
Compare follow-up reliability and useful customer outcomes before and after a small trial. Counting messages sent will not tell you whether anyone received better help.
Start with one enquiry source or one type of quote. Record a baseline using recent work, then run the new process for a few weeks. Keep the trial long enough to include your normal customer decision period.
- Waiting enquiries: how many passed your chosen response target?
- Missed next actions: how many tasks were overdue?
- Useful replies: how many customers supplied details or agreed a next step?
- Quote outcomes: how many were accepted, declined or still undecided?
- Team effort: how much time went into checking drafts and fixing errors?
Compare similar work and note changes in enquiry volume, pricing or staffing. Ten quotes are a small sample, so one extra sale is not proof that AI caused an improvement. Look for a repeatable pattern and check the conversations behind the numbers.
What should you prepare before getting started?
Prepare a few anonymised examples, your usual sales stages and the rules for who handles each next step. You do not need to map every possible situation before testing one clear task.
Choose a new enquiry, a waiting quote and a conversation that needed judgement. Write what should have happened in each case. These examples help you assess whether a proposed setup fits your business.
If the problem is that visitors cannot understand your offer, start with your website and service pages. If too few relevant people find those pages, xEO optimisation addresses visibility in search and AI answers. Sales follow-up is most useful once those enquiries arrive.
A sensible first goal is simple: open one list in the morning and know which customers need attention today. Once that works reliably, you can decide whether another part of the process needs help.
What else should you know about AI sales follow-up?
What does AI sales do for a small business?
AI sales helps a small business organise enquiries, summarise customer needs, prepare replies and track the next action on open quotes. The useful starting point is one repeated task, such as reviewing unanswered quotes each morning. People still decide pricing, exceptions and how to handle important customer conversations.
How often should AI follow up a quote?
Set follow-up timing around the customer’s stated plans and your usual sales cycle. Agree a next contact date when you send the quote where possible. If you test a reminder after several working days, treat that as a starting rule. Pause reminders whenever the customer replies or asks to stop.
Do I need a CRM before I use AI sales?
You do not always need a customer relationship management system before starting. A shared inbox and a tidy spreadsheet may be enough to define the process. You do need reliable customer records, an owner for each enquiry and a way to record replies before any automated follow-up is sent.
Will AI sales close more deals automatically?
AI sales cannot guarantee more closed deals. It can help your team notice waiting enquiries and prepare the next step, but customers still decide based on fit, price and trust. Measure missed follow-ups, useful replies and accepted quotes, alongside the time your team spends reviewing and correcting the system.