AI Email Assistant for New Zealand Sales Teams

Work out which of these you are actually shopping for before you compare anything, because a demo of one will not tell you whether it fixes the other.

Drafting help. You are writing one message to one person, and the software writes a first version or rewrites yours. The problem it solves is typing time on messages you already know how to write. It does nothing about volume, timing or whether the message gets sent at all.

Sequence automation. You have a planned series of messages that should go to many people on a schedule, and stop when someone responds. The problem it solves is that day four of a five-touch follow-up never happens when it depends on a human remembering. It does not improve your writing, and a badly written sequence simply annoys more people faster.

Inbox triage. Too much arrives and you cannot tell what needs answering. The software sorts, summarises and surfaces. The problem it solves is attention, not output, and it is genuinely a different category of product from the first two.

Most New Zealand sales teams of under twenty people have the first two problems and not the third. The third arrives with volume, and volume usually arrives after the first two are solved. It is worth saying plainly that HelloGrowthCRM covers drafting and sequences, and does not do inbox triage.

Drafting help pays off on messages where the thinking is already done and only the typing remains. That is a bigger share of sales email than most people expect.

The follow-up after a call is the clearest case. You have just spent twenty minutes on the phone. You know what was agreed, what the objection was and what happens next. Writing it down is pure transcription, and an assistant working from your notes does it in a fraction of the time, in a consistent structure, with the next step stated. The same applies to the reply you send to the three questions that come in every week, the covering note on a proposal, the confirmation after a booked meeting, and the chase on a quote that has gone quiet.

There is a second, less obvious win: shortening. AI is more reliable at cutting a rambling four-paragraph message to four sentences than at inventing four sentences you did not have. If you write long, use the assistant as an editor rather than an author and you will get more value from it than most people do.

Where it stops earning its keep is anywhere the message depends on judgement about a specific relationship. A negotiation, an apology after something went wrong, a message to a customer you have known for eight years. In those cases a draft costs you more in editing than it saves in typing, and the risk of a wrong note is real.

Everyone has now received the email. It hopes this finds you well. It admires the recipient's impressive work in their sector. It suggests that many businesses in that industry face challenges around efficiency. It names nothing, and it was clearly not written for the person reading it.

That email is not bad because a machine wrote it. It is bad because it contains no information, and it would have read the same way if a person had written it in ninety seconds. The failure is asking a drafting tool to supply substance rather than to arrange substance you gave it. Fed a contact name, a company and an industry, that is all it can produce.

• Feed it your notes, not the record. The objection raised, the timeline mentioned, the approver named, the job number quoted. Specifics are the difference and they can only come from you.
• Write the first line yourself. That is where a generated message announces itself. Referring to something that genuinely happened between you removes the tell immediately.
• Delete every sentence that could be sent to anyone. If a line survives being addressed to a completely different company, it is doing no work and should go.
• Never send an unchecked fact. Details about the customer's own business are exactly what a draft will invent with total confidence, and being wrong about someone's own company is expensive.
• Turn winners into templates. When a draft lands well, save it. A proven template beats a fresh generation, is faster, and keeps a team's messages consistent.

The last point matters more than it sounds. A drafting assistant used well ends up producing fewer generated emails over time, not more, because the good ones become templates and the templates get reused.

A sequence is worth it when the same planned series applies to many people and timing matters more than wording. Onboarding a new customer through their first month. A structured follow-up run after a trade show or an exhibition. A re-engagement pass over leads that went quiet six months ago. In each case the messages are broadly the same for everyone and the value is in day four happening at all.

A one-off send is right when the content depends on what this particular person said. Answering a specific question, responding to an objection, confirming a change to a quote. Most teams get this backwards: they run sequences for prospects who have already had a real conversation, and send one-off emails for onboarding, where a sequence would be far more reliable.

Two things separate a sequence that works from one that irritates. The first is that it must be built on a real segment. A list of everyone who ever gave you an email address is not a segment, and a message written for that list can only be generic. The second is exits, which are important enough to deserve their own section. If you want the sends triggered by what happens in the pipeline rather than run as standalone campaigns, that is the territory covered by sales automation software.

This is the single most common way automated email damages a business, and it is entirely avoidable.

The scenario runs like this. A prospect gets touch two of a five-touch run and replies with a genuine question. A salesperson answers it, and a good conversation starts. Two days later, touch three fires: a scheduled message asking whether they had a chance to look at the last email. The customer now knows they are on a list, that nobody joined up the conversation, and that the earlier message was not really written for them. Everything gained in the reply is spent.

Set the exit rules before you write a single line of copy. A reply from the contact should stop the run. So should a booked meeting, a deal moving stage, a quote being accepted, an unsubscribe, and in most cases a reply from anyone else at the same company. It is also worth having a rule that stops a contact receiving two different sequences at once, which is how someone ends up with three of your emails in one morning.

Then test it. Put yourself on the sequence, reply halfway through, and confirm nothing else arrives. Do this every time you change a sequence, because exit conditions are the part that silently breaks when someone edits a workflow.

One more thing belongs in the same conversation. Sending commercial email to New Zealand recipients carries obligations on you as the sender, covered in the questions below: consent to be contacted, clear identification of who is sending, and a working unsubscribe path that is actually honoured. Build consent capture into how leads arrive, and make sure an unsubscribe removes someone from every run rather than only the one they clicked from. Where those leads come from and how consent is recorded is covered on our lead management page.

The honest answer varies by job. Here are the five that make up most of a sales team's email, and what each one deserves.

The pattern across the rows is consistent. Automate where timing decides the outcome, draft where the content is repetitive, and write it yourself where the relationship or the money is large enough that the hour is obviously worth spending.

HelloGrowthCRM covers two of the three jobs. Drafting, through an AI writing assistant that works from the record and your own notes. Sending, through Campaigns, plus bulk WhatsApp broadcasts and SMS campaigns for the customers who never open email. It does not do inbox triage, and if that is your real problem you should buy something built for it.

What makes the drafting and sending worth having in a CRM rather than a standalone email tool is context and consequence. The draft is written next to the calls, notes and quote for that customer, so it can be specific. The send is recorded against the contact and the deal, so the next person can see it. Workflows create the follow-up task, custom fields let a message use real details from the record, and real-time dashboards and custom reports show what happened to the deals a campaign touched rather than only what was sent. Unlimited contacts mean segmentation is a decision about relevance rather than about plan limits. If you want the calling half of the same follow-up problem, see our dialer and call tracking page, and for how the AI features fit together across the product, the AI CRM overview.

Pricing is NZ$17/user/mo on annual billing or NZ$22/user/mo billed monthly, with no seat minimums. A free plan is available with no credit card, and paid plans include a 14-day free trial. The New Zealand pricing page lists what each plan includes and how GST is treated.

A sensible first test is narrow. Take the follow-up email you send after every sales call, draft the next ten with the assistant during a free trial, and see how many you send with only small edits. If they land well, that is your first template. Teams setting up their first proper system usually start from our small business CRM page, or you can book a demo and we will build a sequence against your own pipeline.