AI follow-up

Message sequences for those who didn't respond. Second and third reminder go out on their own, at set time and with stop condition.

• individuallyScope
• AI and automationsArea
• 6 hours – 2 daysLaunch

Who this is for

Services with a longer decision process.

Problem

No response ends the contact, though a reminder would suffice. This hits hardest services with longer decision cycles: someone asked for an offer, got it and went silent because they had other things on their mind. No one manually comes back to them because nobody maintains such lists, and the impulse "I need to write to them" isn't a process.

What we do

We set up message sequences for leads that didn't respond. If an AI assistant enters the conversation, it introduces itself as AI — it doesn't sign with any team member's name.

• The sequence starts automatically when a submission gets no response
• Time, channel and number of touches are set in advance, not dependent on someone's day
• Stop condition is part of the sequence: response or request to stop breaks sending immediately

What's included

Three elements; the middle one decides if the reminder is helpful or annoying:

• Sequences for leads with no response
• Rules for time, channel, exit and handoff to human
• Automatic CRM launch and response measurement

How long it takes

Launch takes 6 hours – 2 days. Work divides into scenarios, content, automation and measurement. How many silent submissions actually return is known only after full sequence passage — that's why response measurement is included from the start, not added later.

How much it costs

Two numbers decide: how many different contact scenarios have to be written and how many channels they have to run through — e-mail alone is different work from e-mail, SMS and a messenger at once. This is not a one-off launch: the sequences have to be corrected by what the replies show, and so they work for as long as AURA runs this area for you.

What you get

Recovered leads without manual work. The conversation doesn't end with one unanswered message, and the team doesn't need to remember who to follow up with on Thursday. You also see which reminder actually recovers contact, because each sequence is measured separately — so instead of adding more messages, you can turn off those that don't yield anything.

How to calculate the value of one follow-up sequence

The sequence only works on inquiries that already failed to respond once, so its value is calculated by how many of them come back:

• Value of a week of sequences = number of unanswered inquiries per week × share that returns after a reminder (calculated on those who received at least one reminder) × average value of one recovered conversation

You get the middle share from your own data after the first full run of the sequence — before that there's nothing to calculate it from, because nobody has received a reminder yet.

How an automatic sequence differs from a manual reminder and a mass mailing to the whole base

Four ways of reaching someone who didn't respond produce very different results:

• Personal memory — someone on the team was supposed to follow up in a week, got busy with something else, and never did
• One manual email — sent whenever someone remembers, so the time between the inquiry and the reminder is random
• Mass mailing to the whole base — also reaches people who already bought or already replied, so it looks like a mistake
• Automatic sequence — starts at a set point after no response and stops itself the moment someone replies

The difference isn't in the number of messages sent, but in whether they reach the right person at the right moment and stop on time.

Where the sequence uses a language model, and where it's schedule and rules

This distinction determines how much real artificial intelligence is actually in the sequence.

• The language model, if it takes part in the exchange, understands the lead's reply and formulates the next message in context — the only part that can honestly be called artificial intelligence
• The sending schedule, number of attempts, time intervals and channel are rules set in advance together with you, not a decision made anew each time
• The stop condition — a reply or a request to be left alone — is a simple trigger in the system, not an assessment of the situation

It's exactly the rules from the second and third point that decide whether the sequence feels like a helpful reminder or intrusive spam.

When a reminder sequence doesn't pay off

There are situations where an automatic reminder does more harm than good.

• There are only a few unanswered inquiries a week — the team has time to follow up on each one manually anyway
• The decision process takes hours, not days — a reminder sent after a day arrives once the decision is already made
• Every message needs to be unique and fully individual because the contact is very personal — automation here feels like disrespect, not help
• There's no legal basis or consent for automatic contact through that channel — without it, the sequence shouldn't run, no matter how much it might recover
• The contact base isn't kept up to date and some addresses bounce as non-existent — the sequence then fires into a void, and response metrics just hide that fact

What you need to prepare before the start

The sequence starts on your data and your rules, not a generic template:

• Access to the CRM with the list of inquiries left unanswered
• Time and attempt-count rules agreed with the team — how many reminders, at what intervals, how many is too many
• Base message texts for further personalization, written in your own voice, not a generic template
• A decision-maker on the tone of communication and the stop condition, before the first message goes out to anyone

How to check after a month whether the sequence paid off

You calculate the metrics from your own CRM, separately for each stage of the sequence:

• Response share after the first, second and third reminder separately — shows which one actually works
• Number of recovered leads that without the sequence would simply have been written off as dead
• Number of requests to stop contact — a rising number signals too many reminders, not a stronger-working sequence
• Comparison of the recovered-conversation share with the period before implementation, for the same month of the year

What a reminder sequence doesn't do

It reminds and asks whether the topic is still relevant, but it doesn't negotiate price or terms, doesn't lower the offer on its own, and doesn't send another message once someone has asked to be left alone — that stops the sequence immediately, no exceptions. When a lead replies and the conversation comes back to life, a human takes over — the sequence only reopens the door, it doesn't carry the conversation forward on its own.

This mechanism isn't optional: a sequence without the ability to hand the conversation to a human stops being a reminder and turns into a persistent message sent into a void.

The most common mistake when implementing a sequence, and whose data stays with you

The most expensive mistake isn't a poorly written first message, it's too many reminders or no stop condition at all — a lead who was meant to come back instead feels cornered and leaves for good instead of just postponing the decision. The second mistake is sending the same sequence regardless of channel and stage of the conversation, which then sounds like a machine rather than a reminder from someone who remembers the context.

The sequence history and reply statuses are stored in your CRM, right alongside the rest of the lead's conversation history — not in a separate database we'd have exclusive access to.

What you get

• Sequences for leads that have not responded
• Timing, channel, exit and human-handoff rules
• CRM-triggered automation and response measurement

When the problem lies elsewhere

If your problem sounds different, neighbouring areas of the same system stand right next to it — AURA connects them to each other rather than selling them separately:

• CRM and automations — Requests, reminders and client data work on their own, without re-entry from paper to calendar.
• AI Reports — Weekly summary of leads, campaigns and sales, written in simple language and sent on its own. Instead of a dashboard that nobody checks.
• Integrations — We connect what you already use: POS, calendar, spreadsheet, warehouse, booking system. Instead of replacing with new.
• Automatic messages — Confirmations, reminders and thank-yous go out on their own — by SMS, email or through WhatsApp.