◆For decision-makers

Three, which belong together in small and medium-sized companies:

• Knowledge that exists only in people’s heads and mailboxes. Who spoke to whom about what and when is gone after a holiday or when someone leaves.
• Customer acquisition that does not happen, because it sits alongside day-to-day business and nobody follows up.
• An outside image that goes out of date, because every change costs an invoice.

Aurora Hive Sales addresses the first two, Aurora Hive Sites and Live the third.

Two kinds, and they behave very differently:

• The licence — predictable, based on scope and type of operation. What applies to you is in your quote; this help deliberately contains no figures, because they would be wrong on the day of the first adjustment.
• Usage — AI evaluations and paid research. They only arise when someone triggers them, are shown with an upper limit before the click, and can be budgeted per instance.

What does not arise: setup fees for every change, licences per device, maintenance flat rates for small things the customer can change themselves.

The honest answer: to the extent that a company becomes dependent on any system it works with — with three precautions that defuse this point:

• Your data remains accessible. Contacts, companies and vacancies as CSV, reports and contracts as PDF, media as files. Signing in, information requests and export always remain open, even after a licence has expired.
• The domain belongs to you — make sure it is registered to your company and not to a service provider.
• No copy protection that locks you out. An expiry blocks sending, never signing in.

What does not move with you: the structure of the pages built from blocks. A move means reassembling them in the target system.

The measures that carry weight in operation:

• Separate instances per customer — own database, own environment. Two instances cannot see each other.
• Servers in Germany; the exact details are in your data processing agreement (AVV).
• Encrypted transmission, mailbox credentials stored encrypted, sign-in via the provider instead of passing on passwords.
• Access log covering all changing operations.
• Daily backups including a restore test.

What the system cannot do is down to you: individual passwords, deactivated accounts of people who have left, care when exporting.

Four things, three of which the system prepares:

1. Data processing agreement with the operator — a draft is generated from your master data, but it should be checked.
2. Legal notice (Impressum) and privacy notice that match what is actually switched on — the list of services is derived from the configuration, not from a template.
3. Record of processing activities — you already have most of the details for it.
4. A deletion policy — you have to define that yourself.

Point 4 is the one software cannot take off your hands, and it is the first thing asked about in an audit.

Yes, at three levels:

• Not at all: modules that use AI are not activated — then they do not appear in the menu.
• Limited: a budget per instance. Once it is reached, chargeable runs are refused.
• Controlled: in the workflows, every AI suggestion remains a suggestion and only takes effect after human approval. This cannot be switched off.

Two places are explicitly built without AI, even though it would be the obvious choice: column detection during data import and industry detection in research. With contact data, that is not a detail.

For the core of Aurora Hive Sales: one week alongside normal work, not a project.

• Day 1: mailbox, settings.
• Day 2: import and categorise addresses.
• Day 3: commission the DNS records — that takes time externally.
• Day 4: templates.
• Day 5: first small campaign.

What takes effort is not the software but the groundwork: which categories do we use? How do we write company names? Who maintains the data? These three questions cost ten minutes at the start and days later on.

Not in the sense of a training day. The explanation is where the question arises: the ⓘ on the field, the tour in the area, the demo film by the title — and this help centre with over a thousand answers.

What is worthwhile is half an hour as a team before the start: categories, spellings, responsibilities. That is not training but an agreement, and it decides more about the benefit than any induction.

Not with arguments about the system, but with what gets better for them. Three points that work in practice:

• Less searching. The history is on the contact, not in five mailboxes.
• Nothing forgotten any more. Follow-ups remind you; notes do not.
• Cover becomes possible. Someone on holiday no longer holds up the matter.

And one point that honestly belongs here: a CRM makes work visible. Anyone who feels that as control is not entirely wrong — say what you use it for, and stick to it.

There are two questions behind this, and both need answering:

• Will we notice? The system status checks from the inside and in principle cannot report a total failure — that requires an observer that does not sit on the same server. On the vendor side, there is “Check the instance” for this.
• Can we get back to our data? Daily backups of the database, environment files and uploaded files are running. What matters is not that backups are made but that they have been restored — ask about the last restore test.

In an emergency: report early, do not keep working. A restore resets the data to an earlier state.

The programs are developed and run in-house — the same software that is sold runs our own operations. That is not a reference in the usual sense, but a reliable one: whatever does not work, we notice first.

For specific references, ask your contact person. What this help offers instead is verifiability: every process described here appears exactly like this in the program — right down to the button labels.

Yes — 14 days, with a complete instance of your own. Not a demo account with features switched off: the same software, under your own address, with your own data.

Three days before the end, you get a reminder. After that, a human decides what happens — nothing is deleted automatically and nothing is taken down automatically. If you take it on, the same instance carries on with the same data; you do not start from scratch.

Use the time for a real run-through, not a click-through tour: a small address list, a real campaign, a few replies.

Aurora Hive Sales is operated by us, not on your premises. Each customer gets their own installation with its own database and its own address, operated in a data centre in Germany.

What this saves you: operation, backups, updates, certificates and the question of who can be reached at night.

For Aurora Hive Sites, there is a separate route for companies that want to take over operation themselves. That is a separate item with a conversation beforehand — speak to your contact person.

Aurora Hive Live runs on your customer’s web space anyway; there is no server of ours involved there.

Exactly where your data is stored and who works on it is set out in your data processing agreement.

First check whether it really is special: the most common special request is an arrangement that should be solved differently anyway.

If it remains, the way forward is a problem report — and it works best when it describes the problem rather than the solution: “We need an overview of open items per customer” gets further than “Please add a column X”.

What is not possible and will not be coming is stated openly in this help: no shop, no free design outside the blocks, no syncing of contacts and form responses between Aurora Hive Sales and Sites.

Four things that experience shows decide between success and a forgotten system:

1. One person responsible for data maintenance. Not “the team”.
2. Commission the three DNS records early. They take time externally and decide whether you land in the spam folder.
3. Start small. Fifty recipients in the first campaign, not five hundred.
4. Set budgets before the first big research run starts — not afterwards.

And one rule that outweighs everything: run only one system. Two versions of the truth side by side are worse than one incomplete one.