The record keeps itself
Calls, email and messages become the record automatically. Nobody types up a call in Signals.
How the record is writtenProduct, AI-native
Every recruitment CRM says it has AI now.
Almost all of them added it to software built before AI existed.
Signals put it underneath the record, so there is nothing to switch on.
AI-native recruitment software, built that way from the first line of code.
The difference is not how much AI a system has.
It is where the AI sits.
AI-native recruitment software
AI-native recruitment software
AI-native recruitment software is built with AI as the foundation rather than added to an existing system. In Signals, AI captures conversations, writes the record, tags relationships and surfaces what matters without being asked. Added-on AI sits in a panel the recruiter has to open. Native AI sits underneath the record, so the work happens whether or not anyone invokes it.
Why we built it
Recruiters never asked for AI. They asked to stop writing up their calls at seven in the evening.
Twelve years on desks across Asia Pacific taught us that the job is conversations, and that everything after a conversation is admin. Every CRM built before AI treated the record as something the recruiter maintains. That was the only option available at the time. It is not the only option now.
So we started underneath. One place where the conversation lands, the record writes itself and the desk stays current, described in one workspace.
The recruiter should never be the system of record.
The distinction
A candidate calls you about a role.
Added on
- Take the call.
- Open the CRM.
- Find the candidate.
- Create an activity.
- Type the notes.
- Tag the role.
- Update the stage.
- Set a reminder.
- Open the AI panel and ask it to summarise what you just typed.
Nine actions
Native
- Take the call.
One action
The AI did not do the admin faster. The admin never existed.
What native looks like
What native looks like
01 The record keeps itself
02 It knows your desk
03 You ask instead of search
04 It tells you what to do next
It knows your desk
Companies, contacts, jobs and history are connected, so context arrives with the record rather than being searched for.
How the desk connectsYou ask instead of search
Ask Signals in plain English and get an answer from your own data, not a list of results to sift.
How asking worksIt tells you what to do next
The system surfaces the next action on a desk rather than waiting to be interrogated.
How next actions surfaceA day on the desk
A day on the desk
Nine
Client call about a new role
No write up afterwards.
Ten thirty
Three candidate calls
No activity logging.
Twelve
A hiring manager mentions a second headcount in passing
No note taken, and it is on the record anyway.
Two
Shortlist goes out
No manual status updates.
Four
A client replies from their phone
No copying it into the system.
Six
Laptop closes
Nothing owed to the CRM.
Nothing was left for later, because nothing was left.
Limits
What it will not do.
Systems that claim to do everything are asking you to check everything.
It will not invent what it has not heard.
If a detail was never said, it is not on the record. Gaps stay gaps.
It will not act as you without you.
Nothing sends, submits or replies in your name until you say so.
It will not reconcile your numbers.
Billings and fees come from your finance system, not from an inference. Signals shows you what it can evidence.
Three fewer things to check.
Side by side
Where the AI sits changes everything downstream of it.
| Behaviour | AI added on | AI-native |
|---|---|---|
| Where the AI sits | On top of the record | Underneath it |
| Who maintains the record | The recruiter | The system |
| Getting an answer | Search and filter | Ask in plain English |
| Context on a record | Whatever was typed in | Everything that was said |
| When AI runs | When invoked | Continuously |
| What happens if nobody uses the AI | Nothing changes | The record is still complete |
Common questions
AI-native means the AI is part of the foundation rather than a layer on top. In Signals the record is written as conversations happen, so there is no panel to open and no prompt to write for the basics. Software built before AI existed can add AI to a screen, but the record underneath still waits for a recruiter to maintain it.
Added-on AI is invoked. A recruiter finishes a call, types the notes, then asks the AI to tidy them. The typing is still the recruiter's job. Native AI runs before the recruiter reaches the screen, so the call, the email and the message become the record on their own. The difference shows up in how many actions one conversation costs.
Yes. There is nothing to opt into for the record itself. Calls, email and messages become the record whether or not anyone asks for a summary. If your team never opens an assistant, the desk is still documented and the pipeline still reflects what happened. The optional part is asking questions of your data, not keeping it.
From your own communications and your own workspace. Calls, email, messages and calendar activity that your team already runs through Signals, plus what recruiters enter directly. Nothing is collected from outside those sources. Signals does not build records from the open web and it does not fill gaps with data bought from elsewhere.
Yes, and this matters. Every inferred detail shows where it came from, the conversation and the moment inside it. Open it, read the source, then change or delete it in place. A correction replaces the inference and is kept as the recruiter's own entry. Nothing inferred is hidden behind a summary you cannot trace back.
It can draft, it will not send. Signals prepares a reply with the context of the record behind it and it stays a draft in your hands until you send it. Nothing goes out in your name without your action. For recruiters who prefer to write every word themselves, drafting can stay switched off.