The agents inside the platform
Agents that work the hours a team cannot.
85 agents read your replies, write your follow-ups, label your contracts, mine your calls and keep testing your campaigns. Last month 72% of their work happened outside the working day.
The roster · all 85 on shift
live
Measured, not claimed
The busiest hour is nine at night.
Every agent action in one workspace over thirty days, plotted by the hour it happened. The working day is the quiet part. The agents pick up when everyone goes home, and they don’t stop for the weekend.
- Busiest hour
- 21:00
- Between midnight and 6am
- 21%
- Outside 9 to 6
- 72%
- Actions in 30 days
- 2,906
The shape of the system
Every agent runs the same five steps.
It notices, reads, decides, acts and writes down what it did. No step waits on you. Learn it once and every new agent is already familiar.
- 1Something happens
a reply, a form, a signature
- 2It reads
the record and the history
- 3It decides
and writes down why
- 4It acts
sends, books, updates, files
- 5It logs
so you can check it later
Then again, for the next thing that happens.
Replies and follow-up
A reply at 2am. A booked call six minutes later.
The gap between someone raising their hand and hearing back is the whole game. Nobody was awake for this one.
- 102:11Reply lands
from a prospect
- 202:11Read for intent
interested
- 302:12Answer written
in your voice
- 402:12Checked
graded before it goes
- 502:13Sent
from the right mailbox
- 602:17Call booked
on the calendar
The full roster
85 agents, in 13 teams.
Each one has a job, a trigger that wakes it up, and a record of everything it did. Jump to a team, or read it all.
The agent you talk to 12 agents
One assistant on every screen. It knows the record you are looking at, so you never paste context.
-
Agent chat
The assistant built into every screen. Ask it to build something and it builds it, right where you are.
Wakes on you ask
-
Chat connector
Lets your other tools talk to the agent through the connection your apps already use.
Wakes on an outside tool calls
-
Channel operator
Sets up and looks after the phone, email, and messaging lines your agents send through.
Wakes on you connect a line
-
CRM skill
Teaches the agent to work your CRM. Find a contact, move a deal, log a call, update a record.
Wakes on the agent needs it
-
Tables skill
Teaches the agent to build and change your tables. Ask for a new field or view and it makes it.
Wakes on the agent needs it
-
Automations skill
Teaches the agent to write automations for you, including the conditions and every step.
Wakes on the agent needs it
-
Websites skill
Teaches the agent to build and edit your public pages and booking sites.
Wakes on the agent needs it
-
Operations skill
Teaches the agent to handle setup and admin work across the whole system.
Wakes on the agent needs it
-
Agent Store skill
Lets the agent install other agents. You describe the job, it picks the worker and switches it on.
Wakes on you describe a job
-
Control plane
The switchboard behind every action. It checks permissions, tracks spend, and records what happened.
Wakes on any agent acts
-
Tool surface
The safe list of tools an agent may use. Nothing outside the list can be touched.
Wakes on any agent acts
-
Agent builder
A guided setup for your own agent: the trigger, the steps, and what it may do alone.
Wakes on you ask
Your tables and records 6 agents
Your data is not a spreadsheet that sits there. Columns can think, and everything is searchable by meaning.
-
AI columns
A column that fills itself in. Write the instruction once and every row answers it.
Wakes on a row changes
-
Bulk prompt
Ask one question of hundreds of records at once and write the answer back to every one.
Wakes on you ask
-
Meaning search
Search by what you mean. “Unhappy about pricing” finds the record that says “the cost is too high”.
Wakes on you search
-
Record indexing
Quietly reads every new record so search stays current without anyone running anything.
Wakes on a record is saved
-
Table discovery
Reads your workspace and works out which table holds what, so the agent can answer across all of them.
Wakes on the agent needs it
-
Offer workbench
A working surface for building an offer and the market it goes to, with the agent drafting.
Wakes on you ask
Automations 7 agents
Describe the job in a sentence. See exactly what would happen before anything runs. Then switch it on.
-
Automation copilot
Say “text me when a lead goes cold for five days” and it writes the automation. No trigger menus.
Wakes on you ask
-
Dry run
Shows which records would be touched and what would be sent, before a single message goes out.
Wakes on you ask
-
Plain English readback
Reads any automation back to you in normal words, so you can confirm it does what you think.
Wakes on you open one
-
Automation runner
The worker that carries out every automation, including the steps that need writing.
Wakes on a record changes
-
Speed to lead
A new lead arrives, gets qualified, and gets a first message out within seconds.
Wakes on a new lead
-
Database reactivation
Goes through old leads and writes a personal reason to reach back out. Two touches maximum.
Wakes on you start a push
-
Missed call text back
You miss a call, they get a text immediately. When they reply, it drafts your answer.
Wakes on a call is missed
Contracts and signatures 3 agents
Send a contract without dragging boxes onto a page. The agent reads the document and works out where the fields go.
-
Field finder
Reads the contract and finds every blank that needs filling, without you clicking anything.
Wakes on a contract lands
-
Field labeller
Names each blank and works out which party fills it, so a signer only sees their own fields.
Wakes on blanks are found
-
Signature triggers
A signed contract is an event. Onboarding starts, the invoice goes out, the deal moves.
Wakes on someone signs
Replies and follow-up 8 agents
Every reply gets read, sorted, and answered. Interested people hear back in minutes instead of the next morning.
-
Reply reader
Reads every reply and works out what the person actually wants.
Wakes on a reply lands
-
Intent sorter
Turns that reading into a decision the system can act on, with no inbox triage by hand.
Wakes on a reply is read
-
Second opinion
Goes back over replies it was unsure about and grades them again, so nothing good is lost.
Wakes on every few minutes
-
Reply writer
Writes the answer in your voice, using what you already know about that person.
Wakes on intent is known
-
Sender
Sends the reply from the right mailbox, at a sensible hour in the recipient's timezone.
Wakes on an answer is ready
-
Booking handoff
When someone says yes, it moves them straight to your calendar instead of more email.
Wakes on they say yes
-
Review queue
The exception, not the rule. Only the calls it is genuinely unsure about come to you, with the reasoning shown.
Wakes on confidence is low
-
Reply intake
Catches replies the moment they land in your sending tool, so nothing waits for a sync.
Wakes on the inbox pings
Calls and meetings 5 agents
Every call becomes something you can use later, not a recording nobody opens.
-
Call intelligence
Listens to the recording and pulls out what was agreed, what was objected to, and what is next.
Wakes on a call ends
-
Call playbook writer
Builds a prep sheet before the call: who they are, what they care about, what they will push back on.
Wakes on a meeting is set
-
Playbook delivery
Puts that prep sheet in front of the rep at the right moment, on the right meeting.
Wakes on the call is near
-
Meeting sorter
Works out what kind of meeting each one was, so your reporting is not one undifferentiated pile.
Wakes on a meeting is filed
-
Meeting intake
Picks up recordings and transcripts automatically and files them against the right contact.
Wakes on a recording lands
Playbooks and knowledge 7 agents
Your process is already in your calls. These agents write it down, and every step points back to the moment it came from.
-
Process miner
Watches how your team actually works and writes the procedure from real calls, not from memory.
Wakes on nightly
-
Draft assistant
Turns rough notes into a clean procedure someone new could actually follow.
Wakes on you ask
-
Source finder
Looks across recordings and documents for where a process was explained best.
Wakes on you ask
-
Procedure search
Finds the right procedure by describing the situation, not by remembering the title.
Wakes on you search
-
Fact extractor
Pulls the people, companies, prices, and commitments out of conversations and stores them as facts.
Wakes on a conversation ends
-
Knowledge indexing
Keeps everything the company knows searchable by meaning as it arrives.
Wakes on anything is captured
-
Duplicate resolver
Works out that the same company written four ways is one company, and merges it.
Wakes on a new name appears
Research 6 agents
Point the system at a market. It comes back with who to sell to, what to sell them, and why they would buy.
-
Market researcher
Researches a market properly: who is in it, what they struggle with, and where the opening is.
Wakes on you name a market
-
Research scheduler
Keeps research running in the background so the picture stays current instead of going stale.
Wakes on nightly
-
Session runner
Manages a full research run from the first question to the finished report.
Wakes on a run starts
-
Strategy builder
Turns an intake conversation into a written plan for what to sell and who to sell it to.
Wakes on intake completes
-
Offer builder
Writes the offer itself, with starter, pro, and premium versions and a reason each one exists.
Wakes on a niche is chosen
-
Inbox triage
Reads incoming applications and enquiries and sorts the real ones from the noise before you look.
Wakes on every few minutes
Finding the right customers 11 agents
The most valuable work happens before a message is sent. These agents read the actual companies rather than trusting a filter.
-
Audience generator
Proposes the specific groups of companies worth going after, with the reasoning for each.
Wakes on a niche is chosen
-
Audience discovery
Goes looking for groups you have not thought of, based on who already buys from you.
Wakes on you start a search
-
Audience checker
Reads the actual companies on your list to confirm they are what the filter claimed.
Wakes on a list is built
-
Targeting checker
Tests your filters against reality and says when they are too broad or too narrow to work.
Wakes on a filter changes
-
Offer fit checker
Confirms the offer matches the audience, before you spend money finding out it does not.
Wakes on a list is built
-
Automatic validation
Runs those checks on their own, so a bad list is caught early instead of after the send.
Wakes on every few minutes
-
Drift watch
Watches whether the checking itself is still accurate over time, and flags it when standards slip.
Wakes on nightly
-
Company resolver
Turns a messy company name into the real business, with site, size, and location confirmed.
Wakes on a company appears
-
List preview
Shows the real companies and people a filter would produce, before you commit to it.
Wakes on you ask
-
Lead sourcing
Pulls the actual contacts once a list is cleared, and keeps pulling on a schedule.
Wakes on daily
-
Targeting approval
An optional stop. Turn it on for a niche you want eyes on, leave it off and the list ships itself.
Wakes on a list is ready
Campaigns that improve themselves 11 agents
Writing the campaign is the easy part. These agents keep testing it and quietly retire what is not working.
-
Sequence writer
Writes the full email sequence for a campaign, matched to the audience and the offer.
Wakes on a campaign starts
-
Sequence generator
Produces several versions to test against each other rather than betting on one angle.
Wakes on a campaign starts
-
Campaign drafter
Assembles the whole campaign, list and copy and settings, ready for you to review.
Wakes on you ask
-
Strategy presets
Starting points that already reflect what has worked, so nobody begins from a blank page.
Wakes on you ask
-
Weekly research
Refreshes what it knows about the market every week so the messaging does not go stale.
Wakes on weekly
-
Optimizer
Runs continuously in the background, testing subject lines, openers, and calls to action.
Wakes on every few minutes
-
Winner picker
Calls the test when there is a real winner, and only when the result is strong enough to trust.
Wakes on a test matures
-
Quality check
Grades what the agents write before it goes out, so quality does not drift over thousands of messages.
Wakes on before every send
-
Multi-offer testing
Runs several offers side by side to find which one the market actually responds to.
Wakes on you launch a test
-
Message tailoring
Rewrites the message for the specific industry so it does not read like a template.
Wakes on a lead is added
-
Campaign engine
Keeps every campaign moving on schedule and inside your sending limits.
Wakes on every few minutes
Knowing your customers 5 agents
Every contact record builds itself from what has actually happened, instead of waiting for someone to type notes.
-
Lead memory
Keeps a running summary of everyone: what they said, what they wanted, where things stand.
Wakes on anything happens
-
Data cleaner
Fixes names, titles, companies, and phone numbers as leads arrive so your list stays usable.
Wakes on a lead arrives
-
Buying signals
Flags the moments that matter: interest, an objection, a bad fit, or a clear no.
Wakes on a message is read
-
Conversation summaries
Gives you the state of any thread in one line, so you can pick it up without reading back.
Wakes on a thread updates
-
Integration builder
Works out how to connect a new tool by reading how it behaves, instead of waiting for a connector.
Wakes on you add a tool
Voice 3 agents
Phone work handled in a real voice, tied to the same records as everything else.
-
Voicemail
Leaves a voicemail in a natural voice, written for that person rather than read from a script.
Wakes on a call is missed
-
Voice agents
Answers and makes calls, and knows who it is talking to before the call connects.
Wakes on the phone rings
-
Call review
Reviews what was said on the call and files the outcome against the contact.
Wakes on a call ends
Cost control 1 agent
Every agent action is metered, so intelligence is a line item you can see rather than a surprise.
-
Spend tracking
Tracks what every agent costs to run, by workspace, so you always know what the system spends.
Wakes on any agent acts
They start when the work arrives.
The work happens whether you’re watching or not. Every action is written down so you can check it afterwards, and you can put a stop on anything you want eyes on.
Free. 45 minutes. A few partners per market.
- 1Book your strategy call
Six quick questions, then pick a time. Free, 45 minutes.
- 2Get a straight answer
How it works, what it costs, how the guarantee works. Yes, no or not yet.
- 3Launch with our team
We do eight of the ten launch steps. You talk to owners and close the deal.