Auto prospecting means letting software handle the repetitive half of outbound: finding people who match a trigger, checking they fit, and sending the first touches on schedule. It does not mean the whole job runs while you sleep. The parts that decide whether you get replies, who you target and what you say, stay yours.
I run a LinkedIn prospecting agency. My team sends outbound every day, mostly through automation, and the failures I see are almost never technical. They come from automating the wrong step: teams automate volume before they automate targeting, then blame the tool when nobody answers.
This page breaks the work into four parts, says what to automate in each, and is honest about what stays manual.
What automated prospecting is, and what it is not
An automated prospecting system does four things: it builds a list from a trigger, it filters that list against your customer profile, it sends a sequence of touches, and it hands the conversation to you when someone replies.
What it does not do: decide who your buyer is, write a first line worth reading, or answer a reply. Every tool that claims otherwise is selling volume, and volume is the cheapest thing in outbound. Anyone can send 500 connection requests a week. Almost nobody can send 80 that deserve an answer.
The useful mental model: automation moves work from your calendar to your setup. You pay once, in thinking, then the machine repeats it.
The four parts of the system
| Part | What software does well | What stays yours |
|---|---|---|
| Trigger and list building | Watch a source continuously, collect matches | Choosing the trigger |
| Filtering | Apply rules and AI checks at scale | Defining what a good fit looks like |
| Sequencing | Send on schedule, respect delays and conditions | Writing the messages |
| Inbox and handover | Centralize replies, stop sequences on reply | The conversation itself |
Skip a part and the whole thing degrades. Most tools sell part three, which is why so many teams end up with a fast way to contact the wrong people.
Step 1: automate a trigger, not a list
A static list is stale the day you export it. A trigger keeps producing.
Triggers that work for a small B2B team, in rough order of quality:
- Job posts. A company hiring for a role that only exists when your problem exists. The hiring manager is named on the post most of the time.
- Profile visits. Someone looked at you. That is the shortest path from signal to opening line.
- Webinar and event attendance. Public interest in a topic, with a name attached.
- First degree network activity. People commenting on posts in your niche, including on your competitors' posts.
- Funding and headcount changes. Slower, but they tell you a budget appeared.
Pick one and run it for a month before adding a second. A trigger you check every week beats four you configured once. The full list of signals I use, with the first message that goes with each, is in how to identify buying signals, and the wider category these signals belong to is covered in B2B intent data.
Step 2: filter before you send, not after
This is the step that separates automated sales prospecting from spray and pray, and it is the one teams skip because it feels like it slows them down.
Two layers do the job.
Hard rules first. Company size, market, geography, language, and an exclusion list of current customers, competitors and anyone already in another campaign. These cost nothing to apply and remove most of the noise.
Then a judgment layer. Does this person's title match the buying role? Does the company look like the ones that already pay you? This used to mean a human reading profiles for an hour a day. AI filtering does the first pass now, and it is decent at it, as long as you describe your customer in specifics rather than adjectives. "Head of sales at a 10 to 50 person B2B SaaS company selling to mid market" works. "Ambitious growth focused companies" does not.
Keep a sample. Read 20 profiles the filter accepted before you launch anything. If you would not have picked them by hand, your description is the problem, not the tool.
Step 3: sequences that still sound like a person
Automated does not have to mean generic. The structure that works for us on LinkedIn:
- A connection request, with a note only when the note references the signal.
- If accepted, a first message within a day, about them, one question, no pitch.
- A second message three to five days later that adds something useful: how someone else handled the same problem, a short observation, a resource.
- A last one after another week, short, easy to decline.
Then stop. A fourth and fifth follow up buy you almost nothing and cost you the impression you left.
Two rules make automated messages readable. Reference the trigger in the first sentence, so the message could not have been sent to anyone else. And write the way you speak, because a message that reads like marketing copy gets treated like marketing copy.
An example built on a job post trigger:
Hi Sarah, saw the two SDR roles you posted in Berlin. Curious how you are handling list building while the seats are empty, that is usually what slows the ramp. Happy to share what we see working if useful.
Conditions are where automation earns its keep: send this only if the invitation was accepted, wait five days, skip anyone who already replied, pause everything when someone answers. That logic running correctly at 9am every day is worth more than any clever template.
Step 4: decide what stays manual
Three things stay manual in every setup I have run.
Replies. The moment someone answers, automation stops. Anything else gets caught immediately and burns the account and the brand together.
Edge cases. The prospect who replies "not me, talk to my colleague". The one who answers with a question about pricing. Those are the ones worth your time.
Volume decisions. LinkedIn caps how many invitations you can send and restricts accounts that push past it. LinkedIn's own help pages on invitation limits and types of restrictions say restrictions typically lift within a week, but the platform does not publish a precise weekly number, and it varies by account. So treat any tool's promised send volume as a ceiling you do not want to test. Warm a new sender account slowly, keep daily numbers boring, and add sender accounts rather than pushing one account harder.
Nobody can promise an automation setup carries no risk for your account. Anyone who does is guessing.
What to automate first, by team size
Solo founder. Automate sequencing and filtering. Keep list building manual at first, because building lists by hand for a month teaches you who your buyer is. Then automate the trigger you kept using.
Two to five people. Automate all four parts, but keep one shared inbox so replies do not sit unanswered for two days. At this size, response time beats volume.
Agency or multiple clients. Separate workspaces per client, separate sender accounts, and one dashboard. The bottleneck is not sending, it is knowing which campaign is producing without opening five tools.
Where Sendable fits
Sendable, the product I am building with my cofounder, covers the four parts above for LinkedIn. It watches signals (job posts, webinars, profile visits, first degree network activity), filters profiles against the ideal customer you describe, pushes the matching ones into a sequence you built in a visual workflow, and centralizes replies in a shared inbox with multiple sender accounts and one workspace per client. Published pricing: $49 per sender per month from 1 to 4 senders, $39 from 5 to 25, $1,000 a month for unlimited senders.
What it is not: an email tool, a database you can export, or a way to skip the thinking in steps one and three. It is LinkedIn only, and parts of the product are still being finished. If your buyers live in an inbox rather than on LinkedIn, a different category fits better. Details on getsendable.io.
FAQ
What is auto prospecting?
Auto prospecting is using software to run the repetitive parts of outbound: watching a trigger for new prospects, filtering them against your customer profile, and sending the first touches on schedule. The parts that need judgment, choosing the trigger, writing the messages and handling replies, stay with you.
Is automated prospecting worth it for a solo founder?
Yes, once you know who your buyer is. Automating sequencing and filtering gives back several hours a week. Automating list building before you have talked to 50 prospects tends to lock in the wrong target faster, so start manual on the list and automate the sending.
Does automated prospecting get your LinkedIn account restricted?
It can, if you push volume. LinkedIn limits invitations and restricts accounts that go past what a person could plausibly send. Keep daily numbers modest, warm new sender accounts over several weeks, and add senders instead of raising one account's volume. No tool can promise an account will never be restricted.
How many messages should an automated sequence contain?
Three or four touches, spread over two to three weeks, then stop. Later follow ups add very little and make the next campaign to the same person harder. Stopping the sequence the moment someone replies matters more than the number of steps.
What is the difference between an automated prospecting system and a sales engagement platform?
An automated prospecting system starts earlier, at the trigger and the list. A sales engagement platform assumes the list exists and focuses on cadences, dialing and reporting across a sales team. Small teams usually need the first. Teams with several reps and an existing pipeline need the second.