Social Media automation

How to Get Clients on LinkedIn Using mobilerun AI

Priyasep 6, 20267 min read
Share
How to Get Clients on LinkedIn Using mobilerun AI

Most B2B founders eventually turn to LinkedIn for outbound.

And for good reason.

LinkedIn is one of the few places where your prospects have already told you who they are: their job title, company, industry, location, career history, and often what they are currently working on. That makes targeting unusually precise.

Instead of buying a list and hoping the data is accurate, you can search for exactly the people you want to reach:

Head of Mobile Engineering at companies with 200–1,000 employees, working in fintech, based in India, who recently changed jobs.

You can then open the results and immediately see whether those people actually fit your ICP. The problem isn't finding the right people. It's doing the same outreach work every day.

Search. Open profiles. Read them. Write messages. Send connection requests. Check who accepted. Follow up. Track conversations. Repeat. The strategy is very simple but the execution isn't. That's where AI agents become interesting.

Execution is the real bottleneck

If you already do LinkedIn outreach, you know the loop. Run a search. Open a profile. Read it. Write something that proves you read it. Send the request. Log it in a sheet. Check back in three days to see who accepted. Message the ones who did. Log that too. Repeat tomorrow.

None of that is hard. All of it is repetitive, and it has to happen every single day, forever, or the pipeline dries up. This is why teams hire five SDRs to do the same thing in parallel, and why most solo founders quit LinkedIn outreach after three weeks.

The work is not intellectually demanding. It is just work that requires a human sitting in front of the app. Which is exactly the shape of the problem an agent should be solving.

What changes when the agent uses the actual LinkedIn app?

There are plenty of tools that interact with LinkedIn through browser extensions, APIs, databases, or scraping infrastructure. mobilerun takes a different approach.

The agent operates a real Android device and interacts with the LinkedIn mobile app through the device itself. That means you can give it instructions in the same way you would brief a person doing prospecting.

Setting it up takes about five minutes:

  1. Connect a device. Either plug in a phone you already own, or spin up a fresh device in mobilerun Cloud if you would rather keep this off your personal hardware.
  2. Install LinkedIn on that device and log in to the account you want to run outreach from.
  3. Tell the agent what to do, in plain English. That is the whole configuration step. There is no campaign builder, no field mapping, no CSV import.

The workflow is pretty straightforward:

1. Connect a device

Connect an Android phone you already own, or use a device in mobilerun Cloud.

2. Install LinkedIn

Set up the LinkedIn app and sign in to the account you want to use.

3. Give the agent instructions

Describe the audience, your positioning, the tone you want, and what the agent should do. You don't need to build a complicated campaign or manually map every field. The important part is that you define the strategy and the agent handles the repetitive execution.

Personalization without writing every message yourself

Personalization is where most automated outreach falls apart. A generic message like this one (below) does not feel personalised.

"Hi Lisa, I came across your profile and thought our product could be a great fit for your team."

The alternative is manually researching every prospect and writing every message, but it is hard to scale. mobilerun agent sits somewhere in between. For eg, you give it the context it needs:

Our pitch: We help mobile teams reduce QA cycle time by automating testing on real devices.

Audience: Mobile engineering leaders.

Tone: Short, conversational, no corporate language.

Instruction: Reference something specific from the person's profile when relevant. Don't force personalization if there isn't anything useful to mention.

You can also give it a rough template:

"Hi {name}, saw you're leading {function} at {company}. We help teams like yours {benefit}. Worth a quick chat?"

The template is just an example, not something you should swear by. For eg.

Robert H. : Head of Mobile Engineering at a fintech

His profile mentions a recent post about flaky Appium test suites slowing down releases. A contextual message might be:

"Hi Rahul, saw your post about flaky Appium suites holding up releases. That's a problem we work on running tests on real devices instead of relying only on emulators. Worth comparing notes?"

Sarah T. : VP Engineering at a logistics startup

Her profile shows that she recently joined the company and is building out the engineering team. The message could instead be:

Hi Sarah, congrats on the new role. Building out a mobile team usually means QA gets complicated pretty quickly. That's one of the things we help teams with. Worth a chat?

See the pitch has not changed but the context has. That's the important distinction between personalised automation and simply sending the same message to everyone.

Two ways to build the workflow

Your LinkedIn setup will look different depending on whether you use Sales Navigator.

Option A: You have Sales Navigator

Sales Navigator gives you the good filters, seniority, headcount, industry, technology, recent job changes, and saved lead lists that update themselves. Point the agent at those.

Connection requests with notes. Same as above, but your search is sharper, so your acceptance rate is better.

Direct messages without connecting first. Normally your outreach only lands once someone accepts your request. With Sales Navigator you can skip that step entirely for some prospects, using InMail. So the instruction becomes:

"Find VPs of Engineering at Series B fintech companies in Bangalore. Send each one a personalized InMail about our QA automation."

The agent finds them, reads each profile, writes each message, sends it.

No spreadsheet needed. Sales Navigator already logs what you have sent and to whom, and keeps your lead lists in sync. You do not need to duplicate that tracking anywhere. Let the platform hold the record and let the agent do the work.

Path B: if you do not have Sales Navigator

You can run the whole motion on a free account. It just needs more structure, because you are working with tighter limits and you will mostly be relying on connection requests rather than notes.

Step 1: search and queue.

"Search LinkedIn for Head of Product at Series A SaaS companies. Send connection requests to 15 of them."

10-12 a day is the number to hold on to. More on why in a moment.

Step 2: set a recurring job for follow-ups. This is the part people skip, and it is where the actual conversations come from. Ask the agent to check daily for newly accepted connections and message them:

"Every day at 10am, check who has accepted my connection requests since yesterday. Send each of them a personalized opener based on their profile. Wait two days, then send one follow-up to anyone who hasn't replied."

Now you have a real sequence running, without a sequencer.

Step 3: track it in a Google Sheet. Without Sales Navigator you need your own record, and the agent can maintain it:

"Keep a Google Sheet with one row per person: name, title, company, date request sent, accepted yes/no, date accepted, message sent, replied yes/no. Update it after every run."

After two weeks that sheet tells you what you actually need to know,  which job titles accept, which openers get replies, and which segments are a waste of time. That is the feedback loop that makes the next month better.

How much LinkedIn outreach should you send?

LinkedIn caps connection requests at roughly 100 per rolling seven-day window. It is a rolling window, not a Monday reset. Your cap frees up seven days after each request was sent.

A few things people get wrong about this:

  • The cap is the same on every tier. Free, Premium, and Sales Navigator all sit at roughly the same ceiling. You cannot buy your way past it.
  • New accounts are throttled harder, sometimes to 20 or 50 a week, until you build up history and a decent acceptance rate.
  • Daily pace matters too, not just the weekly total. Staying at 15–25 a day keeps you well inside the envelope. Firing 100 in one afternoon does not.
  • Free accounts are limited on personalised notes specifically. If notes are central to your approach, that is the real argument for a paid tier not the invitation count.

Hitting the cap is not catastrophic; you just stop until it resets. But sustained aggressive behaviour is what triggers actual restrictions, and those escalate: limited search, feature restrictions for one to three weeks, and in the worst case a permanent suspension.

Once it is set up, you stop doing anything

This is the actual payoff. The agent runs on your own device, or a device in mobilerun Cloud 24*7. 

You are not sitting in the LinkedIn app. You are not remembering to check who accepted. You are not maintaining a sequence in a tool that costs $80 a month per seat. You set the targeting, you set the pitch, you set the volume, and then you check the replies.

Getting started

If you already do LinkedIn outreach manually, you have the hard part done. You know who you are targeting and you know what to say to them. The only thing missing is someone to do it every day.

Connect a device in mobilerun, install LinkedIn, and describe the loop you are already running by hand. Start at 15 requests a day, watch the first week's acceptance rate, and tune the message from there.


Also Read:

  1. Whatsapp automation
  2. Instagram automation