If you’re looking for some inspiration on how to use AI for ABM (Account Based Marketing) – in this post, you will find 3 practical ways to use AI for your ABM on LinkedIn:
I still remember staring at LinkedIn Campaign Manager for hours, trying to work out which accounts were actually moving, who to pass to sales, and which campaigns I should kill. That pain is what made me lean so hard into AI – both “generic” AI like OpenAI’s API for creative work, and purpose-built AI like Zena (our LinkedIn Ads AI analyst inside ZenABM) for analysis.
Account-Based Marketing (ABM) is basically:
Pick a finite list of high-value accounts
Get them to actually care about you
Prove that the money you spent moving them through the funnel turned into pipeline and revenue
When we started ABM at Userpilot, our first “proper” ABM program on LinkedIn generated over $900k in pipeline in five months – with around $8 in pipeline per $1 spent. It was not glamorous (think: spreadsheets, hacked-together reporting, brutal ops), but it was clearly better than generic cold outbound for our ACV and deal size.
Your exact buyers live there (titles, industries, seniorities are all targetable).
You can build persona- and stage-based campaigns with high message control.
LinkedIn is still the best “billboard in front of your ICP” for most B2B.
The catch: LinkedIn Campaign Manager is not built for ABM. You get campaign-level stats, but not the company-level view you need to run real ABM. That’s exactly where AI + a tool like ZenABM come in.
When people hear “AI for ABM” they often think “write me some headlines”. That’s maybe 10% of the value. Here’s where I actually use AI across an ABM program:
Strategy & ICP
Summarise win/loss notes, CRM notes and customer interviews to refine my ICP.
Cluster best-fit accounts and buyers to find patterns (industry, size, stack, triggers).
Audience & list building support
Build account lists against my ICP description (with Clay)
Generate rules and filters (“what filters would I use in LinkedIn to reach companies like these 50 logos?”).
Ad Creative & messaging
Generate initial value props, angles and ad copy variations by persona & stage.
Turn long-form assets (webinars, case studies) into ad briefs.
ABM Campaign analysis & optimisation with AI
Ask questions like “Which campaigns brought in the most interested-stage accounts?” or
“Where am I wasting budget?” – and get answers in natural language instead of pivot tables.
Finding top engaged accounts for Sales handover & warm outbound
Turn account-level engagement data into highly personalised outbound copy for BDRs (or use a home-made “AI SDR” in Clay/Smartlead).
You can absolutely do parts of this with ChatGPT or the OpenAI API directly. Where it gets really powerful is when you combine a generic LLM with your own first-party data – which is exactly what Zena AI does on top of ZenABM 😉
Watch this interactive demo to see how Zena can help you use AI for your ABM campaigns on LinkedIn easier!
Let’s start with the easy and fun part: ads.
When we ran our first big ABM campaign, we needed dozens of variants:
Different personas (PM, PMM, CS, VP Product, etc.)
Different stages (aware, interested, considering)
Different angles (pain, outcome, social proof, product feature)
Doing that manually is painful. Here’s how I now use OpenAI (or ChatGPT) to scale this without turning everything into generic AI mush.
I usually paste in:
A short description of our product and positioning
The specific persona (e.g. “Director of Product Marketing at mid-market SaaS”)
The campaign intent (e.g. “Feature adoption & in-app guidance”)
The ABM stage (e.g. “awareness” vs “considering”)
Then I ask for:
5–10 variations of single image ad copy
5–10 opening hooks for Thought Leader Ads
3–5 angles I may have missed
Example style of prompt:
“You’re an experienced B2B SaaS marketer. Write 5 LinkedIn sponsored content ad variants for {persona}, at the {ABM stage} stage.
Focus on {pain} and {outcome}. Use clear, conversational language, 1–2 short paragraphs + a CTA.”
I never paste the output straight into LinkedIn. I:
Pick 1–2 strongest angles per persona/stage
Rewrite in my own voice
Add specific proof (numbers, logos, quotes)
Make sure it matches our visuals and landing pages
If you’re running multi-persona ABM campaings, doing this in ChatGPT alone gets messy. With separate projects on chatGPT for different personas and ad formates, you can:
Store personas, pains, JTBD, proof points in a spreadsheet
Use a simple script to generate ad copy for each persona and ad briefs (with the right formats each ad type (single image vs. video)
The goal is not to let AI generate a 100% usable ad creative on its own – it’s to get from blank page to a good 80% starting point, fast.
Generic AI can’t help you much if it doesn’t see your data. That’s why we built Zena.
Zena is an AI chatbot that sits on top of:
Your LinkedIn Ads API data, and
Your CRM data (HubSpot or Salesforce), plus
Your ZenABM company-level engagement & ABM stages
Instead of exporting CSVs, cleaning them, and then uploading to ChatGPT, you just connect ZenABM once – and then you can literally chat with your LinkedIn ad + CRM data.
For example, you can ask Zena things like:
“Which of our campaigns generated the most pipeline this quarter?”
“Who are the top companies clicking our image ads?”
“Which campaigns influenced the most revenue this month?”
And get answers like:
A list of campaigns ranked by influenced pipeline
A table of companies, their engagement, and ABM stage
Breakdown of spend vs pipeline per campaign or per account
That’s AI for ABM in the most literal sense: instead of you doing analysis, you ask the question and Zena does the heavy lifting on top of actual first-party data.
With Zena, you can de-anonymize ad performance on the target account (company) level. It will show you exactly how each company interacted with your ads and how far they moved through your funnel (see: account stages).
You’ll know which companies clicked your ads, viewed your content, or filled out forms – and you’ll see this data tied to your CRM records.
This lets you answer questions like “Which companies engaged with our recent campaign?” and “How much pipeline did we generate from company X last month?” without manual analysis.
Zena can also help you identify high-value targets for sales outreach. For instance, it can quickly show you which companies had the most ad clicks last week, indicating they’re ready for engagement:
Most ICP descriptions are hand-wavy paragraphs in a Notion doc. AI can help you turn that into something structured and data-driven.
Here’s how I do it:
Feed AI your “truth”
A sample of closed-won deals
A sample of closed-lost or churned accounts
Notes from customer interviews and win-loss analysis
Ask it to find patterns
For example:
“What do our top 30 customers have in common in terms of industry, size, tech stack, triggers?”
“How are our worst-fit customers different from our best-fit customers?”
“Which firmographic patterns correlate with high ACV / short sales cycles?”
Translate patterns into LinkedIn filters
“Based on this ICP description, suggest LinkedIn targeting filters and exclusions.”
“Given these 50 logos, what would be a good Boolean search for job titles to target?”
Close the loop with engagement data
Once you run campaigns, use Zena to ask:
“Which industries have the highest engagement and pipeline?”
“Which company sizes are most likely to reach ‘Considering’ stage?”
AI doesn’t “know” your ICP better than you do – but it’s very good at summarising noisy inputs and spotting patterns you might miss when you’re tired and staring at dashboards.
AI tools for ABM on LinkedIn like Zena break down engagements by job title (and other demographics) so you see which roles are most interested in your content. For example, it might show that “Marketing Managers” clicked your latest campaign 200 times, whereas “CTOs” clicked only 50 times.
Since Zena tells you exactly who (which job titles) is clicking your ads, with this data, you can adjust your targeting in future campaigns or tailor messaging to the job functions that matter most.
You can also ask Zena which companies had the highest cost per click and didn’t convert, so you can exclude them from your targeting:
Zena AI also gives you instant insights into how the companies are progressing down the funnel:
AI can completely transform how you build target account lists for ABM on LinkedIn using Clay, turning what used to be a manual, error-prone research process into a scalable, data-rich workflow. Instead of starting with a vague ICP description, you can feed Clay + an LLM (ChatGPT/OpenAI) your closed-won deals, your best-fit customer attributes, your technographic requirements, industries, funding stages, hiring signals, competitor users, and even top engaged companies from your broad-targeting LinkedIn ad campaigns – and let AI generate a precise, structured set of accounts, enriched with firmographic and technographic filters.
In Clay, you can then automate the entire process: use AI to classify each domain by ICP fit (“Is this company similar to our top 50 customers?”), enrich companies with BuiltWith or Clearbit data (“Which of these companies use Amplitude + Segment?”), apply intent-like triggers (“Show me companies hiring PMs or UX roles in the last 60 days”), and even score each company using an AI-generated scoring rubric (“Score this company from 1–100 based on ICP match, tech stack relevance, revenue, and recent growth signals”).
Once Clay assembles the enriched dataset, AI can then summarize each company’s narrative – why it’s a good fit, which product pain it likely has, what tools it uses, and which personas you should target on LinkedIn—and create tiered account lists (Tier 1, Tier 2, Tier 3) based on predicted ACV or ease of entry.
Finally, with one click, you can sync these structured, AI-scored ICP-fit accounts directly into HubSpot or Salesforce, where ZenABM can immediately start tracking LinkedIn account-level engagement. The result is a living, AI-curated target account list that updates automatically, reflects real buying triggers, and is 10× more accurate than any static list built by hand.
Once your ABM campaigns on LinkedIn are running, you want to know:
What’s actually working?
For whom?
At which stage of the funnel?
You can get this by uploading CSV files to chatGPT from Campaign manager of course, but With Zena, I can get this without building a single pivot table.
Typical questions I ask:
“Which LinkedIn campaigns performed best by CTR this month?”
“Rank campaigns by those that drove the most ‘Interested’-stage companies.”
“Which ABM campaigns influenced the most revenue this quarter?”
“How did my ABM campaign performance change month over month?”
“Which campaigns underperformed based on spend vs engagement?”
“Which campaign drove the lowest cost per engaged company?”
Under the hood, Zena is joining:
CRM deals & pipeline
ZenABM’s ABM stages & account scores
On the surface, you just get straight answers like:
“Campaign group ‘Feature Adoption – PMs’ generated $280k in influenced pipeline on $14k spend (20x pipeline per $), driven mainly by these three single image ads…”
Which is exactly the type of sentence your CEO or CRO cares about.
Zena AI will tell you which of your LinkedIn campaigns and ABM campaigns are performing best – in terms of driving engagement, or pipeline.
Zena can shows detailed engagement data per campaign – clicks, impressions, and more – for each company:
It can also give you insights into your top-performing ABM campaigns (spanning several LinkedIn campaigns, targetting e.g. different audiences) over time:
Zena combines ad engagement data with CRM deals to give you pipeline influenced by LinkedIn ads and show you how your LinkedIn ads affect revenue?. It can break this down by month, so you can track trends over time.
AI helps you compare performance across different ad formats. Whether you’re using text link ads (TLAs), image ads, video ads, or carousel ads, tools like Zena can analyze each format’s effectiveness in terms of driving engagements and pipeline:
You might learn that video ads are driving the highest click-through rates or that carousel ads are attracting a particular industry. With these insights, you can double down on the formats that work and refine your creative strategy.
One of my favourite parts of combining AI + ABM data is budget optimisation.
First, you get granular visibility through ZenABM:
Spend by campaign, campaign group, and account
Pipeline per dollar spent
Then you use Zena to answer questions like:
“Where did we spend the most budget this month?”
“Which companies cost us the most to engage, and how much pipeline did they generate?”
“Show spend by campaign + ROI.”
“Which campaigns wasted budget (high spend, low engagement)?”
“What’s my cost per engaged company for this quarter?”
At a more strategic level, I’ll also use generic AI (ChatGPT/OpenAI) to do scenario planning, e.g.:
“We want $1M in pipeline from ABM with a $120k LinkedIn budget. Our ACV is $50k, close rate is 25%, demo-to-opportunity rate is 75%. How many accounts do we need to target and what benchmarks at each stage do we need to hit?”
It’s the same math I described in my ABM LinkedIn guide – but instead of spending an afternoon in Excel, I sanity-check my numbers with AI in minutes.
AI gives you more more visibility into ad spend at both the ABM campaign and target account level:
Tools like Zena AI will report how much of your budget was spent on each campaign:
…and even each individual company you target:
This means you can track ROI precisely – seeing, for instance, that Company X got 30% of your budget but only contributed 5% of pipeline, signaling a need to reallocate spend.
This granular view helps you optimize your ABM campaign budgets for maximum impact.
Tired of spending hours preparing reports? AI can help you generate and share detailed reports in seconds.
You can export your LinkedIn Campaign Manager “Performance” “Engagment” and “Delivery” data for each month into CSV files, upload them into your chatGPT, and ask it to give you performance breakdowns month by month.
Or you can simply ask Zena AI for the report you need (it already has access to your LinkedIn adsAPI data) – for example, “Create a monthly LinkedIn Ad performance report” or “Give me an executive summary of my ABM campaign/ LinkedIn campaign performance in October vs November” – and Zena will pull all relevant data from LinkedIn and your CRM.
You’ll get a ready-to-share report detailing clicks, impressions, top accounts, ROI metrics, and more. This speeds up your workflow and ensures your team can make decisions based on the latest data.
Throughout all these features, Zena’s user-friendly chat interface makes it simple. It transforms complex LinkedIn Ads analytics into plain English insights so you can focus on strategy, not spreadsheets.
If I had to summarise how to use AI for ABM on LinkedIn, it would be:
Use generic AI (OpenAI / ChatGPT) for thinking and creation
Refining ICP & personas
Generating and iterating ad copy
Turning long-form content into campaign creative
Doing the first pass on planning and benchmarks
Use specialised AI (Zena AI + ZenABM) for analysis and decisions
Company-level engagement and ABM stages
Campaign and ad performance tied to pipeline and revenue
Keep humans in the loop
You still need judgement to pick messaging, define strategy, and talk to customers.
AI just removes the grunt work so you can spend your cognitive budget where it matters.
If you’re already running LinkedIn ads and you’re serious about ABM, start small:
Use ChatGPT/OpenAI to generate better ad variants and refine your ICP.
Then plug your LinkedIn Ads + CRM into ZenABM, and let Zena analyse performance and surface the accounts and campaigns that matter.
That’s the combo that finally let me stop living in CSV exports – and start running ABM on LinkedIn like a grown-up.
