Skip to content
NagentNagent
Log inSign upHire your AI Growth Team
NagentNagent
Log inSign up

How are B2B sales teams using AI video ads to warm up target accounts before outbound?

8 Minutes read
Published:
B2B teams upload their target account list to LinkedIn, run short AI-made video ads to the buying group for two to four weeks, then start outbound to the accounts that engaged. Here is the playbook, the timeline, how to measure it and the mistakes that make it backfire.
PB
Pratap BeheraNov 20, 2026·8 min read
How are B2B sales teams using AI video ads to warm up target accounts before outbound?

B2B sales and marketing teams warm up target accounts by uploading their account list to LinkedIn as a company audience, then running short AI-made video ads to the buying group for two to four weeks before reps reach out. AI makes it affordable to cut a version for each role and each problem. Reps then prioritise accounts that clicked, visited or filled a form, and open with the problem the ads covered, never with "I saw you watched our ad."

Below: why this works, the five plays teams use, a four-week timeline, the ad formats, how to measure it, what goes wrong, and where Nagent's video teammate Alpha and outbound research lead NORA fit.

Why warm up accounts with ads before outbound?

Cold outbound asks a stranger for time. Ads first change two things.

  • Familiarity. By the time an email lands, the buyer may have seen your name and your point of view a few times. That will not close a deal, but it can make the email feel less cold.

  • Signal. Accounts that click, visit or fill a form show interest you can act on first.

It also fits how B2B buying works. LinkedIn's B2B Institute says "95% of your potential buyers aren't ready to buy today" (LinkedIn). Ads keep you in front of the 95% at low cost; outbound focuses on the few who show signs of being ready.

AI video matters here because account-based campaigns need many small variants: one per role, per problem, sometimes per industry. Making each with a film crew rarely pays for a list of a few hundred accounts.

What plays are B2B teams running?

Five plays come up again and again.

1. Company-list ads on LinkedIn
Upload the target account list as a company audience and run video ads to people at those companies, filtered by job function and seniority. LinkedIn describes its matched audiences as a way to "run your account-based marketing campaigns to reach key companies and decision makers" (LinkedIn). Its API documentation notes that the first match of an uploaded list can take up to 48 hours, and an audience not used in a draft or active campaign for 30 days is archived (LinkedIn API docs).

2. Role-specific clips
The same problem, cut three ways: the finance lead hears about cost and risk, the operations lead about time, the IT lead about security and setup. With AI presenters and voice-over, each version is a script change, not a new shoot.

3. Website retargeting of target accounts
People from target accounts who visit your site get a short follow-up video. LinkedIn's Insight Tag matches site visitors to members for retargeting (LinkedIn API docs).

4. CRM-synced audiences
Marketing syncs lists from the CRM to the ad network, so ads follow the account's stage. HubSpot, for example, can sync a contact segment or a company audience to LinkedIn Campaign Manager (HubSpot).

5. Personalised video in the outbound itself
Not an ad, but often part of the same programme: an AI avatar video addressed to one account, sent in the sequence. HeyGen, for example, lists integrations with Apollo, Clay and Salesforce for prospect and outreach videos (HeyGen). Use it sparingly; a personalised video that gets a fact about the account wrong does more harm than a plain email.

What does a four-week warm-up look like?

An illustration for a list of 300 target accounts:

  • Week 0: set up. Agree the list with sales. Upload it as a company audience. Write one problem statement per role. Make three to six short AI video variants. Add UTM tags per variant.

  • Weeks 1 and 2: awareness. Run 15 to 30 second videos to the buying roles at those accounts. One problem per ad, real product screens, a low-effort call to action such as a two-minute walkthrough.

  • Week 3: retarget. People who engaged, or visited from a target account, see a deeper 45 to 60 second video and a pricing or checklist offer.

  • Week 3 onward: outbound. Reps start with accounts showing engagement. The first email opens with the problem the ads covered, in the rep's own words. Accounts with no engagement still get outbound, later and with a different angle.

  • Weekly: sales and marketing review which accounts moved, and swap the weakest variant.

Which AI video ad formats work for account warm-up?

  • Problem-first explainer (15 to 30 seconds). A presenter names one painful task, then the real product fixes it on screen.

  • Founder or expert point of view (30 to 60 seconds). One opinion about how the problem should be solved. Credible and hard to copy.

  • Role clip (15 to 20 seconds). One sentence of pain in the role's language, one screen, one CTA.

  • Proof clip. Only with a real, permissioned customer quote. An AI presenter must not pose as a customer.

Keep claims to what you can prove, and use your own screen recordings, not generated interfaces. Burn in captions so the ad works muted.

Who should own each part, sales or marketing?

The play fails most often at the handover, so name an owner for each step before launch:

  • Account list: sales and marketing together. Sales knows which accounts matter; marketing knows which will match.

  • Problem statements per role: product marketing, checked by two or three reps who hear these problems on calls.

  • Video variants and claims review: marketing, with a second person signing off every claim.

  • Audience upload, budget and launch: whoever runs paid media.

  • Weekly engaged-account list: marketing sends it; sales confirms who will act on each account.

  • First-touch emails: the reps, in their own words, referencing the problem rather than the ad.

Write the owners down in the campaign brief. When the weekly list arrives with no owner, the warm-up turns into ordinary brand spend.

How do you measure whether the warm-up worked?

Measure at the account level, not just the ad level.

  • Reach into the list: how many target accounts saw the ads, and which roles.

  • Engagement by account: clicks, landing page visits and form fills from people at target accounts.

  • Outbound lift: reply and meeting rates for engaged accounts against a held-out group of similar accounts that saw no ads. Without a holdout, you cannot tell warm-up from a better list.

  • Pipeline: opportunities and won deals from the list, by variant.

Know your CRM's limits. In HubSpot, for example, an ads interaction means a click or a form submission after engaging with an ad, it is recorded as a timeline event rather than a contact property, and the ad's destination URL must carry HubSpot's tracking (HubSpot). A view alone does not create one.

What goes wrong, and what does it cost?

  • Creepy outreach. "I noticed you watched our video" tells the buyer they were tracked. It costs trust and replies. Open with the problem instead.

  • Lists too small or too stale. A list that matches poorly wastes the budget. Clean the list before upload.

  • Sales not in the loop. Marketing runs ads, sales never hears which accounts engaged. The warm-up is wasted.

  • Generic ads. One ad for every role is the cold email problem in video form.

  • Overclaiming. An AI presenter quoting results nobody had. Under the FTC's endorsement guides, advertisers are liable for "misleading or unsubstantiated statements made through endorsements" (16 CFR 255.1).

  • No holdout. Every meeting gets credited to the ads, and budget decisions rest on a guess.

For outbound tools themselves, see AI SDR agents for outbound prospecting.

Where does Nagent fit?

We make Nagent, so weigh this as a vendor's view. Nagent is multiplayer AI for growth: people and AI agents in one shared context, delivering the number. Sales and Content teammates share one context, so the ads and the outreach can work from the same accounts and the same claims.

  • NORA, the AI outbound research lead, finds accounts that fit your customer profile, researches each into a sourced dossier and drafts the first email. Customer-facing steps wait for a person.

  • Alpha, the video teammate on CREA's Content team, makes the ads, including a LinkedIn video ad preset (1:1 or 16:9, 30 or 60 seconds, captions burned in, no generated shots) and software demos from your own screen capture. Nothing renders until a person approves the storyboard and its credit quote.

  • What it does not do. NIA, who builds paid campaigns, works on Google Ads, Meta and ChatGPT ads today, not LinkedIn, so the LinkedIn audience and campaign are set up by a person in Campaign Manager. Alpha never publishes or touches an ad account.

Frequently Asked Questions

How long should you run ads before starting outbound?

Two to four weeks is a common window: long enough for people at target accounts to see the ads a few times, short enough that the list is still current. Start outbound earlier for accounts that engage quickly.

Do personalised AI videos work in cold outreach?

They can, if the personal details are correct and the video is short. Check every fact about the account before sending, and keep a plain-text version for people who will not play a video.

Can I target a list of companies with LinkedIn video ads?

Yes. LinkedIn matched audiences let you upload a company list and target people at those companies, narrowed by job function and seniority. Allow time for the list to match before launch.

Should sales mention the ad in the first email?

No. Reference the problem the ads covered, not the fact that the buyer saw them. Mentioning tracking costs trust.

Want ads and outbound working from the same accounts and the same claims? Book a Nagent demo.

AI agents for SaaS & Technology

Continue learning

Related agents

Agents that match this read.

Browse all 200+ agents
Select Category
Sales

How are B2B sales teams using AI generated product images in outbound and sales decks?

B2B sales teams use AI product images to show the product in each prospect's setting, refresh decks per vertical, build line sheets before samples exist and explain software with clean annotated screenshots. Here are seven uses, the workflow and the lines not to cross.

Performance Marketing

How are B2B sales teams using AI to run Google Ads and LinkedIn ads?

Mostly in narrow, supervised jobs: Google's own bidding and AI Max fed with CRM outcomes, AI agents that triage search terms and draft copy for a person to approve, and LinkedIn creative variants per buyer role. Here is what works, the guardrails, and what fails.

Creative

Can AI make a video ad for a B2B software company?

Yes. AI can script, voice, present, edit and resize a B2B software ad in hours. It cannot invent your real interface or your customers' results. Here are the formats that work, a step-by-step process, costs and the mistakes to avoid.

Creative

Our AI video ads look inconsistent from one render to the next. How do we keep the same presenter and product?

AI video ads drift because the presenter, product and settings are re-invented on every render. Here are the causes, the fixes for each, a pre-launch check and how Nagent's Alpha locks the presenter and product.

Marketing

Top 10 AI Video Tools for Ads and Social (2026)

Compare the 10 best AI video tools for ads and social, from Nagent Alpha, Synthesia and HeyGen to Creatify, Arcads and MakeUGC, by the job each does best.

Sales

Our pipeline review takes hours and the CRM is always out of date. How are sales teams using AI to fix it?

Slow pipeline reviews and a stale CRM share one cause: status lives in reps' heads, not the record. Teams fix it by letting AI propose CRM updates for reps to approve and prepare the review from the record before the meeting.