ABM · OUTBOUND · PERSONALISATION

AI Outbound & ABM: from account selection to sales sequences

AI can make outbound faster, but speed is not the strategy. I use it to improve account selection, research and personalisation while keeping targeting, relevance and sales judgement explicit.

What it means

I treat outbound and ABM as an operating system: define the ICP, decide which accounts matter, identify relevant triggers or pain hypotheses, create appropriate messaging, execute across channels and feed responses back into the model.

The problem it addresses

The common failure mode is scaling low-relevance outreach. More enrichment and more generated copy simply produce more noise if account selection, timing and offer are weak. AI should improve the quality of the decision before it increases the volume of execution.

How I approach it

At Think & Grow, I designed signal-based outbound programs in Clay with ICP scoring, tiered account lists, fit gates and AI-generated pain lines, executed as HubSpot sales sequences. I also built target lists, HubSpot campaigns and multi-step email, call and LinkedIn sequences for campaigns including AWS Marketplace Roadshow, R&D Tax Incentive ABM and US revenue programs.

A practical workflow

  1. 01

    Define account fit and exclusions.

  2. 02

    Tier accounts and add timing or growth signals where available.

  3. 03

    Generate a pain hypothesis or research angle, then review it for plausibility.

  4. 04

    Choose the sequence and channel mix: email, call, LinkedIn or a campaign path.

  5. 05

    Measure replies, connects, meetings and downstream opportunity movement rather than optimising only for send volume.

Tools and systems

ClayHubSpotClaudeABMEmailLinkedInSales sequences

Measured examples

24Enterprise accounts acquired through ABM and personalised campaigns at HIVERY.
26% reply rateFrom personalised cold outreach at HIVERY.
7% call-to-connectFrom a 30-day ABM experiment at HIVERY.
~620 accountsHubSpot target-account structure built at Think & Grow.

These are separate examples from different periods and programs. They should not be read as a single funnel or as results produced by one AI workflow.

← All capabilities