How a zero-budget AI intent engine replaced 6sense and converted 18% of strategically selected accounts
Client: FigPii / Invesp — CRO SaaS | Timeline: Sep 2024 – Jul 2025
Executive Summary
| Metric | Result | Impact / Context |
| Target Accounts | 78 (strategically selected) | Precision over volume |
| New Clients Won | 14 | Successful pipeline conversion |
| Conversion Rate | 18% | Account-to-client (Well above industry average) |
| ARR Per Client | ~$10K | Immediate revenue impact |
| Account Profitability | +23% | Via expansion & reactivation |
| Sales Cycle Reduction | 15–25% | Through nurture journey design |
| Manual Work Eliminated | 40% | Via AI workflow automation |
Context & Overview
- Client: FigPii / Invesp — Conversion Rate Optimisation (CRO) SaaS
- Market: B2B SaaS — ecommerce and product-led companies
- Challenge: No defined ICP, broad unfocused targeting, zero intent tool budget
- Goal: Build a focused ABM programme that generates pipeline and closes clients
- Tools: n8n, Clay, Apify, Claude, HubSpot
- Channels: LinkedIn, email, direct mail, private events, content-led nurture
The Problem
FigPii was targeting everyone. No ICP. No scoring. No intent layer. Budget was too limited for tools like 6sense or ZoomInfo.
The challenge: Run a 1:1 ABM programme at scale, with full personalisation, on a lean budget — and build the intelligence layer from scratch.
Strategy at a Glance
- Built a custom ICP scoring model from existing customer data — not generic firmographics.
- Strategically selected 78 accounts from a broad pool based on fit signals — not assigned, chosen.
- Replaced 6sense with a self-built AI intent engine using 6 automated sequences.
- Ran private coaching sessions and content-led events to warm executive buyers.
- Built a post-event follow-up playbook feeding directly into bottom-of-funnel plays.
- Automated scoring, segmentation, and outreach so every account got the right play at the right time.
What I Did — Step by Step
Step 1 — Built the Truth Base (Days 0–20)
Analysed existing FigPii customers to find what actually predicts fit. The signals that mattered were specific:
| Customer Signal | Scoring Impact |
| Web traffic above 50,000 | +10 points |
| Running paid ads | +10 points |
| Unstructured landing pages on ecommerce sites | +10 points |
| Having an AB test developer or product owner | +10 points |
| Using tools like Hotjar | +10 points |
| Using a direct competitor tool | -10 points |
Execution: Apify scraped LinkedIn profiles, company pages, and the web. Clay enriched the account list. Claude cross-referenced enriched customer data against the broad prospect pool and identified the signal clusters that actually predicted conversion — not assumed ones.
Outcome: A scoring baseline built on real customer truth, not guesswork.
Step 2 — Scored Into Buckets, Not Tiers (Days 20–40)
Pulled a broad list, enriched it with the same signals as the customer list, then scored every account into one of four buckets:
- SAM — Strong fit, target now
- TAM — Good fit, nurture for later
- Others — Low priority
- Exclude — Disqualified
n8n ran the scoring logic automatically and pushed accounts into the right HubSpot buckets. No spreadsheets. Claude stress-tested the scoring weights against real customer data.
From the broad pool — 78 accounts were strategically selected for SAM. Not assigned. Chosen.
Step 3 — Built the AI Intent Engine (No Budget for 6sense)
Built a 6-sequence AI workflow engine that replicated what 6sense does — using n8n, Clay, Apify, Claude, and HubSpot.
| Sequence | Name | Workflow & Function |
| 1 | Account Identification | Apify scrapes hiring activity, product releases, competitor moves, LinkedIn posts on a schedule. |
| 2 | Intent Scoring | Claude reads the raw signal dump and assigns a priority flag per account — hiring spike, product launch, event-led, content push, silent, or community active. |
| 3 | Contact Intelligence | Apify pulls LinkedIn activity per contact; HubSpot surfaces full interaction history; Claude builds the job to be done per person. |
| 4 | Personalisation | Claude combines persona, contact history, and live account signal — the message only goes out when all three are connected. |
| 5 | Routing Decision | Claude decides which play fires; n8n executes automatically — no human in the loop. |
| 6 | Asset Generation | When a prospect requests something, Claude searches the full content library and generates a branded personalised PDF in under 2 minutes. |
This system eliminated the need for 6sense, ZoomInfo, or any paid intent platform — and cut agency spend entirely.
Step 4 — Private Events & Executive Engagement
Ran private coaching sessions and open content-led activities to warm executive buyers inside target accounts. These were not brand-awareness events — they were designed to identify buying intent, surface objections, and create direct conversion opportunities.
- Private coaching sessions for product owners and AB test practitioners
- Content-led group activities that surfaced jobs to be done per account
- Event attendance signals fed directly into HubSpot and triggered the follow-up sequence
Result: Attendees directly influenced pipeline, increased account engagement, and converted to clients.
Step 5 — Post-Event Follow-Up Playbook & Bottom-of-Funnel Execution
Built a structured playbook that activated immediately after every event interaction. The playbook covered:
- Trial conversion campaigns — personalised to what the attendee engaged with
- Proposal acceleration — tailored assets addressing the specific gap surfaced at the event
- Executive stakeholder outreach — targeted the decision-maker above the event contact
- Direct mail — high-touch physical touchpoint for accounts showing strong buying signals
- Demo conversion campaigns — personalised invitations based on event engagement data
Every play triggered automatically by n8n. No manual triage.
Why It Worked
- Precision over volume: 78 accounts chosen for fit, not because they were in a territory or industry.
- AI did the repeatable work: Scoring, enrichment, routing, and sequencing ran without human input.
- Three-dimensional personalisation: Every outreach combined persona, contact history, and live account signal — not a template.
- Events were pipeline plays, not brand plays: Private sessions created conversion opportunities, not just awareness.
- The system compounded: Account intelligence refreshed every 90 days automatically — the longer it ran, the sharper it got.
Tools & Infrastructure
| Tool | Function |
| Apify | Scraped LinkedIn profiles, company pages, job listings, landing page structure, and tool stack signals on a schedule. |
| Clay | Account sourcing, enrichment, and initial list building. |
| n8n | Automation layer — ran scoring, refreshed intelligence, triggered sequences, updated HubSpot. |
| Claude | Pattern recognition, intent scoring, job-to-be-done diagnosis, personalisation briefs, routing decisions, asset generation. |
| HubSpot | CRM — stored all account intelligence, contact history, deal data, and sequence tracking. |
Client Testimonial
“Meghana excels at selecting high-fit accounts using data-driven insights, not guesswork. She proactively identifies funnel gaps and creates effective campaigns without waiting for requests. Her cross-functional collaboration and strategic focus make her more than a marketer — a true partner.”
— FigPii / Invesp
Talk to Me About Similar Results
- LinkedIn: Open profile
- Email: meg.bhate@gmail.com
- Phone: +48 533 579 077