AI Automation Specialist for Marketing Agency

Company: Vysta

Location: Remote (CEST preferred)

Experience: 4–8+ years in performance marketing/agency environments

About Vysta

Vysta is a performance marketing agency managing $100M+ in annual Google Ads spend for eCommerce brands. We’re building an automation layer that saves hours, reduces waste, and compounds performance across Google/Microsoft, YouTube, CRO, and analytics—with human-in-the-loop controls and platform-compliant tech.

We’re hiring two AI Specialists to accelerate this roadmap. If you’ve actually shipped Google Ads Scripts or API automations, wired BigQuery pipelines, and built alerting/approvals that media buyers trust, you’ll feel at home here.

What You’ll DoCore (both tracks)

• Build human-in-the-loop agents that propose changes and log diffs; buyers approve before execution.

• Start with platform built-ins (Google Ads Scripts & Automated Rules; Microsoft Ads Scripts) before custom apps.

• Maintain clean data contracts and audit trails; measure ROI in hours saved, waste reduced, and revenue lift.

• Keep everything policy-compliant (Google/Microsoft API & script terms).

Track A — Paid Media Automation

• Budget pacing (MCC-level): GAQL reporting + Scripts + Slack alerts; optional budget nudges within whitelisted bands.

• Anomaly detection: model CPA/ROAS/CPC/conversions; alert or pause only on extreme, policy-safe events; always write change notes.

• Hygiene: 404/out-of-stock URL watcher; naming/structure QA; duplicate/overlap catcher.

• Optimization: n-gram search-term mining (draft change sets), dayparting recommender, Auction Insights monitor, PMax query-to-creative mapping, LTV-aware value rules, cross-channel budget reallocation (Google ↔ Microsoft) with caps & logs.

• Governance: change simulation (“what-if” on bulk negatives); pre-flight policy lint; Auto-QBR builder; “why did ROAS drop?” root-cause explorer.

Track B — Creative / CRO / Analytics

• YouTube creative intelligence: hook/first-5s analyzer, thumbnail/title experimenter, creative taxonomy & tagging; link features → outcomes.

• Competitive: Ads Transparency Center miner → swipe-file & trend alerts (regions/formats).

• CRO & LP: ad-to-page scent checker; price/promo drift monitor; Lighthouse/Core Web Vitals watchdog; heuristic CRO agent (50-point checklist → prioritized test backlog with copy drafts); safe personalization recipes; form QA bot.

• Analytics: nightly BigQuery warehouse for Google/Microsoft/GA4/Shopify/Klaviyo/Stripe; MMM-lite (LightweightMMM/Robyn) for planning; incrementality calendars; revenue & cash forecasting; margin monitor.

• Enablement: SOP copilot; onboarding wizard; security/PII guardrails (redaction, data boundaries).

Outcomes You’ll Be Measured On

• Hours saved per account per month (from task/runbook logs).

• Media waste reduced (bad placements, 404/OOO, structural errors) as % of spend.

• Speed-to-change (issue detected → approved → resolved).

• Performance lift (CVR/ROAS gains from dayparting, asset mapping, reallocations).

• Gross margin per client & net revenue retention (upsell triggers driven by agents).

Simple impact model we use:

EBITDA Δ ≈ (Hours_saved × blended hourly rate) + (Spend × waste_reduction%) + (Incremental_margin_from_ROAS_lift) − (Tooling + Dev)

Must-Have Skills

• Google Ads Scripts (JavaScript) and Google Ads API (Python); strong GAQL.

• BigQuery & SQL, job scheduling, API auth, webhook integrations (Slack/Notion).

• Practical LLM orchestration (prompting + retrieval) with privacy guardrails (PII redaction, data boundaries).

• Strong experimentation mindset and basic statistics (time-series, holdouts; MMM familiarity is a plus).

Clear documentation habits; change logs and runbooks your teammates will actually use.