AI Automation Specialist – AI Architect

Role Overview

• Design the agentic systems that run across hospitality design, hospitality strategy, and marketing workflows
• Decide the stack: which tools, which agents, which integrations, what memory architecture, what evaluation harness
• Document the architecture so it compounds rather than evaporates when you take leave
• Build agents using Claude, ChatGPT, Make, n8n, Zapier, the Claude Agent SDK, and adjacent platforms
• Maintain integrations across our stack: Gmail, Google Workspace, Streak, Xero, WhatsApp, ArchiCAD outputs, WordPress, Meta Ads, Google Ads, LinkedIn, Klaviyo
• Build evaluation and observability into every agent. We need to know when one fails.
• Build AI-assisted content systems for LinkedIn, Instagram, blog, EDM and short-form video, calibrated to brand voice per client
• Architect the brand voice infrastructure: founder interviews to voice guide to enforcement layer across all content output
• Build performance reporting pipelines that pull Meta Ads, Google Ads, LinkedIn Ads, and organic data into client-ready insights
• Design ad creative generation and variation pipelines tied to performance feedback loops
• Build SEO and AI-search optimisation systems (GEO / AEO), including keyword clustering, content briefs, on-page optimisation, and schema markup
• Build lead intake systems across website, WhatsApp, referral and inbound email, with classification and routing
• Design qualification scoring against our ICP, and the agent layer that drafts initial follow-ups
• Architect the CRM rhythm in Streak: deal stage updates from meeting transcripts, proposal follow-up cadence, pipeline reporting
• Build fee proposal generation: brief in, branded proposal out, with version control and follow-up logic
• Design LinkedIn lead-scraping, enrichment, and personalised outreach workflows for BDM
• Build the trigger logic for which deals sees personally and which the system handles
• Build the project workspace generation that triggers when a deal is won: folder structure, templates, access, programme, kickoff brief
• Design the reverse brief drafting agent that converts kickoff notes into structured project briefs
• Architect the cost planning workflow: from sketch to v0 cost plan in our FOH/BOH/Services/FFE/Prelims structure, with builder benchmark lookups
• Build the builder budget revision loop: receive revised budget, reconcile against cost plan, draft client memo, schedule meeting
• Design RFI log management, drawing register hygiene, and subcontractor follow-up systems
• Build the weekly client status reporting layer per active project
• Support bookkeeping rhythm with automation: Xero reconciliation triggers, invoice issuance against milestones, debtor follow-up cadence, BAS prep
• Embed AI into the daily operating rhythm of three businesses
• Coach the team on how to use the systems you build. We're not buying tools, we're buying capability.
• Stay across emerging AI tools, agents, frameworks and platform changes
• Bring proactive ideas. Test them. Tell us what worked and what didn't.
• Identify where AI extends judgement vs replaces it. Defend the line.
• AI is integrated into daily workflows across all three businesses in a measurable, defensible way
• The founder recovers significant weekly hours through agent-mediated work
• Project status reporting, cost plan generation, and client comms run on assisted workflows we can prove the ROI on
• Bookkeeping rhythm (reconciliation, invoicing, debtor chase) is supported by automation
• No Filter scales content production without proportional headcount growth
• We have a documented, defensible AI operating layer that the business owns, not the vendors.

Requirements

• You can architect a system and ship the first version yourself. Not one or the other.
• You have hands-on experience building with LLM tooling in production. Agents, RAG, prompt engineering, evaluation, multi-step workflows. The Claude / OpenAI / OSS ecosystem is your daily tool.
• You understand marketing, social media, content production, and growth from the inside. You've built or run marketing systems, not just observed them. You know what good content looks like, what performance metrics actually matter, and where AI does and doesn't lift the needle.
• You understand how sales pipelines and CRMs actually work. You've run leads through one, drafted proposals, chased deals, and felt where the friction sits.
• You understand business operations: project delivery, process design, the rhythm of running multiple workstreams in parallel. You've either run ops or you've built systems for someone who does.
• You're comfortable in a small business context. You know that elegance loses to "live by Friday."
• You can communicate with founders, not just engineers. You can explain why an agent works, why one didn't, and what you'd change.
• You have strong opinions about tools, architecture and sequencing. You're willing to be wrong, but you have a view.
• You care about hospitality, design, growth or creative work as industries. This isn't a generic AI role. Context matters.

Tech Stack

• WordPress