anfloy.Book a call

Anfloy blog

Notes on building AI agents.

Build logs, engineering deep-dives, and frameworks for buying AI, from inside an AI engineering firm shipping production agent systems for B2B teams.

Weekly1,200-2,500 wordsB2B founders + GTM leads

Latest

LinkedIn Content Machine: The Full Claude Code Setup (2026)

The full setup behind 1M+ LinkedIn impressions in 90 days: pillars, a voice that learns, Apify scraping, AI Ark and Claude qualification, Slack cards and the Slack agent, with prompts and code.

GTM EngineeringOct 2026

GTM EngineeringGTM Automation Testing & QA: Validate Before LaunchA practical testing and QA framework for GTM automations: what to test before launch, how to validate against real data safely, and the regression testing discipline most workflows skip.Oct 2026AI EngineeringAI-Powered Sales Forecasting: The Methodology Behind the NumberHow AI sales forecasting models actually work, why confidence intervals matter more than a single predicted number, and the specific reasons AI forecasts fail in practice.Oct 2026AI AgentAI Agent Memory: How to Build Persistent Memory for Autonomous AgentsLearn how AI agent memory works, including short-term, long-term, episodic, semantic, and procedural memory, retrieval, forgetting, and GTM agent memory architecture.Oct 2026GTM EngineeringPrompt Injection Prevention in External GTM BotsHow prompt injection threatens GTM bots processing external content, cold replies, scraped pages, and chatbot input and how multi-model guardrails defend them.Sept 2026GTM EngineeringProduction Orchestration: How to Run AI Agents and GTM Workflows ReliablyLearn how production orchestration coordinates AI agents, workflows, tools, data, retries, state, observability, scaling, and human approvals in real-world GTM systems.Sept 2026GTM EngineeringAsynchronous Webhook Architecture & Queue Management for GTM PipelinesWhy webhooks drop payloads during GTM signal spikes, and the async queue architecture, dead-letter queues, worker pools, and idempotency that actually prevents it.Sept 2026GTM EngineeringIdentity Resolution & Deduplication in GTM: Matching Alphabet to Google & StripeHow identity resolution actually works in a GTM data stack: fuzzy matching algorithms, domain normalization, corporate hierarchy mapping, and contact-level resolution across sources.Sept 2026GTM EngineeringAI Sales Agent Observability, Tracing, and Evaluation (LLM-as-a-Judge for GTM)Learn how to evaluate AI sales agents in production using tracing, observability, LLM-as-a-Judge, groundedness, accuracy, tone, and GTM-specific eval frameworks.Sept 2026GTM EngineeringReverse ETL for GTM Engineers: Sync Warehouse Data to Clay & CRMA practical execution guide to reverse ETL for GTM engineers: syncing Snowflake or BigQuery into Clay and your CRM without hitting record limits, API quotas, or sync conflicts.Sept 2026GTM EngineeringHandling Unstructured Data: RAG & Vector Databases for Enterprise GTMLearn how to build a custom RAG pipeline and vector DB to index enterprise sales collateral and eliminate hallucinations in live deal cycles.Sept 2026GTM EngineeringHow to Build a Multi-Source TAM Mapping and Account Tiering Engine in ClayA step-by-step architecture for building a TAM mapping and account tiering engine in Clay: multi-source deduplication across Apollo, Ocean.io, and LinkedIn, and programmatic tiering from intent signals.Sept 2026GTM EngineeringDownstream GTM Engineering: How to Automate MOFO and BOFO ConversionLearn how GTM Engineers automate post-demo follow-up, champion enablement, deal velocity, stalled opportunities, proposals, and mutual action plans.Sept 2026GTM EngineeringSQL and dbt for GTM Engineers: A Practical GuideLearn how GTM Engineers use SQL and dbt to build GTM data models, scoring systems, customer health models, reverse ETL workflows, and revenue infrastructure.Sept 2026GTM EngineeringABM as a System: How ICP, Negative ICP, TAM, and Signals Actually ConnectABM isn't a campaign type, it's a system built from four pieces you likely already have: ICP, negative ICP, TAM mapping, and signal detection. Here's how they connect.Sept 2026GTM EngineeringHow We Run 24 GTM Agents: The Full Backend Build (2026)We run 24 GTM agents on our own pipeline. What each one runs on, how its prompt is shaped, what it costs, and the thing that broke.Sept 2026GTM EngineeringPost-Sale GTM Engineering: Churn, Renewals & OnboardingLearn how GTM Engineering extends beyond pipeline with post-sale systems for churn prediction, renewal automation, customer onboarding, expansion signals, and sales-to-CS handoffs.Sept 2026GTM EngineeringCustomer Data Platform (CDP): What It Is and How It Fits GTM EngineeringA practical explanation of what a CDP actually is, how it differs from a CRM and a data warehouse, the three CDP architectures available, and when GTM engineering actually needs one.Sept 2026

Let’s build what your company needs.

Day 0 is a working demo on your real data, free, before anything is agreed. Day 8 is handover with the documentation and the keys.

Book a call