Find experts by the stack they build on.

Bespoke, code-first builds rather than an off-the-shelf platform. Chosen when requirements outgrow packaged tools and the work justifies engineering it directly.

OpenAI

177

OpenAI's GPT models and API, the default foundation for a large share of production AI builds. A safe pick when you want a well-supported, general-purpose model.

Claude

137

Anthropic's family of language models, favored for careful reasoning, long documents, and drafting. Providers build on it where accuracy and tone matter.

n8n

91

An open-source, self-hostable automation platform. Experts reach for it when a workflow needs custom logic, private data handling, or code steps that hosted tools cannot run.

Make

68

A visual automation builder for multi-step, branching workflows. A common choice when a process is too involved for simple triggers but does not warrant custom code.

Zapier

57

The most widely adopted no-code automation tool, connecting thousands of apps. Best for getting reliable, standard integrations live quickly across a business.

AWS

52

Amazon Web Services, including Bedrock for foundation models. The choice for teams that need cloud-native infrastructure, scale, and data residency control.

HubSpot

36

A CRM and marketing suite that anchors many go-to-market teams. Providers wire AI into its records, pipelines, and sequences to automate revenue operations.

Airtable

31

A spreadsheet-database hybrid that often becomes the operational backbone of an automation. Experts use it as the shared data layer behind agents and workflows.

Gemini

29

Google's multimodal model family, strong on long context and tight integration with Google Workspace and Cloud. Chosen by teams already living in that ecosystem.

Salesforce

23

The enterprise CRM standard. Engagements here tend to be larger and more governed, embedding AI into established sales, service, and data workflows.

Retrieval-augmented generation: grounding model answers in your own documents and data so responses stay current and cite their source. The backbone of most knowledge and support builds.

The most widely used framework for composing LLM applications, from prompt chains to retrieval pipelines. A marker of a firm that builds against model APIs directly rather than wiring no-code tools.

An all-in-one sales and marketing platform popular with agencies. Experts layer AI onto its CRM, funnels, and messaging to automate client acquisition.

Power Platform

13

Microsoft's low-code stack (Power Automate, Power Apps, Copilot). The natural fit for organizations standardized on Microsoft 365 and Azure.

The Model Context Protocol, an open standard for connecting AI models to your tools and data through a common interface. A recent, telling signal of a firm building on the current agent stack.

LangChain's framework for stateful, multi-step agent workflows with explicit control over branching, memory, and human checkpoints. Favored for production agents that need to be inspected and debugged.

A framework for orchestrating multiple role-based agents that collaborate on a task. Chosen when a job splits cleanly into specialist agents rather than one monolithic prompt.

The databases (Pinecone, Weaviate, Qdrant, pgvector) that store embeddings and power retrieval at question time. Their presence signals a firm building real retrieval systems, not demos.

Evaluation suites, guardrails, and observability that measure whether an AI system actually works and keep it from regressing. The sharpest marker separating AI-native engineering from automation resale.

A design platform for conversational agents and voice assistants. Used to prototype and ship chat and phone experiences without building the flow engine from scratch.