Apps & AI Tools

The Tools We Build With

We are fluent across the AI tools that matter — and we run them ourselves before we put them in front of a client. Below: the tools we work with, and Claude as a worked example of how deep that fluency goes.

Why this matters

Tools serve the method, not the other way round

A family business does not need every tool. It needs the right few, wired into a clear operating picture, with a person who owns each output. What follows is not a shopping list — it is the working knowledge behind that judgment. We keep up with the tools so a business owner does not have to.

Prototype. This is a first-pass layout for a new p1next section. Logos are placeholders; the tool list and the Claude write-up are for review, not yet published.

The Toolkit

What we work with

Grouped by the job it does. The ones we lean on most are marked with a ★.

AI models & assistants

The models

CLClaude Anthropic — primary model & agent platform
GPChatGPTOpenAI — general assistant
GEGeminiGoogle — long-context work
NLNotebookLMGoogle — source-grounded research
GRGrokxAI — real-time context
Generative & creative AI

Making things

HFHiggsfield Image · video · audio · 3D
CVCanvaDesign & brand templates
MCMermaid ChartDiagram-as-code
MRMiroVisual boards
Knowledge & productivity

Where knowledge lives

NONotion Knowledge base / source-of-truth
DRGoogle DriveDocument / evidence layer
GMGmailIntake (draft-only)
CAGoogle CalendarScheduling / time
M3Microsoft 365Office suite
GUGuruKnowledge management
JFJotformStructured intake / forms
COCoupler.ioData pipelines
TSThoughtSpotNatural-language BI
PDPDF ViewerRead / annotate
Dev, infra & web

Where it runs

CFCloudflare Workers / Pages — our own sites
GHGitHub Source control / canonical truth
NENetlifyStatic hosting
WXWixSite builder
SBSupabasePostgres backend
GDGoDaddyDomains
Automation & publishing

Moving the work

BLBlotatoMulti-platform publishing
COCoupler.ioScheduled dataflows
Worked Example

Claude, in depth

We use many models and point each at the job it does best. Claude is the one we build on most, so it is the clearest example of what "fluency" means in practice — not just chatting with an assistant, but running a family of models across a range of surfaces, from a phone to a full operating system.

The model family

One family, five points on a capability curve

Claude is not a single model but a family. They share the same design — capability paired with control — and differ in how much reasoning power, speed, and cost each brings. The skill is choosing the model that fits the task, not defaulting to the most powerful one.

Flagship

Fable 5 · Mythos 5

The most capable models, for the most demanding reasoning and long-horizon, multi-step work. Mythos 5 is a limited-access sibling of the same tier.

1M-token context
Workhorse

Opus 4.8

Highly autonomous; strong on agentic work, knowledge work, and long tasks. The everyday heavy-lifter.

1M-token context
Balanced

Sonnet 5

Near-flagship quality on coding and agentic tasks at a lower cost — the sensible default for high-volume work.

1M-token context
Fast

Haiku 4.5

The fastest, most cost-effective model, for simple and high-volume tasks where speed matters more than depth.

200K-token context
The development track

From chat assistant to agent platform

Claude started as a conversational assistant and has become an agent platform. The through-line is capability paired with control: as the models got more capable, the surfaces around them gained the ability to read files, use tools, and complete multi-step work — with review and safety kept in the loop rather than traded away. That is the shift that makes AI usable as business infrastructure, not just a smarter search box.

The product surfaces

The same model, wherever the work is

The same intelligence shows up in different places depending on how you need to work — a chat window, a coding terminal, an autonomous coworker, or an API your own systems call. Fluency means knowing which surface fits which job.

claude.ai (web)
Browser

The core chat and work app — the front door for most people.

Desktop & Mobile
macOS · Windows · iOS · Android

Claude native on the desktop and in your pocket.

Claude Code
CLI · IDE · Web · Desktop

Agentic coding in the terminal or editor — it can build, test, and run work, including headless and on the web.

Cowork
Desktop · Web · Mobile (beta)

Hand Claude a task and it works across your files, calendar, email, and the web — running in the background while you do something else.

API / Developer Platform
Cloud

Programmatic access to the models — how we wire Claude into a business's own systems.

Agent SDK
Your own infrastructure

The building blocks for custom agents that read files, run tools, and act — deployed where you control them.

Claude in Slack
Slack — Team / Enterprise

A shared @Claude teammate in a channel the whole team can steer, pick up, and redirect.

In the browser & Office
Chrome · Microsoft 365

Claude acting where the work already happens — the browser and the Office suite.

That range is the point. A business owner does not need to learn all of it — that is our job. We match the model and the surface to the task, keep a person accountable for every output, and build the result into a system the business can actually run on. See how that fits the wider practice on AI Advisory.

Want the right tools pointed at the right constraint?

The first conversation is free and helps both sides decide whether the work is a fit.

Request a Consultation