AI agents and automation for growing B2B teams. We build the tools, connect the systems and help your people put AI to work.
Claude Code
Codex
Cursor
Claude
OpenAI
Gemini
n8n
Make
Zapier
Supabase
Slack
ClickUp
Salesforce
Zoho
SAP
Power BI
PythonAutomations that work across 9,000+ app integrations and 66,000+ triggers and actions.
Custom apps, automations, AI agents, system integrations and hands-on training, built by one partner from first idea to everyday use.
Client portals, internal apps, dashboards and AI products built with Claude Code, Codex and Cursor. From the first useful version to a tested, maintainable application in your accounts.
A shorter path from business need to working software.
From a new request to a ready-to-work team: intake, setup, approvals and handoffs connected around your existing process.
Less admin. Fewer forgotten steps.
Read documents, organise knowledge, prepare responses and triage requests, with review and escalation where judgement matters.
Less repeated checking. More useful context.
Connect finance, sales, people and operations. Practical experience across Zoho, SAP, Salesforce, Tally and enterprise ERP environments informs the field mapping, permissions and exceptions.
Cleaner records. Fewer manual reconciliations.
Practical AI workshops, agentic coding sessions and training around your team’s actual work. Learn to use the tools, evaluate the output and maintain what you build.
Useful skills your team keeps after the session.
As teams grow, manual work grows faster. These are the four problems founders and operations leads bring to us most. Hover or tap a card to see what changes.
Every new project starts with an hour of copying briefs, building tasks and chasing owners before any real work begins.
See the fixOne form now creates the workspace, tasks, owners and team handoff automatically.
Each department keeps its own spreadsheet, so weekly reviews start with an argument about whose figures are right.
See the fixFinance, delivery and capacity data reconciled into one live leadership view.
Handoffs depend on someone remembering every step, so follow-ups stall and opportunities quietly go cold.
See the fixTriggers assign owners, send reminders and flag anything that gets stuck.
Good work is buried across folders and file names, so people start from zero instead of reusing what already works.
See the fixAI tagging makes every past asset searchable by idea, not by file name.
Six systems built for real teams, from a one-form project setup to a five-system leadership view. Pick a project below to see what changed.
About an hour of copying, setup and handoffs removed from every new project. The team starts with the same complete structure, every time.
Leadership reviews capacity, delivery, churn and margin in one place, instead of debating which department spreadsheet is right.
Designers search past work by idea, angle or format and jump straight to its Figma source, instead of an afternoon of digging through shared drives.
An unattended job now prepares, archives, tracks and announces each month’s client workspaces: about 1,000 a year at that rate.
Client requests, time-off cover, data syncs, digests and follow-ups run automatically, with error handlers and a watchdog that flag failed or missing runs.
A paid-media checklist became an agent review across Google, Meta, LinkedIn and Bing, with the source data and thresholds attached to every finding.
Case studies on this page are selected work from Yash’s career before First Principles. Client names are confidential; screenshots use illustrative data where noted.
That’s the first principles approach, and it’s how we build every workflow, app and agent for your team.
“I help growing teams turn operational friction into useful software, and give their people the skills to use it.”
Connect with Yash on LinkedInYou own the code, accounts and documentation. We work within your access controls and can work under your NDA and security review.
We agree a baseline such as time per task, recurring volume, errors or rework, then compare it after launch. Running costs are part of the assessment.
Documentation and a handover come with every build, and workshops help your team use and maintain the systems themselves.
n8nMake.comClaude CodePower BIPythonUSD per month. Every retainer includes maintenance of live systems. Tool and API costs are separate.
Keep what’s live running and improving.
A sprint’s worth of app building and automation each month.
Dedicated capacity across your workstreams.
Need one defined change? Discovery Teardowns and Automation Sprints are scoped and quoted after a call.
Need a focused fix? Ad-hoc work is $52/hour. Scope, capacity and response expectations are agreed before work starts.
Feedback from founders and teams in the US and Israel.
Custom apps, client portals, internal tools, AI agents and connected workflows. We also integrate ERP and business systems, and run practical AI workshops. Bring the business problem; we’ll work out the simplest useful solution.
Usually, no. We start with the tools you already use, then connect or extend them. Custom code and new platforms earn their place only when they solve a real limitation.
We agree a baseline such as time per task, recurring volume, errors or rework, then compare it after launch. Recovered time is capacity; cash savings depend on what your business does with it. Running costs are included in the assessment.
You own the code, accounts and documentation. We work within your access controls, request only the access needed and can work under your NDA and security review.
You receive documentation and a handover. Retainers cover monitoring, fixes and ongoing improvements with agreed priorities and capacity. Workshops can help your team learn to use and maintain the systems themselves.
Bring an app idea, a workflow or a team learning goal. We’ll map a useful first step together.
TypeScript
Brief & documents attached