Norton Lam · Minneapolis, MN · updated June 2026

AI Work

I help businesses and nonprofits figure out where AI can actually help, then build the tools to do it. Here's what I'm working on.

Start a Discovery Session → See how it works

New: 10 AI pro tips most users miss, from 900+ real working sessions, with copy-paste prompts. Read the tips →

Selected projects

My own AI products, and work I've done for clients.

FlagshipMulti-agent

The Business X-Ray

An X-ray for your business: it shows what you can't see from the inside. Every owner has blind spots, and your data knows things you don't. A committee of AI executives goes through your finances, operations, marketing, tech, and contracts all at once, a Facilitator moderates the debate to determine what matters most, and you walk out with one prioritized plan of what to fix first, plus the full transcript of how every call was made. It works for businesses and nonprofits, and it's the most thorough thing I've built.

For-profit
Nonprofit
Multi-agent

For businesses → For nonprofits →

Client onboarding● Live

AI Discovery Session

A client interview, run by AI. You answer on your own time, and it asks about your business, your workflows, what's painful, and what you want. Then it sends back a summary, the full transcript, and a context file you can hand to any AI tool. No scheduling, no prep calls. You get a clear picture before we talk.

See how it works → Start a Discovery Session ↗

Multi-agent77+ agents

Engineering Org — AI-Native Engineering Teams

Nine AI engineering teams that cover the whole software process: product and analysis, architecture, building, testing, accessibility, infrastructure, UX, SEO, and Cowork tooling. That's 77+ senior-level agents. They work with your engineers or on their own, and they produce real work your team can use. Set up for your stack, and consistent from project to project.

Learn more about the Engineering Org →

Open sourceModel-agnostic

Voter Research — AI candidate research prompt

A free, open-source prompt that turns any capable LLM into a non-partisan election researcher. Point it at any race, in any district or country, and it produces sourced candidate memos, a fact-checked claim ledger, and an issue comparison table. Local races get almost no fact-checking coverage; this makes any AI chat tool do that work on demand.

Learn more → View on GitHub ↗

Nonprofit opsHuman in the loop

AI grant-writing assistant

Grant research, writing, and reporting for a STEM nonprofit. Funders want personalized applications, so it helps with that instead of churning out generic ones. It saves the Executive Director hours per grant, but she always reviews it before it goes out.

Prospect researchPython

Funder discovery pipeline

Finds likely grant and foundation funders for a nonprofit's fundraising team and ranks them, with a shared page for checking the matches. Before, that research was difficult and time consuming. Now it is a simple scan to ensure there is a fit.

SEO contentEditorial automation

SEO blog content engine

An SEO blog aimed at life-science researchers. Every post is accurate, research-use-only, and tied to a real product. The pipeline writes the drafts, sends them for review, and posts them, so it takes a minimal amount of the moderator's time.

Technical SEO3,000+ pages

SEO meta engine, 3,000+ product pages

SEO titles and descriptions for a chemical supplier's whole catalog, over 3,000 pages, written and posted by a script. It would have taken weeks by hand. Now it runs in minutes and can be rerun anytime.

AutomationAPI integration

Lead-generation pipeline + plugin

Pulls biotech and pharma prospects from open science databases and drops them in a spreadsheet for review. It was packaged as a plugin so anyone on the team could run it with very little effort.

AccessibilityWordPress

Website redesign + rebuild

A full rebuild of a nonprofit's website on WordPress. Fixed the accessibility (WCAG 2.2 color and contrast), restructured the pages for SEO, and ran UX, accessibility, and QE checks along the way.

Start here

Readiness

Is my business or nonprofit ready for AI?

Ten ranked signs: six that say go, four that say fix this first. Based on real Discovery Sessions.

Read the readiness guide →

Diagnostic

A full look, or another targeted fix?

Ten symptoms that say the problem is bigger than the department complaining about it, ranked by how strongly each one predicts it.

For businesses → For nonprofits →

Local

AI consulting in the Twin Cities

Who I work with, how a project goes, and why I build it and hand it off so you own it. Twin Cities based, remote is fine too.

Twin Cities AI consulting →

Process

What actually happens in a Discovery Session

A plain walkthrough of the interview and the three documents you get.

See how it works →

Craft

10 AI pro tips most users miss

Field-tested habits from 900+ working sessions, ranked by return on effort. Copy-paste prompts included.

Read the pro tips →

Experiments & Building

CoworkDaily driver

Cowork — AI desktop automation

Anthropic's Cowork mode. I use it for file management, making documents, and multi-step workflows across Dropbox, Google Drive, Slack, and more. This site and the Business X-Ray were both built and run in it. It's changed how I work every day.

ScheduledDaily

Daily Brief

A daily job that scans my Cowork projects, emails, and logs, and prioritizes what to do next based on the next biggest client impact.

Local modelsOff-API

Local open-source models

Running Gemma, Qwen, and GPT open models on my own machine to see what they can do off the API. Useful for privacy-sensitive work and offline tinkering.

Get startedDiscovery Session

Want to see where AI fits for your organization?

I work with small and mid-size businesses and nonprofits. The easiest way to start is a Discovery Session, a short AI interview that exposes where you're losing time and money. Or just email me if you'd rather talk.

Start a Discovery Session → aiconsultant@nortonlam.com