ak-labz
AI Software for Operators

Give your team back hundreds of hours with AI-native software built around your business.

For operators of 10 to 100 person businesses doing $5M to $50M. You value results, speed, and experts who understand business and build software.


What we build

AI fits in your business in two places. We build for both.

Tasks every business has

Meetings, email, documents, knowledge searches, reports. The universal grind. We pull action items from meeting transcripts and push them into the tracker your team already uses. We classify incoming email by intent and draft first-pass replies in your team's voice. We extract fields from invoices and route them to the right approver. We index your existing files and email so anyone can ask in plain English and get a cited answer. Hours back per person, every week.

Read the blueprint

Workflows specific to yours

The multi-step thing only your business does: how RFPs get answered, how orders move from inbox to invoice, how claims get triaged, how loans get underwritten, how leads get qualified. We sit with your operator and map the workflow as it actually runs. Then we build AI for the part that hurts most. That might mean extracting fields from inbound docs, drafting a first-pass response in your team's voice, classifying and routing, or pulling context from your CRM and ERP. Your operator stays in the loop where judgment matters.


How it goes

Working AI software in your team's hands in 4 weeks. Built around how you actually work.

Building AI into your business has two halves: the AI itself, and the work to connect it to your data, your tools, and the way your team actually works. Most projects fall apart on the second half. We do both.

01

Outcome alignment

Week 0 · free

A 60-minute working session. We walk through the operations, then pick one result we would commit to delivering. You leave with that outcome named, or an honest "not us."

02

Workflow teardown

Week 0

We break the workflow that produces that outcome into the activities it takes today: who does each, how often, where the time goes. The output is a one-pager you can argue with.

03

Build spec

Week 0

We spec the build on your data: what we would ship, how it lands in your stack, what each half costs. You sign off before we write code.

04

Build

Weeks 1-4

Two builders write the system. You review weekly. We integrate into your stack and run alongside the current workflow until your operator buys the output. No new dashboards to learn.

05

Retainer

Ongoing · paid

Monthly retainer once the system lands. We tune as operations evolve, surface new opportunities from the same workflow analysis, and fix what breaks. Standard model, not optional.

We work with clients across multiple projects, not one and done. The relationship grows: we know your team and your operator. The system grows: each build sits on the one before it, not next to it. The impact grows: by the second project, we're moving bigger numbers because we already know your stack.


Who is building

We have built software for hundreds of millions. Now we build for you.

We are obsessed with one thing: building software that gives operators their time back, so your team can focus on the work that actually matters. No account manager between you and the people writing the code.

K. Iwasaki

K. Iwasaki

Co-founder · Product builder / Operator

MBA, Michigan Ross · MS, UC Berkeley

Andrew Gray

Andrew Gray

Co-founder · Product builder / Architect

BS, Neuroscience & Mathematics, Boston University

Senior product leader specializing in AI software. Spent seven years at SmartNews, a Silicon Valley startup, growing from Machine Learning engineer to Director of Product Management leading the entire product team. Revamped the US app for millions of users, anchored by the AI-powered news recommender. Six years at Microsoft in business management and finance. Ships multiple AI apps on the side.

Senior machine learning engineer specializing in large language models and AI-driven products. 11 years of experience developing solutions across voice AI, image generation, and chatbots. Currently works on Instagram Reels retrieval. Previously led the personalized news recommender at a major news aggregator. Has launched two LLM-powered ventures and served as a machine learning coach.


Next step

One 60-minute call.

Free. You leave with either a scoped POC offer, or an honest "we are not the right team for this."