Your next AI project or product, delivered in 3 to 4 weeks. Pick anything on the menu.

Apps

A native iPhone app, live on the App Store.
Starting at$4,0003 to 4 weeks
Book this
UX
Flows, screens, what to cut from v1
Design
Visual system, icon, store images
Architecture
Data model, APIs, what runs where
Front end
Swift and SwiftUI, in your repo
Back end
API, database, jobs, the AI behind it
Auth and users
Sign in with Apple, user management
Billing
Subscriptions and paywalls
Analytics
Events, funnels, crash reports
Communication
Push notifications and email
Launch
TestFlight, App Review, your listing
Picture.Repair, first called Kamera, reached the App Store 10 days after its first commit. Read: The Difference Between an AI Feature and an AI Product
A production web app on your domain.
Starting at$5,0003 to 4 weeks
Book this
UX
Flows, wireframes, what to cut
Design
Visual system, responsive screens
Architecture
Data model, services, room to scale
Front end
React and TypeScript, fast anywhere
Back end
APIs, database, jobs, AI with evals
Auth and users
Sign-in, teams, roles, admin
Billing
Stripe plans, usage and invoices
Analytics
Event tracking, funnels, dashboards
Communication
Transactional email, SMS, in-app
Launch
On your domain, monitored, and yours
TinyNews opened public sign-up under four days after its first commit. Read: What a Good AI Proof of Concept Looks Like
A native desktop app your team installs.
Starting at$8,0003 to 4 weeks
Book this
UX
Menu bar or full window, shortcuts
Design
A native look, app icon, every state
Architecture
What runs locally, what syncs
Front end
Native UI that feels at home
Back end
Sync, APIs, the AI behind it
Auth and users
Sign-in, licenses, team seats
Billing
Licenses or subscriptions
Analytics
Usage tracking, crash reports
Communication
System notifications and email
Launch
Signed, installable, updates itself
Amply, a native Mac remote for Spotify, drawn pixel for pixel. Read: When to Build vs. Buy AI (The Framework I Actually Use)

Go-to-market engineering

A CRM that fills and routes itself.
Starting at$8,000a month
Book this
Self-filling CRM
Records from email and calls
Enrichment
Accounts and contacts, filled in
Intent signals
Site visits, hiring, funding, tech
Lead scoring
Your ICP turned into a score
Routing
The right lead to the right rep
Data quality
Dedupes and cleanup, always on
Integrations
HubSpot, Salesforce and your stack
Reporting
Pipeline and data health dashboards
Led HubSpot Data Enrichment from under $1M to over $10M ARR in six months. Read: From SDR to GTM Engineer: The Career Shift That's Already Happening
A tuned ICP, and a fresh list of target customers every week.
Starting at$8,000a month
Book this
ICP tuning
Your best customers, as a scorer
Sources
Data vendors, job posts, the web
Enrichment
Firmographics, tech stack, contacts
Research
Agents that research each account
Contacts
The right people, found and verified
CRM sync
In your CRM, without duplicates
Review
A human check until you trust it
Reporting
What each list turned into
LeadShogun turned anonymous site visits into personal outreach, and got me hired at Clearbit. Read: Build a company research agent that replaces 12 open tabs
Outreach that starts the moment a lead shows up.
Starting at$8,000a month
Book this
Inbox warming
Domains, warm-up, sending limits
Email sequences
First lines from real research
Messaging
Automated SMS, voice and follow-ups
Triggers
Form fills, signals and new lists
Replies
Sorted and routed to your reps
Opt-outs
Suppression lists and calling hours
CRM
Every touch logged
Reporting
What books meetings, what doesn’t
Peregrine calls new leads in about 4 seconds and live-transfers the good ones. Read: Build a personalization writer that actually sounds human

Company AI

Your own custom Claude Tag, built around your business.
Starting at$15,0003 to 4 weeks
Book this
Company content
What matters about your business
Source of truth
Access to CRM and customer systems
Collaboration
Works across teams and channels
Slack app
In the channels you choose
Permissions
Only what each person can see
Citations
A source on every answer
Evals
Tested on your team’s real questions
Analytics
What gets asked, what it can’t do
Rollout
One channel first, then the company
Led HubSpot’s Agent Hub from private beta to general availability. Read: Agents vs. Automation: How to Know Which One You Actually Need
Siloed company knowledge, ready for your agents and your team.
Starting at$15,0003 to 4 weeks
Book this
Sources
Docs, tickets, calls and wikis
Ingestion
Parsed, chunked and kept in sync
Embeddings
Vector embeddings of all of it
Search
Hybrid search, with reranking
Agent access
For your agents, over MCP or API
Citations
Every answer links to its source
Permissions
Access rules carried from the source
Evals
Scored on your real questions
Monitoring
Cost, latency and answer quality
Corpus answers from a person’s own records, with the source one tap away. Read: The Stack a Modern AI Strategy Actually Runs On
Open models on your infrastructure. Your data stays yours.
Starting at$16,0003 to 4 weeks
Book this
Assessment
Which workloads are worth moving
Privacy
Nothing sent to OpenAI or Anthropic
Model choice
Open models tested on your tasks
Fine-tuning
On your data, when it beats a prompt
Infrastructure
GPUs in your cloud, sized to traffic
Serving
An API your apps can call
Benchmarks
Cost and quality against today
Monitoring
Latency, errors and spend
Handoff
Runbooks your team can follow
At HubSpot: 98% lower model cost, with prompt engineering, caching and self-hosted, fine-tuned open models. Read: Business Flywheels: Use AI to Autonomously Improve Performance

Training

PMs who prototype, test and ship with AI.
Starting at$5,0002 weeks
Book this
Week one
Your team’s needs, your org’s setup
Week two
Up to 4 hours, on site or remote
Prototyping
In v0, Lovable or Claude Design
Customer signal
From calls and your data warehouse
PRDs
Requirements, drafted with AI
Product wiki
Kept current by watching your Slack
Living view
The product, for stakeholders
Playbooks
Yours to keep, written for your team
Group Product Manager at HubSpot, and I built UXLoops for exactly this. Read: Why AI Projects Need a Product Manager, Not a Project Manager
The perfect message to the perfect customer at the perfect time.
Starting at$5,0002 weeks
Book this
Week one
Your team’s needs, your org’s setup
Week two
Up to 4 hours, on site or remote
Personalization
Hyper-personal targeting, at scale
Campaigns
One strategy across every channel
Landing pages
How to build them, start to finish
Live content
AI classifiers tailor each visit
Context
Just-in-time customer context
Playbooks
Yours to keep, written for your team
Ran marketing at Delphi Digital and growth at Clearbit. Read: How to Train Your Team to Use AI Without Making It Mandatory
More time selling, less time on data entry.
Starting at$5,0002 weeks
Book this
Week one
Your team’s needs, your org’s setup
Week two
Up to 4 hours, on site or remote
Meeting prep
Research and a brief for each call
Insights
Structured data from email and calls
Coaching
Guided selling and call coaching
Methods
BANT, MEDDIC and the ones you use
CRM updates
Automatic, with no data entry
Prospecting
Prospects found, intent spotted
Follow-up
Next steps and follow-up, automated
I’ve sold cars at Tesla and my own products since. PhoneScreen is used by 80+ hiring teams. Read: Build an objection handler that knows the deal before your rep does