Independent practice. Manufacturing, services, and software.
I build the go-to-market systems you don't have, then hand them over.
Outbound, revenue operations, and internal tools for B2B companies. Every engagement is fixed in scope and ends with you owning what I built.
What I build
Start with whichever one sounds like you.
"We have capacity we can't fill."
Outbound installs
Infrastructure, a verified target list, the sequences, and the training to run it in house.
"We can't trust our own numbers."
CRM and RevOps rebuilds
Cleanup, a qualification model built around how you sell, and reporting you can defend in a meeting.
"My experts spend their day reading paperwork."
Internal AI tools
Software built on your own history that gets an expensive person to an answer in a minute.
Most engagements combine two, because outbound tends to expose whatever is wrong underneath it.
Why built, not rented
Buy the infrastructure. Own the intelligence.
Most of this work gets sold to you as output you have to keep buying, or as software where the useful logic stays locked inside someone else's product. I write it into a repository with your name on it, then teach your team how to change it.
Selected work
The same three, with the numbers attached.
Outbound
Two books built from scratch. Neither company hired anyone to do it.
Both were living on referral and relationship revenue, and nobody had ever gone looking for the rest of it.
Revenue operations
Most CRM work is configuration. The question is whether the numbers coming out are true.
Cleanup and enrichment so the records are worth reporting on, automation so they stay that way, and a qualification model built around how the company sells rather than how the software ships.
The part an agency does not do
A configured CRM still lies. At one company, closed-won lived on three different stage IDs across five pipelines, so any report filtering on one of them undercounted wins by 30%. At the same company, 23% of the deal table belonged to two deactivated owners with no open pipeline, quietly inflating every per-rep number. I also reported that their ad spend was not sourcing anyone, which was the opposite of what I had been hired to demonstrate. They kept the build and changed the strategy.
Internal AI tooling
Software that prices a bid against five years of what the company actually sold.
Built for a commercial glazing contractor: an RFQ scoring and pre-estimating tool, a quick-quote engine that populates takeoffs and proposals, and a customer-facing configurator that prices storefront elevations on the spot.
How it is built
The model only extracts. Scoring and pricing are deterministic code against a published rate card, so every number is auditable and I could calibrate it against jobs where the answer was already known. Where it still cannot price something it reports nothing instead of a confident wrong answer.
Start with the part that is costing you most.
Capacity you cannot fill, numbers you cannot defend, or people spending their day on work a system should be doing. Tell me which one and I will tell you what I would build and what it would take.