AI services
Every engagement is a named result with a fixed scope, agreed in writing before work starts. The first conversation is free. The assessment is the usual second step, and its fee is credited in full to any build within 90 days. Package fees are quoted after the assessment, because the scope depends on what it finds. Hourly consulting is available at $200/hr for anything that doesn't fit a package.
Pricing and engagement terms →- Who it's for
- Anyone deciding whether to spend on AI, or how.
- What you get
- 30 minutes with the person who would do the work, not a salesperson
- A short written note afterwards: what I heard, whether AI fits, the first step I'd take
AI Readiness Assessment
A ranked map of where AI pays off in your business, with the cost, the risk and the data rules for each, plus one working result.
- Who it's for
- Companies new to AI, or with AI tools nobody uses well, in any industry.
- What you get
- Interviews with the people who do the work, and a look at the tools and documents you already have
- Three to five opportunities, ranked by payoff against effort, each with a cost range, the risks, and which data it may and may not touch
- The data and approval guardrails I'd put in place before anything connects
- One working proof: a configured workspace or one tested skill your team can use the same day
- A readout call, and a written report you own
AI Workspace Setup
Your team's AI tools configured and connected to your documents, with access rules and usage guardrails, ready to use.
- Who it's for
- Teams with ChatGPT, Claude, or Copilot licenses and no shared setup.
- What you get
- Accounts, workspaces and roles set up so the right people see the right things
- Your reference documents connected, so answers come from your material instead of the open web
- Usage and spending guardrails, and a one-page "what to put in, what to keep out" rule for the team
- A walkthrough for the people who will use it
MCP & Tools Integration
The model reads and acts on your CRM, tickets, documents and databases through connectors with rules, so it works from your real data instead of guessing.
- Who it's for
- Companies whose AI answers are generic because the model can't see their systems; software teams adopting agents.
- What you get
- MCP servers for the systems you name, with read and write permissions set per tool
- Every action the model can take listed, and the ones that matter gated behind a person
- Logs of what the model read and did
- A test set that proves each connector before it goes live
Skills & Prompt Library
The ten things your team asks AI to do most, turned into tested, reusable skills, so everyone gets the same quality.
- Who it's for
- Sales, support, marketing, admin and operations teams whose AI output varies by who wrote the prompt.
- What you get
- A short list of the tasks that cost the most time, agreed with the team
- A skill for each: the instructions, the examples, the format, and the checks
- Each skill tested against real examples from your work before it ships
- A shared library the team opens from the tools they already use, and a way to add the next one
- Who it's for
- Any team with a process that is the same every time and still done by hand.
- What you get
- One process, mapped step by step, with the human decision points marked
- The automation, built so the dangerous steps cannot run without a person's approval
- A before-and-after measurement: time, errors, cost
- Monitoring, so you know when it stops or drifts
- Who it's for
- Teams already using AI in front of customers or in documents that matter.
- What you get
- A written definition of "good" for each output, agreed with the people who judge it today
- A test set built from your real examples, and a score for the current setup
- The changes that raise the score: skills, data connections, model choice
- The scorecard re-run on a schedule, so quality is a number you can watch
Team Training
The people who will use AI know what it is good at, what it is bad at, and how to get consistent results from it in their own work.
- Who it's for
- Any team, technical or not. Software teams get a version about agents, MCP and the method.
- What you get
- A half-day session or two shorter ones, on your work, not generic examples
- Hands-on with your own tools and skills, not slides
- A one-page reference the team keeps
- A follow-up a few weeks later to answer what came up
Custom AI product build
Your AI product, built to run in production: human approval where it matters, cost ceilings, metering, and a single accountable builder.
- Who it's for
- Companies with an idea or a prototype that has to become real, safe and billable.
- What you get
- Scope, plan and price in writing before work starts
- The product, built in small reviewable steps with tests and a written record of every decision
- Human gates on every action that spends money, contacts someone, or changes a record
- Deployed to your cloud account, with the keys in your hands
AI partner retainer
An AI lead on call: ongoing tuning, new skills and automations as they earn their place, cost monitoring, and someone accountable when something drifts.
- Who it's for
- Companies that have AI running and want it to keep getting better without hiring for it.
- What you get
- A monthly plan of what gets built or tuned, agreed with you
- The output scorecard reviewed every month, and the fixes made
- Cost and usage watched, with ceilings set and kept
- A named person to call