Get your developers shipping with AI, safely

Most teams have AI coding tools and few rules for using them. We set up the editors, the review bots, and the connections to your own systems, then train your developers on real tickets from your codebase.

Starts at
$2,500
Timeline
1–4 weeks

What you get

  • AI code editor rollout

    Claude Code, Cursor, GitHub Copilot, OpenAI Codex, or Devin Desktop (formerly Windsurf), with team settings, privacy controls, and shared rules files

  • AI code review

    An AI reviewer on every pull request, tuned to your standards and measured against your own past reviews

  • Custom MCP servers

    Connect AI tools to your internal APIs, databases, and docs, with access controls and logging

  • Hands-on training

    Your developers work real tickets with the tools, not demo apps

  • A written AI coding policy

    What code and data can go to which tools, how AI-written code gets reviewed, and who owns it

  • Before-and-after numbers

    Review time, cycle time, and defects, measured before and after so you can see what changed

How it runs

You always know what’s happening and who it’s waiting on. At each checkpoint, the next step starts only after you say it’s right.

  1. Before you sign

    A free call

    Together

    Your stack, your tools today, and where developers lose the most time.

  2. Week 1

    Assess and plan

    Us

    We look at your repos, CI, and security rules, then write down what to set up and how we’ll measure it.

    Checkpoint: You OK the plan and the security settings

  3. Weeks 2–3

    Set up and pilot

    Us

    Editors, rules files, the review bot, and any MCP servers, piloted with a few developers first.

    Checkpoint: Your pilot team approves it for everyone

  4. Week 4

    Train the team

    Together

    Hands-on sessions on your own codebase, plus the written policy.

  5. After

    Measure and tune

    Together

    A 30-day check on the numbers, and rules tuned to what the team actually hits.

Packages

Seats for the AI tools are billed by their makers. We help you pick the plan. Extra engineering time is $150/hr.

  • Developer AI workshop

    Remote half day, up to 12 developers, on your own codebase. Full day on-site in Richmond: $5,000.

    • Prompting, rules files, agents, and review habits that hold up in production

    $2,500

  • AI code editor setup

    Cursor, Claude Code, or Copilot set up for one team

    • Team policies, privacy controls, shared rules files, and MCP config

    $2,500

  • AI code review setup

    An AI reviewer on every pull request, plus a 30-day evaluation

    • We pick and tune one or two tools, measure what they catch and how noisy they are, and train your reviewers

    $3,500

  • Team AI enablement

    Setup, workshop, and 60–90 days of support for up to about 10 engineers

    • Before-and-after numbers on review time and cycle time

    $9,500

  • Custom MCP server

    Connect AI tools to your own APIs, databases, or docs

    • Most single-system servers land between $3,500 (read-only) and $8,000 (with sign-in or write access)
    • Scoped and priced in writing after a call

    Custom quote

  • Monthly AI engineering consultant

    Ongoing help keeping your team’s AI tools, review bots, and MCP servers current, with standards, reviews, and training as the tools change

    • Scoped to your team’s size and needs, and priced in writing after a call

    Custom quote

  • Paid discovery sprint

    For a build you have in mind but haven’t scoped yet. We map the work with you and write a fixed-price scope.

    • The fee is credited toward a build booked within 60 days
    • The written scope is yours to keep either way

    $1,250

A good fit if

  • Your developers use AI tools, but everyone does it differently
  • Code review is a bottleneck
  • You want AI tools to reach your internal systems safely
  • Security or compliance wants rules before AI touches the code

Look elsewhere if

Questions

Is our code safe with these tools?

We set up business plans that don’t train on your code, turn on the privacy controls, and write down which repos and data each tool can see. Your security team signs off before anyone outside the pilot uses it.

Which tool should we use?

It depends on your stack and your rules. We’ve used all the main ones on real projects, and we’ll tell you which fits and why, even if it’s the one you already pay for.

What is an MCP server?

It’s a small service that lets AI tools like Claude or Cursor use your own systems, like a ticket tracker, a database, or internal docs, with the access you allow.

Talk it through first

Book a free 30-minute call. You’ll get a written price within a day.