Andrew RadosevichPhase 1: Readiness Assessment
Before any money changes hands, the consultant scores your data accessibility, workflow documentation, leadership alignment, internal champion availability, and use-case clarity. A score of 75+ means you can move directly to a build sprint; under 50 means foundation work pays back faster than tooling. Most reputable consultants offer this assessment for free.
Phase 2: Paid Diagnostic
The Diagnostic is the most important deliverable in the entire engagement. Over 2–3 weeks, the consultant interviews stakeholders, reviews systems, maps the top 2–3 workflows where AI will deliver measurable ROI, and produces a written 90-day plan with named owners, vendor recommendations, and KPI targets. Expect to pay $5,000–$10,000 fixed-fee.
Phase 3: First Production Workflow
Pilot engagements like radstart take a single high-priority workflow from the Diagnostic and ship it to production in 30–60 days. Deliverables: working integration with your existing tools, team training, runbook for ongoing operation, and a measurement dashboard. Fixed-fee, typically $7,500–$15,000 depending on complexity.
Phase 4: Expansion & Advisory
After the first workflow proves out, most SMBs continue with monthly advisory ($3,000–$8,000) to expand to 2–4 more workflows over the next 6 months. Some prefer to take the playbook in-house after pilot. Either is correct. The goal is a working AI operating layer, not a permanent consulting relationship.
Common questions
How long does the whole process take?
From first call to first production workflow live: typically 60–90 days. Full operating-layer rollout (3–4 workflows): 6 months. Most SMBs see ROI from the first workflow within 60–120 days of going live.
Can we do just the Diagnostic without committing to ongoing work?
Yes, and reputable consultants will agree to this in writing upfront. The Diagnostic produces a standalone 90-day plan you can execute internally if you have the capacity. There should be no pressure to continue if it isn't the right fit.
What if we don't have an internal champion?
Then identifying and protecting time for one is the first deliverable of the engagement. AI implementations without an internal owner fail, no consultant can fix that for you. We won't ship pilots into teams without a named, time-protected champion.