Consultation with me is a paid working session: architecture, vendor choice (Retell vs LiveKit, cascade vs speech-to-speech), and a written build plan, so you do not spend a quarter on the wrong voice or RAG stack.
The intro call is free. Consultation is what happens when you need a document your CTO or founder can act on the same week: stack, sequencing, and what not to build.
I have sat in the Retell-versus-LiveKit conversation more than once. Sometimes the answer is Retell plus GHL. Sometimes the answer is do not use AI here, put a form and a human on the exception path. Named vendors with reasons beat a vision deck.
If the plan is clear, I can execute it. You are not forced to retain me for the build. Fractional reviews and written tickets for an internal team are a normal outcome.
Who it is for
- Founders choosing between Retell, Vapi, and a custom voice stack
- Teams whose agent hallucinates policy or cannot call tools reliably
- Operators who need a two-week plan, not a six-month vision deck
What you get
Written recommendation
Stack, sequencing, and what not to build. Named vendors with reasons, including "do not use AI here."
Risk list
Latency, tenancy, 2FA and portals, evals, and compliance landmines specific to your domain.
Optional build follow-on
If the plan is clear, the same person can execute it. See full-stack, backend, agentic, or voice services.
How I work
Intake
A 30-minute call or a written brief. No slide theater.
Review
Code, call recordings, or architecture. Whatever you can share under NDA.
Decision memo
A short document your CTO or founder can act on the same week.
What this looked like before Hume at Use Cortana
Callers felt the delay on the LiveKit cascade. The consultation was not "add another TTS vendor." It was leave the cascade, move to Hume speech-to-speech, keep Twilio for telephony, and orchestrate Frontend to Restack to Python. That sequence became the production path.
Proof
- Voice stack migration advice that became Hume S2S at Use Cortana
- Insurance automation scoped as address to scrape to multi-carrier, not a generic RPA pitch
- Home-service and realtor funnels scoped as CRM plus Retell, not a custom telephony rewrite
Case studies
- Vibe coding Claude + Cursor setup
- Vibe coding Loom demo
- When to leave a voice cascade
- HVAC and roofing playbook
- LLM evals in production
Stack and docs
Related
Common questions
Is the intro call free?
Yes. A short intro to see if I am the right hire. Consultation is a paid follow-on when you want a documented architecture or vendor decision.
Will you implement after consulting?
Often. I consult and build. You are not forced to retain me for the build if you have an internal team.
Can you advise a team I already have?
Yes. Fractional architecture reviews, pairing on RAG or voice, and written tickets your developers can pick up.
What do I get at the end of a consult?
A short decision memo: stack, sequencing, what not to build, and named vendors with reasons. Enough for a founder or CTO to act the same week.
How long is a typical consultation?
A focused review is often a few paid hours after the free intro. A written discovery sprint is longer when there is a codebase or call recordings to read. Typical rate is $20.00/hr.
Do you sign an NDA?
Yes, when you need one before sharing code, recordings, or customer data. Send it with the brief.
What should I send before the call?
A one-page brief helps: users, the job to be done, current stack, and what already failed. Recordings or a repo under NDA are better than a pitch deck.
Is this only for AI products?
No. I also review full-stack and backend plans: quoting systems, CRMs, and whether a no-code stack should be replaced. AI is one option, not the default.
Can you review a vendor or agency proposal?
Yes. I will say if Retell, a custom LiveKit build, or no AI at all is the honest path, and where the proposal is padding timeline or stack.