> ## Content Index > Fetch the complete content index at: https://intelligence.artificialrealitymedia.com/llms.txt > Use this file to discover other available public pages before exploring further. # How a holistic health practitioner cut client intake from hours to 45 minutes - URL: https://intelligence.artificialrealitymedia.com/ai-client-intake-health-coaching/ - Published: 2026-09-02T12:00:00.000Z - Updated: 2026-09-02T12:00:00.000Z - Description: Dr. Nikki Siso rebuilt her health intake as fifteen Claude skills: forty-five minutes, a ten-page report, and clients who disclose more to the machine. - Author: Galina Fendikevich - Tags: Health & wellness, #For AI Buyers, #What's working, #Import 2026-09-03 05:20 A client intake that took hours by hand now takes forty-five minutes and produces a ten-page report within minutes of the client finishing it. Dr. Nikki Siso, a certified holistic health practitioner who runs Conscious Kitchen near Austin, rebuilt her assessment as a set of fifteen [Claude](https://claude.ai/?ref=intelligence.artificialrealitymedia.com) skills and reports it is more thorough than the version she ran herself. She also suspects some clients tell it things they held back from her. ## What actually changed in the intake? Time first, then depth. Siso describes an intake that used to consume hours per client and still missed things. > "Before AI, it was taking me hours to go through with each client. It would never be as thorough as what it is now." What she reports now: > "The ability to really clearly know the root causes and identify them within forty-five minutes — no way before. No way. This now moves way faster. It saves me hours of work." > "They get a report within minutes of completing the assessment. Ten-page report explaining exactly what low-hanging fruit they can work on right away." Those are Dr. Nikki Siso's figures and her own judgment of quality, given to Artificial Reality, the AI buyer-evidence podcast hosted by Galina Fendikevich. They are not an independently measured outcome. ## Why does she say the assessment is more thorough than she was? Breadth, not intelligence. Her claim is that the system holds more disciplines at once than one practitioner can train in. > "There's no way one human could know all of the facts, even in one discipline. I now have access to knowledge from across fifteen different disciplines that I now know work, because I've tested them on my own body." The disciplines she names as encoded in the system include functional toxicology, integrative neurology, IFS parts work, trauma-informed coaching and polyvagal theory. She also says the build gave her access to peer-reviewed studies out of Japan and Sweden, and to agents trained in functional toxicology. Her comparison with the alternative: > "If you were to specialize in one, like cognitive behaviour therapy, that's psychology training for a decade — like nine years of studying — to really fully understand, to have read every single peer-reviewed study out there. AI can do that in minutes." ## What she deliberately left out One design decision shapes everything the assessment produces, and a reader should have it before the output. Siso directed her system not to draw on sources approved by regulators or by pharmaceutical companies, because she did not want material she considers to carry a vested interest. Artificial Reality reports that as her stated design decision rather than endorsing or disputing the reasoning. It is a significant editorial choice about what evidence the system may consider, and it shapes what comes out. If you are evaluating an assessment built this way — hers or anyone else's — ask what the system is allowed to read before you ask what it concluded. ## Do clients tell a machine more than they tell a person? In Dr. Nikki Siso's experience, some do, and her explanation is the absence of a witness. > "I think because there was a human element, maybe there's a little bit of a fear of judgment, of really sharing what is happening. Some clients would keep things a little bit more secretive." She relays one client's account of the assessment: > "I had a client go through this and they're like — literally within three questions, it got to the root of my belief system that's been challenging me my whole life. She's like: I worked with therapists my whole life, they didn't get to this." That is a single client account relayed by the practitioner, not a measured result. Artificial Reality treats it as a hypothesis worth testing rather than a finding. ## The four choices that made it work None of the load-bearing work in Siso's build was technical. Four of her instructions did most of it: - **Compassion trained in on purpose.** "I had to train it, first of all, to really be compassionate. Which is an interesting concept, right?" - **Agreeableness trained out.** Her diagnosis is that the model is "trained to be very pleaser naturally," so she trained it "to be more direct, but also be gentle." - **What is working first, corrections after.** A list of faults is "overwhelming," so the system leads with what the person is doing right and then raises what needs addressing. - **No pushing.** "If someone doesn't want to answer a question, don't push, just flag it in my report." That last instruction is the division of labour in one sentence. The system collects and notices. The practitioner decides when to press. ## What to take from this if you run a practice Automate the intake before you automate anything client-facing. Intake is the most repetitive, most documented and easiest-to-specify part of the work, and it is the one stage where Siso found that a person in the room may be a cost rather than a benefit. Then write down the two things she wrote down before building: which sources the system may use, and what it does when a client declines to answer. Those are judgment calls about your own practice, not prompts, and no vendor can make them for you. ## Where this comes from S1E7: How a non-tech founder built a 15-agent AI business on Claude — the full interview with Dr. Nikki Siso. Listen or watch: [YouTube](https://www.youtube.com/watch?v=R%5FzWAgAEIMw&ref=intelligence.artificialrealitymedia.com), [Spotify](https://open.spotify.com/episode/08s9u7LSBTtGGx94b4z2MZ?ref=intelligence.artificialrealitymedia.com) or [Apple Podcasts](https://podcasts.apple.com/us/podcast/how-a-non-tech-founder-built-a-15-agent-ai/id6780699025?i=1000777283240&ref=intelligence.artificialrealitymedia.com).