Most dental groups approach AI as a procurement decision. Which tool, which vendor, which price. Dr Sonia Szamocki argues that is the wrong variable entirely.
Sonia is founder and CEO of 01Health. Before building clinical AI infrastructure from scratch, she was an A&E doctor, trained in medicine at Oxford, and worked at BCG and BCG Digital Ventures. She has navigated this problem from inside elite strategy through to the messy reality of building a clinical AI platform, and she has a framework for AI adoption that almost nobody in dentistry is articulating.
In this episode, Dr Randeep Singh Gill and Sonia get into why AI readiness is an organisational design question, not a technology one. The core argument: the decisive question is not which AI you buy, it is whether your organisation can absorb what AI delivers at all.
They cover why you start with the problem and not the tool, the difference between deterministic and probabilistic systems and why it breaks the governance most groups rely on, why between 50 and 90% of real AI work is evaluation rather than building, and how to introduce AI without triggering the change fatigue that kills adoption. Sonia makes the case for building from within rather than doing AI to your organisation, explains how to evaluate vendor claims independently, and argues that the heaviest lifting in adoption belongs to the vendor, not the customer. She walks through the four questions every DSO should ask a vendor before signing, why you should build on imperfect data rather than wait for perfect foundations, and what the UK can learn from the point-solution sprawl that has already burned more mature US markets.
Essential listening for DSO operators, dental practice owners, dental AI founders, and healthtech investors trying to separate durable advantage from expensive distraction.
In this episode:
Why most dental groups are asking the wrong question about AI
Why you start with the problem, not the tool
The difference between deterministic and probabilistic systems, and why it changes governance
Why evaluation, not building, is now 50 to 90% of the real work
How to introduce AI without triggering the resistance that kills adoption
Why building from within beats doing AI to your organisation
How to evaluate AI performance claims independently
The proprietary data layer, and who is positioned to extract value from it
Why you should build on imperfect data instead of waiting for perfect foundations
Why the vendor should carry the heaviest load in adoption, not the customer
The four questions every DSO should ask a vendor before signing
What the UK can learn from US point-solution sprawl
Chapters:
00:00 The wrong variable: why most groups misframe AI
02:02 Start with the problem, not the tool
05:15 Deterministic vs probabilistic systems
06:13 Why evaluation is now 50 to 90% of the work
09:22 Change fatigue and the resistance that kills adoption
11:49 Who restructures care vs who just automates
14:29 Evaluating AI performance claims independently
16:13 The proprietary data layer, and who owns it
24:49 Building a data strategy on imperfect foundations
26:30 The vendor question: solution or problem transfer
29:03 The four questions every DSO should ask a vendor
32:46 Build versus buy, and the cold start problem
34:32 What the UK can learn from the US market
38:17 Lightning round
About the guest:
Dr Sonia Szamocki, founder and CEO of 01Health.
Website: https://01health.ai/uk
LinkedIn: https://www.linkedin.com/in/dr-sonia-szamocki/
01Health LinkedIn: https://www.linkedin.com/company/01health/
Instagram: @01.health / @aerox.health
Read the full written analysis: https://www.techdental.com/insights
Connect with TechDental:
Host Dr Randeep Singh Gill: https://www.linkedin.com/in/drrandeep/
Web: https://www.techdental.com
Email: [email protected]
🎧 Spotify: https://bit.ly/41UsqRO
🎧 Apple Podcasts: https://bit.ly/41pKL9b
📩 Subscribe to the TechDental newsletter: https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7320560217455271939