Rahul Mehendale

A Disruptive Innovation and Entrepreneurial Bigot who thrives in the Future of Health world across LifeSciences and Healthcare @Deloitte

[Rahul Mehendale @globalcitizen99]

https://t.co/jwOO8IK8bm

98 real patients. In 44 pre-visit AI summaries reviewed, doctors said 75% helped preparation and 57% might change their approach. AI’s next job: make the appointment better, not shorter. @AdamRodmanMD @alan_karthi @EricTopol #HealthAI #PrimaryCare

[Rahul Mehendale @globalcitizen99]

https://t.co/1B0UdL1iSr

Healthcare outsourcing has an AI paradox: better automation can mean lower vendor revenue. Sagility describes up to ~2% annual deflation from pricing, offshoring and automation, including AI. Who keeps the savings? @pfersht @bobjherman @nikillinit #HealthcareAI #AI

[Rahul Mehendale @globalcitizen99]

https://t.co/6SWPz5wiTd

Mount Sinai tested 20 AI models: one safety reminder cut unsafe choices from 16.6% to 10.1%, a 39% relative drop. Yet ~1 in 10 test responses remained unsafe. The answer is layered safeguards, not prompts alone. @mahmud_omar_md @EricTopol @peteratmsr #HealthAI #AISafety

[Rahul Mehendale @globalcitizen99]

https://t.co/A2W8rt1NsI

“Math is destroyed” may be a preview of biology. $1.8B is now backing predictive biology. Next: simulate before synthesizing, predict before trials, model before care. The moat may be the reality loop. @DaphneKoller @alexrives @demishassabis #AI #Biotech

[Rahul Mehendale @globalcitizen99]

https://t.co/hZvfsg9Tr3

A bed-making robot hit 91.9% average success on individual steps. The healthcare opportunity isn’t “replace nurses.” It’s unbundle the shift: automate bounded chores, recover from failure, measure labor minutes returned. Reliability before robots-at-scale. @Ken_Goldberg #PhysicalAI #HealthTech

[Rahul Mehendale @globalcitizen99]

https://t.co/ygohPowlBR

Utah is testing a powerful AI model: autonomy that must be earned. Doctors approve the first 100 acne prescriptions, then review shifts backward and eventually toward sampling. @luiswenus may be testing more than AI prescribing. He is testing new care economics. #HealthcareAI #DigitalHealth

[Rahul Mehendale @globalcitizen99]

https://t.co/c1JO3BpyrL

72% of health orgs report unapproved AI. As agents begin taking actions, access control is no longer enough. Every agent needs an identity, bounded permissions and an audit trail. Autonomy should expand only when performance earns it. @jhalamka #HealthcareAI #AgenticAI

[Rahul Mehendale @globalcitizen99]

https://t.co/cHIzH8B7Kp

300M people ask ChatGPT health questions weekly. Healthcare’s front door is moving from the appointment to AI. The next winner may not replace doctors, but own the trusted handoff: Listen -> Context -> Guide -> Escalate -> Learn. #HealthcareAI #DigitalHealth

[Rahul Mehendale @globalcitizen99]

https://t.co/vDgLnIXNMZ
Drug development can take 10+ years, cost up to $2B and still fail >90%. ARPA-H’s SURPASS attacks the architecture: simulate -> run -> learn -> adapt. The AI opportunity may be a clinical-trial system that gets smarter with every study. #ClinicalTrials #AI

[Rahul Mehendale @globalcitizen99]

https://t.co/fFRWFXPXMZ

AI’s biggest impact may not be automating the company. It may be making today’s company obsolete. Biology offers a glimpse: tiny interdisciplinary teams + AI + automated experimentation are radically compressing time.

Amgen reports 3× faster protein engineering and 50% shorter discovery timelines.

#AI #AgenticAI #FutureOfWork #Biotech

[Rahul Mehendale @globalcitizen99]

https://t.co/9VDK0bwaxi

AI may be working perfectly and still make healthcare worse. BCBS estimates coding intensity added $942M while care largely did not change. The lesson: AI will optimize what we pay for. Make the objective better outcomes, not more codes. #HealthcareAI #ValueBasedCare

[Rahul Mehendale @globalcitizen99]

Technology can make disruption possible. It cannot make the ecosystem willing. Reflections from the Christensen Institute Innovation Summit at Harvard on AI, Jobs to Be Done and why the basis of competition is shifting.

https://t.co/ch97ai3FDu

[Rahul Mehendale @globalcitizen99]

https://t.co/xAlwzDGYGv

40% of Roche pipeline decisions already had tracked AI/computational input. Now it is building autonomous labs. The real moat may not be the model, but the loop: predict -> experiment -> measure -> learn. @DaphneKoller, does proprietary biology become the advantage? #DrugDiscovery #AI

[Rahul Mehendale @globalcitizen99]

https://t.co/jrNK3ogXYx

Do you really need to refill that job? UChicago Medicine automated routine RCM work and moved people to harder cases. The AI workforce question should be: Refill, redesign or hold? Redesign the work before the org chart. #HealthcareAI #FutureOfWork

[Rahul Mehendale @globalcitizen99]

https://t.co/eXDMg0KeMI

What happens when AI uses healthcare apps for us? The moat may shift from the screen to context, permission and trusted execution. If another agent owns the patient relationship, are you still the product, or have you become an API? #HealthcareAI #AgenticAI

[Rahul Mehendale @globalcitizen99]

https://t.co/fVuXXqYLIv

Are we waiting too long to look for pancreatic cancer? Mayo AI used routine data from nearly 40K people to flag risk up to 3 years early. The opportunity: risk first, targeted tests second. The model matters. The workflow may matter more. #HealthcareAI #Cancer

[Rahul Mehendale @globalcitizen99]

https://t.co/C5Ohx3WqwP

AI can have positive ROI for one department and negative ROI for healthcare. BCBSA estimates increased complex coding added $942M in spending without corresponding evidence of more care. Measure system value, not task productivity. #HealthcareAI #HealthEconomics

[Rahul Mehendale @globalcitizen99]

https://t.co/rJcOQdQjax

950 AI agents. 21 hours. 200,000+ genes. 20 leads. The interesting breakthrough is bigger than one enzyme: scientific search is becoming massively parallel, while humans remain the experimental truth gate. #Biotech #AI

[Rahul Mehendale @globalcitizen99]

https://t.co/zeron2zvey

Anthropic + OpenEvidence are taking medical AI to ~100 countries. The breakthrough is not just access. It is localization: global evidence + local disease + available resources + clinician judgment. AI scales expertise. Context makes it useful. #HealthcareAI #GlobalHealth

[Rahul Mehendale @globalcitizen99]

https://t.co/AqsB5Bxt4U

Oura targets a $15.6B IPO valuation with 5.7M paid members expected. The bigger health-AI story is not the ring: Sense -> Understand -> Guide -> Act -> Learn. The moat is turning continuous signals into better decisions. #HealthcareAI #Wearables