Healthcare AI Guy Weekly | 10/29

HLTH 2024 recap, AI business model benchmarks, Researchers say an AI-powered transcription tool invents false content, and more!

Good morning, readers —

It’s been another busy week in healthcare AI. Let’s dive right in!

  • HLTH 2024 recap

  • AI business model benchmarks

  • Researchers say an AI-powered transcription tool invents false content

  • 8 new tools/partnerships, 7 funding updates & link-worthy content

Read time: 6 minutes

Our Picks

Highlights if you’ve only got 2 minutes…

1/

HLTH 2024 recap

To no suprise, AI took center stage at one of healthcare's largest innovation conferences — HLTH 2024. From major product launches to sessions on ethics and safety, AI dominated the conversation. Microsoft's Chief Scientific Officer, Eric Horvitz, described this moment as “remarkable” for AI in medicine, calling it an exciting time for technology, especially with the rise of administrative applications. Here are some of the conference AI highlights:

  • Atropos Health unveils chat-based AI co-pilot for real world evidence: Atropos Health, a real-world data platform, announced that ChatRWD, a genAI solution that produces high-quality publication-grade real-world evidence, is now more broadly available to clinicians and researchers. (link)

  • Amazon One Medical rolls out AI tools: Amazon One Medical, working with the online retail company's technology teams, developed AI tools for its clinicians to help them save time. (link)

  • Blue Shield of California, Salesforce to streamline prior auth: BCBS and Salesforce are co-developing a tool to allow physicians and patients to receive prior auth answers in near-real-time, even during visits. (link)

  • CHAI developed “nutrition label” for AI solutions: The Coalition for Health AI – published its draft frameworks for certifying independent Health AI Assurance Labs and standardizing the output of these labs with CHAI Model Cards. (link)

  • CopilotIQ and Biofourmis merge to create end-to-end platform: We included last week but both merged to create “the first end-to-end platform” for delivering in-home care from pre-surgery to acute, post-acute, and chronic condition management. (link)

  • GE HealthCare launched an AI Innovation Lab: GE Healthcare to accelerate progress across areas like clinical decision-making, cancer recurrence predictions, and model training for medical imaging with the Innovation Lab. (link)

  • Luma Health debuted multi-modal AI tool: Luma introduced its LLM-powered Spark solution, which unlocks new capabilities like automated fax processing and conversational chat for phone calls. (link)

  • Suki to partner with Zoom: Suki is bringing its AI documentation capabilities to Zoom’s telehealth platform. (link)

  • Nvidia, Aidoc collaborate to develop guidelines to speed up AI adoption: Aidoc and Nvidia to develop a free, open-source guide on the implementation of AI called BRIDGE: the Blueprint for Resilient Integration and Deployment of Guided Excellence for AI adoption. (link)

2/

AI business model benchmarks

Bessemer Venture Partners published another great read on AI in healthcare. The report largely focused on the investment market landscape and scaling early stage AI companies. Our favorite findings were on the AI services-as-software business model and average KPIs/benchmarks. Here are some interesting takeaways:

  1. AI services-as-software companies quickly hit $10M ARR and on average grow as fast or faster than tech-enabled clinical services.

  2. AI services-as-software companies are seeing record sales cycles and have CAC payback periods similar to SaaS. Some companies have seen <6 month sales cycles, much faster than traditional Healthcare sales cycles of 12-18 months.

  3. Healthcare AI businesses have similar efficiency scores as traditional software, despite varying gross margins by startup. (link)(tweet)

3/

Researchers say an AI-powered transcription tool invents false content

Several researchers raised concerns after finding that OpenAI's Whisper transcription tool suffers from frequent hallucinations and invents text that never appears in recordings despite being deployed in healthcare settings. This is healthcare AI’s driverless car problem. Here are some details:

  • A Michigan researcher found fabricated text in 80% of examined transcriptions, while another reported hallucinations in ‘nearly every’ Whisper output.

  • Hallucinations ranged from non-existent medical treatments to racial commentary and violent content.

  • Over 30,000 medical professionals use Whisper-based tools despite OpenAI's warnings against high-risk applications, according to the AP report.

Whisper was also the most popular open-source speech model according to Hugging Face, with over 4.2M downloads in the last month alone. Hallucications are one of AI’s current gaps and those over relying on the technology need to be wary of this. Proper safeguards, human oversight, and rigorous testing are essential, particularly in high-stakes fields like healthcare. Balancing innovation with responsibility is key to ensuring AI's potential benefits aren't undermined by its risks. (link)

Tools & Partnerships 🔧

Latest on business, consumer, and clinical healthcare AI tools and partnerships…

TOOLS

  • AI reaches expert level in medical scans: Researchers at UCLA just developed SLIViT, a new AI model that can analyze complex 3D medical scans with expert-level accuracy in a fraction of the time required by human specialists. The model matches clinical expert accuracy while reducing analysis time by a mind-blowing factor of 5,000. (link)

  • Talkdesk debuts AI agents: Talkdesk implemented agentic AI across its entire platform – now branded as Talkdesk Ascend AI. The AI agents will be used to quickly process large amounts of data, create personalized customer experiences, respond to live interactions, and manage complex contact center operations. (link)

  • Health Note’s new AI Scribe: Health Note launched its Health Note AI Scribe to streamline patient outreach, self-scheduling, front-office management, and EHR administration. It’s available through Epic and Athenahealth, and is already delivering over 500k clinical summaries per month. (link)

  • Myriad Health’s new AI scribe: The practice management software vendor released MyScribe AI for AI ambient scribing and coding suggestions. (link)

  • Artera realeses admin AI agents: The patient engagement and AI conversational chat solution released two new administrative AI agents – one to help staff manage patient communications and another to help healthcare organizations analyze patient engagement data. (link)

PARTNERSHIPS

  • HCA Healthcare + Commure/Athelas: Athelas/Commure partner with HCA Healthcare on one of the largest deployments of AI in healthcare. Commure Scribe and AI Copilots to go across 2300 sites of care, 45,000 providers. This partnership is on the back of the recent acquisition of AI scribe Augmedix. (link) (tweet)

  • Recursion + Google Cloud: Recursion and Google Cloud have expanded their tech partnership to use cloud technologies to power Rcursion's drug discovery platform. This strategic partnership includes exploring generative AI capabilities, including Gemini models, to support the RecursionOS, drive improved search and access with BigQuery, and help scale compute resources. (link)

  • Tanner Health/Healthliant Ventures + Corti: The GA-based health system and its investment arm are partnering with the AI clinician assistant vendor to support medical coding workflows. (link)

Deal Desk 💸 

Spotlight on latest capital raises, M&A, and investments…

FUNDING

  • Infinitus, an AI platform provider to automate manual healthcare phone calls, raised a $51.5M Series C financing led by Andreessen Horowitz. (link)(tweet)

  • Tennr, a New York-based health-care workflow automation platform, raised $37M in Series B funding. Lightspeed Venture Partners led, alongside insiders a16z and Foundation Capital. (link)

  • Counsel Health, a Menlo Park, Calif., health care virtual communications startup, raised $11M in seed funding. Andreessen Horowitz led, and was joined by Asymmetric Capital Partners, Floodgate Fund, and Pear VC. (link)

  • Dyania Health, an AI patient chart review startup, raised $10M in series A funding round led by HealthX Ventures. Tech Square Ventures and Cleveland Clinic Ventures also backed the round. (link)

  • Pharos, an SF-based hospital quality reporting platform, raised $5M in seed funding. Felicis led, and was joined by General Catalyst, Moxxie, and YC. (link)

  • Pip Care, an AI-powered patient care plan platform for patients undergoing surger announced a $5M Series A. (link)

MERGERS & ACQUISITIONS

  • Clarify Health + Prealize Health: Clariy and Prealize Health to join to help payors and providers anticipate utilization trends and proactively allocate resources. (link)

market snapshot as of 10/28/24

Other Relevant News 🔍

News, podcasts, blogs, tweets, resources, etc… 

  • Compared with what? Measuring AI against the healthcare we have (link)

  • Scaling specialty medical expertise with AI (link)

  • A strange way to fix healthcare: game theory (link)

  • A 'multibillion-dollar bottleneck': 3 top AI issues for health systems (link)

Visuals of the Week 📸

Funny memes, cool pics, and interesting data from around the web…

Healthcare & life sciences unicorn club (+$1B valuation)

Trust in doctors and hospitals: 70%+ in 2020, 31.4% now.

That’s it for this week friends! Back to reading — I’ll see you next week.

Stay classy,

— Healthcare AI Guy (aka @HealthcareAIGuy)

PS. I write this newsletter for you. So if you have any suggestions or questions, feel free to reply to this email and let me know

How was this week's newsletter? Tap your choice below👇

Login or Subscribe to participate in polls.