30 Days of (practical) AI – Day 7: The (failed) Pitch Deck

Today, we wanted to have GPT think like an entrepreneur. To do this, I leveraged two GPT-4 plugins: Scholar Assist and Smart Slides. My idea was to have GPT search recently published scientific papers and make a connection between two seemingly unrelated topics. It was up to GPT to connect the dots and form a node between the two ideas. Along with developing an idea for a product, I wanted to make GPT think like a highly talented entrepreneur who had previously walked the walk:

Here is GPT’s response including the company idea:

Based on the recent technological developments from the academic journals, I’ve identified two promising scientific discoveries that can be interconnected to open up a new market opportunity:

1) Scientific Discoveries:

a) Recent Development in Disease Diagnosis by Information, Communication and Technology

• Published Date: 2021-02-05
• Authors: Shabana Urooj, Astha Sharma, Chitransh Sinha, Fadwa Alrowais
• Summary: The integration of Information, Communication, and Technology (ICT) in the health sector has shown significant potential in improving individual and community health, disease detection, prevention, and overall strengthening of healthcare systems. Large ICT establishments are now offering a variety of Artificial Intelligence (AI) based solutions, including wearable therapeutic devices, healthcare management systems, predictive healthcare diagnostics, disease prevention systems, and automated tactics. AI plays a pivotal role in healthcare instrumentation, enabling machines to sense, comprehend, act, and learn to perform both organizational and clinical healthcare functions.
  Link to the paper

b) AI-based data corrections for attenuation and scatter in PET and SPECT

• Published Date: 2021-08-27
• Authors: Alan B. McMillan, Tyler J. Bradshaw
• Summary: Recent advancements in artificial intelligence technology have enabled improvements in attenuation and scatter correction in PET (Positron Emission Tomography) and SPECT (Single Photon Emission Computed Tomography). These technologies allow for accurate and quantitative imaging without the need for a CT (Computed Tomography) image, expanding the capabilities of PET/MRI, PET-only, and SPECT-only scanners. The use of AI in scatter correction leads to faster image reconstruction and improved patient throughput.
  Link to the paper

2) New Product Idea:

AI-Enhanced Diagnostic Imaging System (AEDIS)

Description: Leveraging the advancements in AI and ICT, our team proposes the development of an AI-Enhanced Diagnostic Imaging System (AEDIS). This system will integrate the capabilities of AI in disease diagnosis with the advancements in PET and SPECT imaging. AEDIS will offer:

• Rapid Imaging: Faster image reconstruction, reducing the time patients spend in the scanner.
• Enhanced Accuracy: Improved attenuation and scatter correction, leading to clearer and more accurate images.
• AI-Powered Analysis: Real-time disease detection and diagnosis, with AI algorithms analyzing the images as they are produced.
• Integrated Health Records: Seamless integration with electronic health records, allowing for immediate updates and easy access for healthcare professionals.
• Wearable Integration: The ability to connect with wearable therapeutic devices, providing a comprehensive health overview for patients.

Market Opportunity: With the increasing demand for rapid and accurate diagnostic tools in the healthcare sector, AEDIS will cater to hospitals, diagnostic centers, and clinics. The integration of AI will not only enhance the diagnostic capabilities but also improve patient experience and throughput, making it a valuable addition to the healthcare infrastructure.

By combining the advancements in AI-powered disease diagnosis with the latest in imaging technology, AEDIS will revolutionize the way healthcare professionals diagnose and treat diseases, offering a comprehensive solution that is both efficient and effective.

From here, we looked to expand just a bit more on the idea:

AI-Enhanced Diagnostic Imaging System (AEDIS) – In-depth Overview:

1. Core Technology Integration:

• AI-Powered Analysis: At the heart of AEDIS is a sophisticated AI algorithm trained on vast datasets of medical images. This AI can detect anomalies, predict potential diseases, and even suggest possible treatments. The real-time analysis ensures that any irregularities in the images are immediately flagged, reducing the chances of human oversight.
• Advanced Imaging: Incorporating the latest advancements in PET and SPECT imaging, AEDIS offers unparalleled clarity and detail in its scans. The AI-driven corrections for attenuation and scatter ensure that the images produced are of the highest quality, making diagnosis more accurate.
• Seamless ICT Integration: AEDIS is designed to integrate seamlessly with existing Information, Communication, and Technology systems in healthcare. This ensures that patient data can be easily accessed, updated, and shared across platforms, improving coordination among healthcare professionals.

2. User Experience:

• Rapid Turnaround: One of the major pain points in diagnostic imaging is the waiting time for results. With AEDIS, the AI-powered analysis ensures that preliminary results are available almost immediately after the scan.
• Patient Comfort: The system is designed to be as non-invasive as possible, reducing the discomfort often associated with prolonged scans.
• Integrated Health Dashboard: Patients can access an integrated health dashboard that not only displays their scan results but also offers insights into their overall health, drawing data from connected wearable devices.

3. Scalability and Future Expansion:

• Modular Design: AEDIS is designed to be modular, allowing for easy upgrades as technology advances. This ensures that the system remains at the forefront of diagnostic imaging technology.
• Expansion into Telemedicine: With the rise of telemedicine, AEDIS can be adapted to offer remote consultations, where specialists can analyze and discuss scan results with patients in real-time.

Total Addressable Market (TAM) Prediction:

To estimate the TAM for AEDIS, we need to consider the number of diagnostic imaging procedures conducted annually and the potential growth in this sector.

1. Current US Population: As of my last update in 2021, the US population was approximately 331 million.

2. Diagnostic Imaging Procedures: According to data from previous years, roughly 75 million MRIs and CT scans are performed annually in the US. If we consider other imaging procedures like X-rays, ultrasounds, and the potential market for PET and SPECT scans, this number can easily exceed 200 million procedures annually.

3. Pricing: The average cost of an MRI scan in the US is around $1,200. If AEDIS can offer enhanced imaging at a similar price point, the potential revenue from MRIs alone would be $90 billion annually. Including other imaging procedures, the potential revenue could exceed $240 billion annually.

4. Market Penetration: If AEDIS captures just 10% of this market in its initial years, the TAM would be $24 billion.

5. Growth Potential: With the increasing aging population and the rise in chronic diseases, the demand for diagnostic imaging is expected to grow. Moreover, the integration of AI and the potential for remote consultations make AEDIS a future-proof solution, ensuring steady growth in the coming years.

In conclusion, based on the current US population data and the demand for diagnostic imaging, the TAM for AEDIS is substantial. With its advanced features and the potential to revolutionize diagnostic imaging, AEDIS is well poised to capture a significant share of this market.

Ultimately this led us into instructing GPT to throw together pitch deck instructions on how to develop slides. I tried providing context that these instructions would then be fed into another plugin that will make a deck to be used.

I then copied over the instructions given to me in my previous response from GPT while leveraging the Smart Slides plugin to create our pitch deck:

DRUMROLL PLEASE……..

As you can see above, IB analysts have absolutely nothing to worry about. The current status of AI pitch deck builders ranges from pretty shitty at best to hardly understandable. Ultimately I wanted to include this effort as an example of a narrative currently floating around on social media. “AI is going to come for our jobs” “AI will replace humans” “We won’t need any human insights anymore, we can just automate”. I don’t agree with any of the aforementioned doom and gloom. On the contrary, I think we are hitting on an important point of human nature. LLMs are extremely talented at analyzing processes, activities, and documents that have been posted about online thousands of times. From all of these posts, each LLM is able to discern the most important takeaways and riff off of any prompts being fed into the chat box. What the AIs fail in is exactly where we thrive as a species. Humans are uniquely good at connecting the dots between two seemingly unrelated ideas. We are each so talented at this because we ultimately understand what it means to be a human. The desires, pains, loves, and cravings push us to read through two scientific papers and make a connection between the two on the off chance that this idea could make someone’s life just 1% easier. We can be powered by LLMs to work more efficiently, but only humans will forever have the ability to make said connections and tap into the human spirit.