Diagnostics & imaging AI · Chest X-ray and CT brain AI
Harrison.ai
Harrison.ai makes Annalise, AI software that reviews chest X-rays and head CT scans and flags a wide range of findings for radiologists, and Franklin.ai, AI for pathology slides.
Strong fit
- Radiology groups and hospitals that want AI to flag many findings on chest X-rays and head CT scans
- Emergency departments that want urgent findings on X-rays and head CTs seen sooner
- Teleradiology providers that read high volumes for hospitals short of radiologists
Weak fit
- Buyers who want breast screening or cardiac imaging AI
- Organizations that want AI to read and report exams with no radiologist
- Buyers who need a published price list
Bottom line
Harrison.ai is worth using for radiology teams that want one AI tool to catch a wide range of findings on chest X-rays and head CT scans. It suits busy departments that want fewer missed findings. Buyers focused only on fractures or lung nodules may prefer AZmed or Riverain.
Score breakdown
Weights: Outcomes 35% · Product 30% · Implementation 20% · Pricing clarity 15%.
Harrison.ai, based in Sydney, Australia, makes Annalise, AI software that reviews chest X-rays and head CT scans. It also makes Franklin.ai, AI for pathology slides.
Brothers Aengus Tran, a doctor trained at the University of New South Wales, and Dimitry Tran founded it in 2018 after growing up in Vietnam and moving to Australia.
Annalise reviews a chest X-ray or a head CT scan and lists findings for the radiologist, so fewer problems are missed on busy days. In some markets it is cleared for up to 124 findings on chest X-ray and 130 on CT brain.
Franklin.ai reads tissue slides, starting with prostate biopsies. Its second foundation model, Harrison.rad.1.5, drafts radiology reports for X-rays.
Those products are in wide use. Hospitals and radiology groups in more than 25 countries use Harrison.ai, and it holds 12 FDA clearances.
It raised a $112 million Series C in February 2025 and says more than 40% of its business comes from the United States.
In 2026 it cut about 15% of its staff to focus on sales. It supplies the technology for Frontier Radiology, a U.S. teleradiology practice it does not own.
Other tools fit narrower needs. Qure.ai suits chest and emergency imaging AI across many countries. Riverain Technologies suits a focus on lung nodules on chest images.
AZmed suits fracture and X-ray detection, and Aidoc suits a health system that wants AI across many imaging types.
Competitor landscape
| Vendor | Overall | Ease of implementation |
|---|---|---|
| Aidoc | 7.9 | 7.5 |
| ScreenPoint Medical | 7.9 | 7.6 |
| Harrison.ai | 7.7 | 7.5 |
| Qure.ai | 7.4 | 7.1 |
| Riverain Technologies | 7.3 | 6.9 |
| Subtle Medical | 7.2 | 6.8 |
| DeepTek | 7.1 | 6.8 |
| AZmed | 6.9 | 6.7 |
| Enlitic | 6.9 | 6.5 |
| Avicenna.AI | 6.9 | 6.6 |
| Ferrum Health | 6.8 | 6.5 |
| Koios Medical | 6.7 | 6.3 |
Pricing
| Item | Detail |
|---|---|
| Model | Software subscription quoted by Harrison.ai. |
| What usually drives cost | How many exams you read, which products you use (chest X-ray, CT brain, or pathology), the number of sites, and connections to your image and reporting systems. |
| What to ask before you buy | The price per exam or per site, which FDA-cleared findings are included in the U.S., and the cost to connect it to your systems. |
| Published pricing | Harrison.ai does not publish a price list. |
Prerequisites for purchase
| Need | Why it matters |
|---|---|
| What you need to get Harrison.ai to function | |
| Lead radiologist for AI | Decides how findings are shown. |
| A connection to your image and reporting systems | Findings appear where reading happens. |
| Starting reading times and miss rates | This shows change later. |
| Policy on AI findings | Radiologists need clear rules. |
| IT support for integration | Studies must route to the AI. |
| What will maximize your value | |
| Start with chest X-rays | Highest volume and widest coverage. |
| Move urgent findings to the top of reading lists | Gets critical cases read first. |
| Audit flagged findings each quarter | This shows where AI helps and where it flags too much. |
| Deal-breakers | |
| You need breast or heart imaging AI | |
| You expect AI to report with no radiologist | |
| No radiologist owns it | |
Value creation time frame
| # | Stage | Typical range |
|---|---|---|
| 1 | Contract and integration | 1-3 months |
| 2 | Chest X-ray live | 2-4 months |
| 3 | CT brain and more sites | 4-9 months |
Leadership
Recent news
- October 2026: Harrison.ai apologized after a promotional video on its YouTube channel showed one patient's private information on an external system's monitor for about 3 seconds. It removed the video and is adding safeguards to how it films and reviews marketing content.
- September 2026: The founders told Forbes Australia that contracted annual recurring revenue had quadrupled over the past 18 months. They said the 34 roles cut in Australia in April were meant to shift spending toward sales, leaving just under 200 employees, and that Frontier Radiology had 20 radiologists.
Methodology
| Weight | Factor | What it measures |
|---|---|---|
| 35% | Customer outcomes | Whether buyers get measurable operational or clinical-workflow results after go-live |
| 30% | Product | Capability depth, reliability, and fit for the job the category buys |
| 20% | Implementation | How hard it is to stand up, integrate, train, and stabilize |
| 15% | Pricing clarity | Whether a buyer can model total cost without a mystery quote |
| Label | Meaning |
|---|---|
| Highly recommend | Strong outcomes and product with manageable caveats |
| Recommend | Solid fit for the right buyer; know the tradeoffs |
| Conditional | Only with a specific use case or heavy caveats |
| Not recommended | Avoid for most buyers in this category |
Read our full methodology for how we weight scores and assign recommend labels.