Five years ago I asked: “Can AI write a strategy?” The AI in question was GPT-3, which had written a strategy document for a healthcare scale-up called Sword Health. Large language models have come a long way since then, so let’s revisit the question with the current iteration, GPT-4. Can AI write a strategy now?
AI strategy written in 2020
The investment memo and GPT-3’s response were posted on Twitter in July 2020. Here is the beginning of the memo:
Sword Health is a full stack digital physical therapy provider. They are the first group in the world to create remote autonomous digitally guided therapy that provides faster recovery and better long term outcomes for a variety of musculoskeletal injuries. They are in the early stages of commercializing their solution, with Humana being their first US-based customer.
In 2020, GPT-3 suggested that a digital healthcare provider could expand its business by taking over physical care centres, writing:
First, we would acquire or acquirer an ambulatory primary care center that has PT, sports medicine, and physical therapy expertise. Then build a Sword center within that facility, allowing Sword to scale without having to scale its own brick and mortar centers.
I criticised the AI for picking a growth path that would be complex and expensive to scale, and not fully digital. It sounded nice but wouldn’t work in practice.
AI strategy written in 2025
This month I gave GPT-4 the same memo as input and asked it to complete it using the following prompt:
Please imagine that you are the author of the memo, and you need to expand it into a request for funding. How would you complete the memo, using the same writing style as the author?
Here’s the first part of what it produced:
The capital from this round will primarily be used to accelerate US market penetration. This includes building out a direct sales team focused on enterprise health plans and self-insured employers, expanding our clinical operations to handle growing patient volume, and standing up regulatory and compliance infrastructure to meet US payer and provider requirements.
In parallel, we will invest in product development to further differentiate our offering. This includes advancing the autonomous digital therapist platform, integrating biometric data for personalized care loops, and developing predictive analytics to stratify patient populations based on risk and expected outcomes. We believe this will further reduce our CAC by improving conversion and retention rates through superior patient experience and outcomes.
This answer is a lot better. It’s no longer recommending a non-scalable growth method. However, the strategy is still quite generic, as the first paragraph could apply to any healthcare company looking to grow. The second paragraph refers to their therapist platform, but doesn’t have any recommendations that are unique to the market that the company is looking to enter.
GPT-4 also wrote some paragraphs on competitive advantage:
Unlike competitors focused on digitizing traditional PT sessions, Sword has built a fully autonomous platform from the ground up, enabling fundamentally lower costs and higher scalability. This is not a telehealth wrapper—it is a new model of care.
Because of our digital-first approach, we collect continuous, high-resolution data across the patient recovery journey. This data powers our proprietary recovery algorithms and forms the foundation of our defensibility. Over time, this will enable us to shift from reactive care to predictive prevention.
Here is where the strategy becomes more appropriate for a digital company, making use of data and algorithms to build out an advantage. Other than that, it’s still quite generic.
Written strategies compared to reality
Sword Health is a real company. In 2020, they did not take over any physical therapy centres. Instead, they grew rapidly with their digital solution as people did their physical therapy at home during the pandemic. They raised $25 million in series B funding in February 2021, which was the first funding round after the date of the memo. At the time they wrote about their company:
[Sword Health] is the first virtual solution to pair its members with a licensed physical therapist and the Sword Health Digital Therapist, an FDA-listed device with a tablet and motion sensors, to deliver a clinically proven, personalized treatment plan that is more effective, easier and less expensive than the traditional gold standard of care.
Note how much more specific this text is compared to the AI suggestion, and the combination of digital and human therapists in their proposition.
In October 2021, they strengthened their competitive advantage by getting clinical validation for the effectiveness of their programs.
In later moves, they expanded their areas of attention from muscoskeletal disorders to women’s physical health, injury prevention and mental health. In 2024, they introduced an AI care specialist that supports the users during their physical therapy sessions.
Sword Health’s real strategy is much more interesting and unique than the AI suggested-ones are.
Can AI write a strategy now?
So how does the strategy written by GPT-4 stack up? There’s nothing wrong with the text that AI has written, but it’s not a proper strategy. It’s a set of plausible-sounding statements that list general moves that any company could make in that market. It lacks specificity, and it’s not going to give any company a real competitive advantage.
So like in 2020, even though the output looks a lot better, my conclusion in 2025 is still the same: No, AI cannot write a strategy. You still need to do your critical thinking and strategy creation yourself.
What should you do instead? Have a look at my four facets of a good strategy for guidance on what a good strategy should look like. And if you would like to have a sparring partner to help you create an actionable growth strategy that gets results, please get in touch.
© Veridia Consulting, 2025
