数字健康创始人不应再担心自己是否“足够 AI”

Everybody wants to be “ AI-first” because the market is telling them they need to be. But what is AI-first? Is it AI-native? Is it software that has successfully threaded AI across its business? I’m not sure anyone can confidently answer the question, and it’s creating a lot of confusion in the market among founders, buyers, and investors.

The money is still flowing regardless –

especially in healthcare. Rock Health reported that AI-enabled companies took 54% of digital health funding in 2025. At the same time, KLAS found healthcare organizations using AI increased from under half of respondents to more than two-thirds over the course of 2025. Menlo Ventures, among many others, have reported seeing the same thing in practice That’s the market right now: plenty of capital, real urgency, and a wave of founders trying to determine whether they’re building something durable or just something timely.

I believe there are two distinct categories of companies trying to persevere through this moment: AI-native companies that will need to become defensible software companies quickly, and software companies that need to figure out how to bolster their defensibility with AI almost as fast.

Neither are safe, per se. But if you’re a software company, I believe you have more time than the market is telling you. It’s not a lot of time, but it’s a head start.

Software companies likely already own workflows. They sit squarely within their clients’ daily processes, have proprietary data, and key integrations, all of which are necessary moats in any market. Workflow depth, the data that lives inside that workflow, and the integrations that make you a pain to remove or replace – these are all the factors that vertical SaaS companies strive to attain because they elevate you to a system of record.

AI-native companies are facing a distinctly different challenge. Their bar for success is higher than the market makes it seem, as investors rush in to bet on their “horse.” Companies built on AI models have momentum, but it is just a head start; they need to build those moats quickly before they end up either as a useful feature inside somebody else’s stack, or as a point solution in an era of point solution fatigue.

You can already see this playing out.

Specialized models, especially in a complex workflow like claims submission, can look like novel improvements over generic models initially, but that’s a fragile and fading value proposition if you do not control the data within that workflow, the customer relationships around that workflow, or the workflow itself. As point solution fatigue builds, the grace period for niche AI solutions is narrowing.

That forces a hard question: do you control enough of the system to prove that ROI consistently? If not, your window is closing.

So why is the market still rewarding AI bluster? Because we’ve seen this movie before.

I have no doubt that AI will fundamentally change digital health, but the hype cycle is reminiscent of previous “revolutionary tech” waves like mobile-first and web3. A decade ago, the entire market of MacPractice was just an EHR that could work on Apple products. DrChrono was the mobile-friendly EHR. The legacy EHRs cannibalized those “differentiators” quickly, but it didn’t mean there wasn’t an opportunity to capture market share with them for some time.

AI is a tool, a Swiss Army knife. Founders need to ask themselves: do you want to be a company that makes screwdrivers or DeWalt?

In healthcare, providers, health plans, and employers are among the first to embrace a more grounded view of AI because their bank accounts have begun forcing them to.

KLAS found

that organizations are still most willing to deploy AI where the work is well-defined and the risk is manageable. Healthcare leaders are using AI in admin work, transcription, and revenue cycle automation because the value is clear. Their palate for advanced agentic tools is much more picky.

In healthcare, enterprises are moving slower than the hype cycle would lead you to believe. That said, they’re not moving slowly – they are saying yes to AI, as long as it can solve for specific workflow needs. That’s what software enterprises need to consider: are you innovating to stay relevant, or innovating with a long-term goal in mind?

This is where I’m finding some comfort in the confusion as a founder. When you cut through the noise, you realize the job hasn’t really changed. The market rewards businesses that show real usage, real retention, and real control over their domain. AI will become ordinary, much in the way mobile-first did. It doesn’t matter if you’re “AI enough” as long as you’re using AI as a tool to deepen your moats.

About Blake Walker Blake Walker is the co-founder and CEO of Inbox Health, a company dedicated to transforming the patient billing experience in healthcare. He has focused his career on design and innovation in the patient billing space and played a pivotal role in developing technologies that simplify medical billing for patients and healthcare providers alike. Under his leadership, Inbox Health has become a trusted partner for more than 3,000 healthcare practices and more than 2 million patients a year and was recently named to the Inc. 5000 list of fastest-growing private companies in America.

Reader Interactions