Sell AI as labour automation

The pitch that keeps landing in hospitals is not “buy our artificial intelligence and maybe revenue goes up.” It is “buy our software and you need fewer hours of manual work next quarter,” because hospitals are still dealing with staffing shortages, higher labor costs, and margin pressure. (hfma.org) That sales framing fits the math of the industry. The American Hospital Association says hospitals and health systems are still under workforce and financial strain, and its 2025 and 2026 workforce scans describe redesigning roles and workflows as a top operating priority. (aha.org) The reason labor automation is such a clean wedge is that U.S. healthcare spends about 25% of total expenditures on administration, or roughly $1 trillion, according to an American Hospital Association market scan summarizing McKinsey’s 2021 work. (aha.org) McKinsey’s estimate, cited by Fierce Healthcare and discussed in JAMA, was that more than a quarter of that administrative spending, about $265 billion, could be cut without reducing care quality or access. That is why software that removes data entry, coding rework, prior authorization chasing, or denial follow-up gets attention faster than software promising vague “growth.” (fiercehealthcare.com) (jamanetwork.com) You can see the same logic in the Council for Affordable Quality Healthcare index. Its 2024 report put the remaining savings opportunity from moving routine administrative work from manual to electronic at $20 billion across eligibility checks, claim status, prior authorization, and other transactions. (caqh.org) That is why a vendor promising a 3 to 4 times return on investment and 25% to 33% cost savings is speaking the buyer’s language, even if those numbers come from a sales thread rather than a public study. In hospital operations, a faster claim, a shorter coding queue, or fewer full-time equivalents needed for repetitive work is easier to budget than a speculative new revenue stream. (modernhealthcare.com) Revenue cycle management is where this shows up first. Modern Healthcare reported in October 2025 that leaders see artificial intelligence as a way to reduce administrative burden, cut costs, and speed cash flow, but many still struggle to fund projects unless the return is measurable. (modernhealthcare.com) The budget problem is real. In that same survey of 116 healthcare leaders, 47% said they were exploring artificial intelligence with no concrete deployment plans, and 49% said 10% or less of their technology budget was dedicated to artificial intelligence initiatives. (modernhealthcare.com) That is where at-risk pricing comes in. If a vendor is selling a module tied to collections or denial recovery, the hospital can ask the vendor to take part of its fee only when money is actually recovered, which turns a hard budget ask into a shared-risk contract built around a concrete outcome. (modernhealthcare.com) The lower-friction move is to start with boring work that already has a line item and a backlog. The American Hospital Association’s artificial intelligence action plan points hospitals toward administrative workflows first, because those projects usually need less clinical change management than bedside tools and can show savings faster. (aha.org) That is also how new vendors displace incumbents. They do not need to replace the whole hospital software stack on day one; they just need one workflow where implementation is light, manual effort is expensive, and the savings show up before the next budget cycle. (aha.org) (hfma.org)