In pharma, the biggest barrier to scaling AI is not capability, but trust. The Take5 Report by HFS Research, in partnership with MathCo, across industries including pharma and life sciences, asked senior AI and data leaders which mode of AI deployment they trust most. 45% pointed to “copilot” tools that draft, summarize, or recommend, with a person reviewing the work prior to execution. Only 14% say they trust AI more when it is embedded directly into a process, acting on its own with little or no human review. That threefold gap in trust toward human-reviewed AI is the real driver behind AI adoption in pharma today.
The Pattern: AI Investment Follows Trust
The survey also asked pharma respondents another question – where does most of your AI portfolio sit today? The answers pointed in the same direction as what respondents say about trust. 45% said copilot tools account for the majority of their AI portfolio, while only 23% have advanced to embedding AI directly into business processes to drive automation. This highlights that most pharma investment leans toward augmentation; tools that draft, summarize, or recommend, with a person making the final call, rather than toward systems built to run a task or workflow independently. This preference for human review is not unique to pharma. Across industries, most large enterprises are still working out how much of their AI should act on its own versus how much should stay under human review. Pharma simply leans further in that direction than the rest, as evidenced in the same survey. Among all the respondents across industries, only 29% report a majority people-productivity portfolio, versus pharma’s 45%. Pharma is markedly more augmentation-focused than the other industries surveyed, and notably more cautious about letting AI act unsupervised.
However, the intent is shifting. Asked where they want to move over the next two years, few pharma leaders pointed toward greater automation than toward more copilot tools – a modest but real shift. Leaders already know, at least directionally, where the stronger returns sit: a clear majority say automation delivers stronger measured results than copilot-led initiatives. Still, today’s investment favors the safer, more supervised option, even as respondents acknowledge automation is the one moving the needle on performance.
The Reason: Trust Is Shaped by Regulation
Pharma’s continued preference for copilots over automation is not a sign of indecision. It is a rational response to real regulatory exposure where trust breaks down. AI in pharma operates inside a GxP- and FDA-regulated environment, where an incorrect automated decision can carry clinical, regulatory, or patient-safety consequences.
Governance, risk, and compliance concerns rank among the top constraints on AI adoption in this survey, cited by 27% of respondents as a top three barrier to scaling. Keeping a human in the loop is not caution for its own sake; it is a defensible way to manage risk. Pharma has simply not yet built the accountability and auditability infrastructure that would make unsupervised automation as trustworthy as a reviewed recommendation, and until it does, the copilot will remain at the default.
The Cost: Delayed Automation, Delayed Value
None of this makes the current approach wrong, but it carries a cost. If automation delivers stronger measured returns for a clear majority of organizations, every quarter spent defaulting to copilots is a quarter of value left uncaptured, and that gap compounds. Automation that could be running an entire workflow today is instead limited to assisting the person who runs it, and the returns pharma is already leaving behind only grow larger the longer the trust gap stays open.
The deeper issue is not a lack of ambition; it is a lack of readiness. AI needs enterprise context; the tribal knowledge, workflows, and decision logic that make an organization function to act reliably. Yet only 18% of senior AI and data leaders in pharma say they treat that context as a business asset that grows more valuable the longer it is maintained. Without that foundation, wider automation is not a risk pharma can responsibly take yet. That’s not because of regulation alone, but because the trust to support it has not been earned, and that trust must be built deliberately, not assumed.
Bridging the Trust Gap
The trust gap in AI adoption across pharma is not a maturity failure; it is a response to real regulatory and operational risk. Without intervention, it becomes a ceiling on the value pharma can realize from AI. That ceiling is measurable: it is the gap between the returns automation already delivers, and the returns organizations are not yet capturing.
Bridging that gap starts with fundamentals: stronger AI accountability structure, auditable decision trails, and named ownership of enterprise context. The same capability 77% of pharma leaders call critical in a delivery partner, yet only 45% currently assign it to a single accountable owner internally. Put those in place, and the choice between copilot and automation; it can engineer for both.
Curious how leading pharma organizations are approaching AI trust and automation? For leaders looking to benchmark their position and explore the broader trends shaping enterprise AI, the Take 5 Report, developed in partnership with HFS, offers deeper insights into where the industry stands today and the opportunities that lie ahead.
Click here to read the report.
