Identify problems across the organization

Why does this matter?

  • To build trust. Involving diverse stakeholders in problem identification can help build trust among frontline clinicians and mitigate downstream challenges related to AIInterdisciplinary field, usually regarded as a branch of computer science, dealing with models and systems for the performance of functions generally associated with human intelligence, such as reasoning and learning. adoption and effective use.
  • To level the playing field. Sourcing problems across the organization ensures that expensive AI-based solutionsThe combination of the AI product and its use in healthcare delivery setting (including user experience and workflow of use) are accessible to all and avoids creating “pockets of excellence Subset of highly resourced business units, departments, and professional groups (e.g., specialist and procedural physicians) will build internal expertise in AI adoption that may be inaccessible to lower-resourced settings.” and a digital divide.
  • To strengthen relationships with front-line workers. A top-down approach to AI adoption can threaten worker autonomy, drive a wedge between managers and frontline workers, and alienate clinicians already struggling with high burnout rates. 1
  • To work on the right problem. Sourcing problems from frontline clinicians involved in healthcare delivery can increase the likelihood of identifying problems well suited for AI. Remember that frontline clinicians are most familiar with how healthcare data is generated in healthcare.
  • To see problems from different angles. Proactively seeking input and participation from different stakeholders can help leaders better understand problems. Check out this guide for more on how to unpack the dimensions of a problem.
  • To broaden the portfolio of AI productThe machine learning model or set of heuristics that is tested and validated by a group of clinicians and data scientists to address the identified pain point safely and effectively.. Employees across the organization are becoming familiar with AI outside of work. Provide them with a mechanism to present problems that can potentially be solved by AI. 

How to do this?

Step 1: Get broad senior leadership buy-in to

  • Gain visibility into priorities and challenges throughout the organization.
  • Secure funds to address the problems.
  • Obtain political and social capitalResources that can take the form of connections, norms and trust, available to individuals and groups through their social networks that facilitate cooperation and coordination among individuals and groups. to navigate internal organizational barriers.

Successfully addressing solicited problems is unlikely without the buy-in of senior leadership. It is important to engage senior leaders from various departments, including information technology (e.g., CIO, CMIO), clinical quality (e.g., CMO, CNO), and operations (e.g., COO, hospital president). Relying solely on the support of technology leaders could result in downstream adoption challenges.

Step 2: Develop the problem solicitation process

  • Create an online form accessible for everyone across the organization that’s short, 2-3 pages, and free of jargon.
  • Have minimal eligibility criteria.
  • Ask questions to characterize the problem and identify potential opportunities for improvement.
  • Avoid seeking prescriptive solutions or requiring that AI be part of the solution. AI hype can lead employees to assume that AI is the best solution to a problem when there may be better or cheaper approaches.
  • Check out the Duke Health innovation RFA for inspiration here

“There’s often so much wisdom that can be found in the community. Communities often know how to solve their own problems. Ideally, you would have community engagement at that early stages of problem formulation”

Community Expert

Step 3: Secure a budget for soliciting problems

  • A budget helps demonstrate the organization’s commitment to addressing problems identified by the frontline clinicians.
  • Ensure the budget is broadly deployable across the organization rather than tied to specific business units or use cases. This allows for greater engagement from frontline clinicians and a broader potential impact.
  • Ensure sufficient funding for initial development and testing. The funding range varies based on organization, risk profile, and use case. Example: Duke Health’s innovation tournament provides $25k-$75k per project per year. These projects are high-risk, and it is expected not all will succeed. On the other hand, organizations may allocate larger amounts for projects with a high probability of success. 

Step 4: Leverage broad and narrow communication channels

  • Plan communication and marketing to reach workers at all levels of the organization, including the frontlines.
  • Disseminate problem solicitation through executive-led company-wide channels. This can include email, internal newsletters, all-hands meetings, town halls, and other highly visible channels.
  • Disseminate innovation tournament materials through business unit channels. This can include chairs of departments or leaders of clinical service lines disseminating information to their staff.
  • Set expectations with senior executives to ask direct reports to pass information about the innovation competition to their teams and front-line workers.
  • Present at team meetings to diverse audiences across business units.

Step 5: Challenge existing power structures

  • People most advantaged by current power structures will likely feel most empowered to identify and submit problems for prioritization. 
  • Put in extra effort to invite non-physicians and non-specialists to submit problems.
  • Encourage women, nurses, pharmacists, medical assistants, physical therapists, social workers, occupational therapists, primary care providers, and more to submit problems.

Step 6: Facilitate submissions

  • Create an easily accessible online portal for employees across the organization to submit problems. Provide detailed information about a review timeline and when applicants can expect to hear about the next steps.
  • Include the contact information of a team member with dedicated effort to support individuals submitting problems on the problem solicitation form.
  • Make time to meet with employees across the organization to discuss ideas, review drafts, and provide feedback on potential problems.


  1. Kellogg, Katherine C., Melissa A. Valentine, and Angele Christin. “Algorithms at work: The new contested terrain of control.” Academy of Management Annals 14.1 (2020): 366-410.

adopt health AI