How to Use AI for Strategic Planning in Healthcare

You probably know that Artificial Intelligence can enhance a number of strategies in the healthcare industry. I’ve written about many before.

But what if you could apply Artificial Intelligence to the planning process itself? Imagine how much more value you could get by eliminating the constraints of traditional strategic planning.

The purpose of strategic planning is to prepare an organization for the future by setting goals and a process for achieving them. This requires capturing data and insights and then synthesizing that into a strategy. The data typically fall into four key areas:
1. What the company/organization offers (resources & competencies)
2. Who the company/organization serves (markets & customers)
3. Who competes with the company/organization (competitors)
4. The market environment the company/organization operates in (laws & regulations, economics, technologies, demographics)

Collating all relevant available data is the first step in an effective strategic plan. For healthcare strategic planning, this includes a wide range of disparate data sources including:

Internal data
• Internal records, systems, and processes
• Previous results data
• Government filings
• Consulting reports and white papers
• Clinical research findings
• Competitive intelligence
• CRM data
• Call centre data
• Market research data on patients
• Market research data on target physicians
• Market research data on target payers
• Complaint data

Publicly Available Free Data
• Media articles
• Public filings
• Corporate PR announcements and annual reports
• Corporate website statements
• Industry-specific websites
• Government freely available data
Commercial Data
• Claims data
• EHR/EMR
• Disease registries
• Sensor data from devices
• Physician and Patient associations
• Think tank publications
• Commercial industry newsletters
• Investment banking research reports

The problem with traditional strategic planning processes

The sheer quantity of data available for each of the areas above is more than any normal strategic planning team can analyse and synthesise. And moreover, traditional strategic planning tools are incapable of integrating and synthesizing this amount of data.

So in most cases, strategic plans are based on too little data to enable good strategic decisions.

In the established model, the problems are often defined and described mainly via reports made by corporate strategists and management consultants. But, these regularly represent the findings from a market research project with a limited number of respondents or limited data sources. And they may not actually be applicable to the entire relevant population.

Compounding the issue, the strategic team often makes assumptions about what will happen in the future based on data that is out of date, incomplete, inaccurate, biased and doesn’t cover all the issues required.
And when data is poor, strategic planners rely on gut feel and experience which doesn’t always count for a lot in a changing world without precedent. This leads to strategic plans that are erroneous and potentially damaging to the organization.

On top of this, strategic planning tends to be an annual, if not twice a year event, which does not always take into account market changes that can happen suddenly (like Covid-19). Events impacting strategy occur constantly, even if teams are not aware of them. Waiting until the following year to incorporate them is too late. The damage is done by then.

The impact of all this can be seen in surveys conducted by McKinsey and Bain for companies with revenues of $500 million and above.  According to the McKinsey work, only 23% had a formal strategic management process (and only 13% said it played a meaningful role in their corporate strategy). The survey also found companies generally do not focus their process on new growth opportunities.  Just 32% in the Bain survey believed they had a strong strategic planning process. The rest found it ineffective in gaining strong results when the plan was followed (which is, after all, the test of whether the strategic plan was indeed good or not). Interestingly the respondents that did report the strategic process was good were not satisfied with how it performed in the real world from their results. This begs the question of whether it was indeed satisfactory. I would argue if it doesn’t get results, then it is likely unsatisfactory.

So, what we are seeing is ineffective processes, with a lack of available and appropriate data, and ineffective analysis and synthesis, which happens too infrequently to work consistently in the real world.
But what is the solution, and where does AI fit into it?

As you know (if you read my articles or do my trainings), Artificial Intelligence relies on data — big data, and the more the better. And, as you can see from the previous data lists, healthcare companies have access to significant amounts of data. All data (in any format) can be combined and analyzed with AI. And it can use that data to make predictions with far higher accuracy than what humans alone can do.

Imagine if you had an ongoing strategic planning process that was automated (so it doesn’t need significant amounts of human time), highly intelligent and highly accurate in terms of growing the company results.
“Impossible,” I hear you gasp. Strategic planning takes teams months of time. That’s exactly why such a system would be so valuable.

What if I told you such a system is being developed?

It is spearheaded by a Professor from Harvard, and I am interviewing the team who developed it to inform my AI-cademy. Although I have seen a lot of Artificial Intelligence in my day, this is a game changer – for those who can afford it. If you are in a small company, you can stop reading about here. However, for those who can afford such a system, it will not only fix the numerous issues with the current process but save human time and gets far better results.

Here are some of the features that make this AI system so exciting:
• Full automation, operating 24/7 without human intervention
• Instant big data import, data analysis and strategic conclusions so that instant response to market events or competitor strategies can be implemented
• Constant automated data search (the types of data will be specified by humans)
• Constant automated data currency review and update from real data, so no data is out of date
• Automatic data validity check and data cleansing
• Automatic review of data relationships
• Automated review of reasoning and conclusions
• Constant automated review of assumptions underlying strategic decisions
• Automatically prompted creation of strategic implementation plans
• Constant automated reviews of strategy implementation for results analysis and adjustment for more rapid progress

On top of the capabilities above, they are planning include the following in the future:
• Assessing strategic thinking in manager
• Development of the manager’s ability to think strategically
• Gather and apply data on resources and competency
• Generate overall strategic vision
• Propose competitor focused strategies
• Analyse and propose multi-SBU portfolios and how to adjust to maximize strategic goals
• Define and recommend market segments for the organizations’ strategic attention (outside existing ones)
• Prompts for functional area sub-plans
• Suggest operational decisions for implementation
• Check for availability of resources and competencies
• Check effects of environmental factors
• Guide managers to the right strategies
• Guide understanding of internal value chain
• Guide understanding of industry value chain
• Identify value chains for strategic implementation (both low cost leadership and differentiation)
• Identify key success factors
• Quantitative assessment of strategy risks
• Recommended strategies for achieving competitive advantage
• Recommend external combination strategies
• Monitor the implementation and effectiveness of a strategy
ROI of this approach

As you would imagine, the costs of investing in this kind of system are not insignificant, but they would be offset from the additional revenue generated from the data-driven, results-based strategic planning. Obviously, the costs of launching and maintain the platform are easy to track, as you are paying the company to customise it for your needs as they do for each implementation. The challenge is to definitively prove the results are from the strategy offered.

The system is custom set up for each client. There is an upfront investment to set up the system and re-engineer the clients’ internal systems to accommodate the product. Plus there are organizational infrastructure and team training costs to ensure that the employees (at all levels) can work effectively with it.

Consultation and advice are provided throughout this process and, if required, there may be adjustments to increase the effectiveness. Then there is an ongoing subscription charge to use the service. In addition, there will be additional data costs if the full range of data available is to be used.

So, what will the ROI be?

There are two cost areas that the return can be examined from. One is the cost of traditional strategic planning in human time and resources as well as data costs and external costs (consultants, off-site meeting costs etc). The other is the incremental gains following implementation compared with the usual gains. These are a mix of enabling agile rapid strategic decisions, the successful predictions of profitable new segments or predicting competitor actions correctly, and steadily increasing profitable revenue gain.

The latter is easy to put a number on. The former components are less so, but they are also important aspects. So you need to compare the time, money and resources used in setting up and operating this approach with the outcomes achieved from it to determine if the return equals or exceeds its average cost of capital.
Given this is an AI data-driven system, the system will continue to learn over time. So when you are looking at the ROI, you should consider it over a period of years.