Generalists vs. Tech Leaders: AI Adoption at Any Stage

analytics-insightThis post was originally published by Analytics Insight at Analytics Insight

Generalists Vs Tech leaders

Deep learning and open source software adoption is redefining the healthcare and medical industries.

Artificial Intelligence (AI) is steadily making its way into all industries. From healthcare to retail, and consumer to enterprise applications, businesses are starting to unlock all the power and benefits AI and machine learning have to offer. One only needs to look at its growth: the AI services market is expected to grow 17.4% year-over-year in 2021, with revenues reaching $37.9 billion by 2024, thanks to a CAGR of 18.4%, according to Statista. Even with the pandemic-laden IT budget setbacks of 2020, AI is here to stay.

While the past year has been largely about survival and prioritizing mission-critical IT initiatives, innovation and optimization are making their way back into the fold as work begins to stabilize. This is good news for companies already on their way to AI-enablement, but it’s also a wakeup call for companies that have paused or haven’t yet embarked on their AI journeys. Whether you’re at a company where AI projects have been the norm for years or just getting ready to launch, there’s a few things to keep in mind as you go.

The healthcare and life sciences industries have been on the pulse of AI for years now, and as such, a recent survey, “The 2021 AI Healthcare Survey Report,” seeks to explore the trending technologies, tools, practices, and attitudes towards AI. Part of this research uncovers some of the fundamental differences of AI behaviors between AI novices and experts. While this particular survey is healthcare-focused, the findings can be applied across industries—and there’s a lot to learn from them.

Build or Buy: Open Source, Cloud Providers, and Consultants

In healthcare, there’s a trend toward leveraging open source software (53%) and cloud computing (42%), which happens in place of relying on consultants or commercial SaaS offerings. This appears to be driven by data privacy concerns and regulatory constraints. Those survey respondents considered technical leaders showed less interest than general respondents about working with consultants, except in cases where they could work with experts in healthcare data engineering, integration, and compliance, or in situations where external consultants could accelerate projects for a faster time-to-market.

44% of technical leaders also consider it very important to avoid data sharing with software vendors. This comes into play when you consider the most prioritized types of software identified in the survey were open source software and public cloud providers. Significantly, public cloud providers have a long history of guarantees for healthcare regulatory compliance and strong provisions for privacy and security. Major headlines in 2020 repeatedly highlighted the rise of ransomware attacks against healthcare facilities, often with disastrous outcomes. While security concerns should be top of mind for organizations in all stages of adoption in all industries, general respondents appeared to be more open to working with outside consulting companies, which makes sense as they navigate how to get their AI projects off the ground.

Tools of the Trade: BI, NLP, and CV

Deep learning technology has made considerable inroads into medical applications. Computer vision, in particular, has proven its value in medical imaging, to assist in screening and diagnosis. 23% of mature organizations, those that have been using AI models in production for more than two years, indicated that they use computer vision (CV), but other areas are becoming more prevalent, including data integration (45%), NLP (36%) and BI (33%). Although generalists are headed in the same direction as the experts, who have already or plan to integrate NLP and BI in the next year, the adoption rates are not surprisingly slower by approximately 10-20%.

Similarly, when it comes to deploying AI models into production, many of the same factors are important to both novices and experts alike, but the level of importance varies slightly by experience level. For example, when technical leaders in healthcare are evaluating tools, solutions, and services, they look for approaches that emphasize healthcare-specific models and algorithms, as well as production-ready codebases. They also highly value no data sharing with third-parties. Comparatively, those in the exploratory and early stages of AI valued these same criteria, but not necessarily at the same degree of importance. The biggest discrepancy was data sharing: those newer to AI put less of an emphasis on sharing data, likely because this group is more open to outsourcing help from consultants and third-party entities.

The Customers: Clinicians First 

When asked to identify intended users for their AI tools and technologies, over half of respondents identified clinicians as target users with healthcare providers as a close second. This is a big leap from AI being used primarily by data scientists and IT professionals, as was common in years past. This trickle-down effect of users persists even further when you consider the customers of mature organizations’ AI tools. In this group, a vast majority of respondents cited clinicians and healthcare providers were users of their AI technologies, with nearly 60% indicating that patients were also users.

As advances and applications of AI technologies grow, so do their intended user bases, so it’s important for all organizations to consider who they’re tailoring usability to. A patient who is interacting with a chatbot to schedule an appointment is a lot different than a radiologist using NLP to analyze the results of an X-Ray—and those are considerations that need to be evaluated when imagining the user experience. All organizations should be taking this into account, whether they’ve been deploying solutions for years now or are just getting started. As AI becomes more commercialized, newer players will take the lead from more mature companies that have had to evolve their customer base over the years.

There’s no one-size-fits-all approach to a good AI strategy. Make sure your goals are clear and you have the right resources in place—talent, tech, and operations. It’s also helpful to have a roadmap from companies at similar stages of their AI journeys. Taking notes from the healthcare industry is a surefire way to get your AI initiatives off the ground, or find ways to accelerate it further.

Spread the word

This post was originally published by Analytics Insight at Analytics Insight

Related posts