Hybrid Intelligence: How AI roles can lead to commercial success (or failure)

Industry 4.0 has arrived, ushering in a new era of direct data exchange and automation in economies and markets globally. This new world is manifesting itself through cloud computing, IoT, intelligent data tools, and a broad spectrum of other transformative technologies.

However, as is common in times of upheaval, organizations are struggling to keep up with the rapidly shifting social, political, and economic realities that stem from having access to vast amounts of data. Far too often, corporate leaders adopt a policy of “hiring a data scientist” or “getting the IT department to figure it out” without fully comprehending how a good data team could become their most valuable asset.

In 2019, researchers working for Pactera Technologies revealed a shocking state of affairs when they estimated that 85% of AI projects fail to deliver on their original aims. Mostly, this is because projects are led by data scientists or IT professionals rather than business strategists, who are side-lined due to their lack of technical knowledge.

Today, I’m going to take a dive deep into AI implementation and how it can trip up even the best-prepared firms. I’ll offer a new perspective on AI and explore solutions to ensure your machine learning projects consistently land within that elite 15% of AI undertakings that do deliver on their original goals.

Understanding how AI can go wrong — Industry 4.0 in context

It can be easy for AI to feel, at best, like little more than the latest buzzword — a descriptor for simple algorithms that do little to improve your business’s capabilities. At worst, AI can become a barely understood black box that threatens your workforce.

Just as companies that thrived in the 20th century had to rethink their processes as we moved into the digital age, businesses again need to adapt to the new age of AI-driven information exchange and machine-to-machine communication. Unfortunately, adjusting to this status quo isn’t as simple as digitizing paper files — impactful AI execution requires an ability to grasp the relationships between AI tools themselves, the way they’re implemented, and the people who will interact with them.

Even machine learning experts can run into difficulties when implementing AI in the real world. In particular, the healthcare industry has been slow to utilize neural networks and other advanced AI tech. Partly, that’s down to inflexible regulations and concerns about limited decision traceability, but adoption has also been hampered by weaknesses that creep in due to wide gulfs between management and development teams.

Why good AI projects go bad

It seems like everyone is looking to implement AI into their business model. The problem is, managers and executives often throw together IT teams to develop AI functionalities and then walk away. When a project ultimately fails or proves to be superfluous, everyone is baffled.

What’s important to understand is that AI project failure doesn’t occur in a vacuum. People are more than just secondary considerations to hardware and software when it comes to AI — the reason things don’t work out as they should is almost always down to human rather than technical failure.

In a recent seminar I attended, a participant shared an anecdote about a project in which a tech team was tasked with finding an automated solution to a business challenge. Time and funds were poured into the project to no avail. Eventually, someone identified a low-tech solution, instantly solving the problem. In the end, a low-tech solution solved the issue, requiring no additional resources.

The takeaway? Non-tech execs and employees need to involve themselves in the development of AI tools to ensure objectives are contextualized and developers aren’t operating in an environment that needlessly blinds them to project weaknesses. This is a challenge seen across the pharma industry where tech teams are tasked with leading the AI projects but they are not close enough to the business challenges to identify the optimal solutions.

What business leaders need to do

So, what does a well-rounded approach to AI look like for businesses? In the past, firms that have successfully solved these issues have done so by hiring those rare professionals capable of bridging gaps between data science and business strategy.

Of course, the problem with hiring these kinds of multidisciplinary workers is that they are currently not that common and currently you either have commercial teams or tech teams. However, the most effective way to assemble a team that can zero in on the best opportunities that Industry 4.0 has to offer is to teach your existing commercial teams and strategists how to take on “hybrid intelligence” roles. In other words, you can leverage your existing workforce through skill training rather than hiring anew.

To make that work, professionals need to educate themselves about what AI is, the strengths and limitations of different approaches, and how it can be applied without getting them to endure learning code and math as there are existing teams for that. After that, they can apply their new knowledge in context — the logical next step is to identify specific recurring problems, financial burdens, or goals to address via machine learning and other Artificial Intelligence techniques.

Understanding the basics of creating a fool proof AI roadmap

Initially, finding the best way for your organization to utilize AI will require a strategy of its own — an adequate preparation work package should align your aims with organizational capabilities and take shareholder expectations into account. Once the planning phase is complete, it’s time to hit the ground running.

Subsequent scoping should involve both business strategists and data scientists. Together, this multidisciplinary team can scour relevant data for clues on how to proceed and come up with a development model that shows infrastructure and data engineers what kind of software environment is necessary for the project.

Finally, developers will team up with infrastructure engineers to integrate machine learning protocols. This is the point at which teams will decide if a custom tech stack is required or whether you could source an existing solution. Importantly, business analysts will also begin to receive data so that project success and KPIs can be gauged by product owners in real-time.

Preparing your commercial teams for delivery

Whether your team needs to commercialize an AI solution for external distribution or integrate a new tool into internal workflows, a culture of flexibility needs to be fostered. A key benefit of Artificial Intelligence techniques is often the elimination of monotonous tasks, and this can feel threatening for current workers.

However, open discussion and education will help to assuage concerns. What’s more, business teams that possess a deep understanding of AI will be better able to use newly-developed custom AI/ML tools for creative problem-solving.

Training in AI strategy for pharma business challenges

Beyond a willingness and enthusiasm for bringing AI onboard, teams will need specific skills to become adept in handling AI tools and interpreting outputs from them. In the healthcare sector, for example, strategists require a good understanding of customer engagement and experience to utilize AI effectively — the healthcare space is one of many markets where future growth will largely come from improved customer engagement and strong customer experience as well as the proliferation of existing ideas and services.

An example of an effective AI integration solution

Addressing systemic AI implementation issues isn’t easy, but there are good resources out there to help. One such system I like to recommend is one I developed specifically for this purpose based on regularly being called in to rescue projects where the process was led by Tech teams when I could not find any other solution on the market that addressed this. It is an in-depth online training package that covers everything from AI basics to the ins and outs of identifying the right business challenge to solve, planning the optimal solution, integrating the right data to use and how to plan the integrated data science and business strategy. This can help a business reboot its approach to healthcare and data ecosystems. With this training, for example, you’ll find out how to do the following:

● Create an effective strategy to solve your specific pressing business issues
● Identify the optimal approach to solve your specific challenge with AI
● Help your team plan the process an create an ROI
● Streamline business activities
● Add intelligence to sales and marketing strategies

It’s certainly worth checking out this program if you want to gain the credibility to guide data and analytics teams, collaborate with vendors and tech professionals (while speaking the same language) to reach profitable conclusions, and accurately demonstrate your return on investment.