In Malcolm Gladwell’s book, the Tipping Point, Gladwell shares how the Cambell Soup Company’s Prego pasta sauce brand went on a journey to find the best sauce with the goal of beating the market leader (Ragu at the time). What Prego discovered was that there are many perfect sauces and people have preferences to meet personal needs. One sauce would not suffice and today the pasta sauce epiphany cuts across every grocery aisle and product; filling shelves with an overabundance of mayonnaise, mustard, bread and other choices to meet a consumer’s needs.
I’m sure most of us agree we wish finding the perfect technology were as simple as buying pasta sauce! Even with choices, today’s technology options provide partial solutions that don’t always meet all of industry’s requirements. Buyers should not be wary, however, as the important thing to remember is that today’s technology has potential to be highly scalable and extensible if you make the right choices.
Increased attention on content automation has advanced significantly over the past two years, largely due to the commercialisation of Generative AI and the potential this technology might provide in authoring documents for medical writers. The enthusiasm for Gen AI and the potential impact on medical writing has gone from excitement to a flood of information across the industry. The innovation overload with promises that AI will do all the writing makes it challenging to navigate the jungle of technological promises, leaving a medical writing organisation having to choose a tool that might address one or two document solutions, apply editorial support solutions, or be so paralysed by the options that they do nothing at all.
Getting started:
Just understanding what technology can and cannot accomplish requires a scrutinising eye to self-educate, challenge suppliers, and ensure the solution presented creates an impact and adds value. Each organisation has its unique business needs to consider: the types of documents being authored, volume of work, standards and templates being used, approach to writing style and communication plus any existing use of technology.
Most organisations in Life Sciences have a technology function that can partner with medical writing to own the burden of technology assessment; most usually assessing core IT-related functional requirements to ensure a provider can pass the muster of technical assessments.
Additional support from the IT function plus the data sciences function can ensure the AI supplier is delivering on their promises. Knowing whether the supplier is leveraging third-party large language models or strictly applying Robotics Process Automation with BPM process automation is important to understand both the supplier’s leverage and approach. An additional consideration is checking that there are controls to ensure that the outcomes from technology can be consistently produced. Apart from expertise in technology, domain and process, organisations need a technology supplier to demonstrate robustness in testing, consistency in performance of AI models, and assurances of business continuity in cases of any disruption in the underlying technology.
Knowing where you want to go:
Whether to go with Generative AI or a rules-based solution might be less important than knowing what you want to accomplish. At the very core, consider the impact you would like to effect. Is there an overload of work, is quality review a burden, do you want to standardise and simplify processes through technology, is the administration of content overshadowing the authoring efforts, do you want to reduce administrative activities to allow medical writers to focus on medical assessment and judgement? Indicating a metric such as efficiency or quality is a good starting point but should be further expanded with use-case examples of where and how those metrics can be impacted by technology.
Some flexibility and understanding of available technology should additionally be considered allowing you to access existing solutions versus waiting for something not yet invented. Prioritising the metrics you want to impact will allow you to stay focused on your assessment of solutions, and using a scorecard can allow you to assess technology in terms of the metrics you want to impact. Taking this one step further, the provider of the solution should at least understand how the solution is impacting your metric and be able to quantify a business impact. This scorecard and business impact calculation will allow you to independently derive a Return on Investment (ROI) once cost becomes part of the conversation.
Assessing a new solution on the benefits for the team and impact on the process of medical writing will ensure your decision-making is not burdened by technical decision points (e.g. API integration) that will add value eventually but should not drive your decision-making in how technology will help the medical writing team.
Setting a course:
Your journey in technology might be very early on or complicated by a network of existing technologies. You might have a process you want to impact immediately (such as quality review, content reuse, editorial actions) or want to focus on a specific document. Knowing your priority is important but also understanding if you need a roadmap of capability (e.g. adding document types / addressing other steps of the medical writing process) should be considered. Whether your roadmap is short and with limited documents and milestones or long and intended to drive value across the content continuum can ensure you select a partner that is prepared not only with solutions but also the same vision that you have.
Either way, establishing a plan, securing stakeholder buy-in from partners in IT, ensuring there is executive sponsorship and medical writing subject matter experts prepared to evaluate solutions will ensure you build the right team to drive a collective decision. Budget naturally enters into the equation but should not be a starting point as this constrains the evaluation process. Cost and budget are worthy discussions once other considerations have been properly assessed.
If you are looking to consider a roadmap-based approach, be prepared to pilot a solution as an opportunity to evaluate its effectiveness. When Generative AI is introduced, medical writing resources often need training on Generative AI. Ask your partner if they can support this. Training will reduce the potential distrust and ensure medical writers maintain an open mind to the advantages the technology brings.
Another important consideration in the evaluation of a pilot solution is the implementation roadmap: is it considering a series of documents and solutions in a roadmap approach? Planning out the sequence of releases of either additional documents or capability should include ongoing assessment of the outcome experience and impact. Be sure to calculate this before you move on to the next release. The long-term view and the solution components should in many ways be interconnected so that if you enable a specific document, this contributes to the automation of another document further in your roadmap. This can ensure that the goal of automating the content continuum is logical and sequenced, and solutions can both build onto other documents and in some cases be repurposed for different documents as well.
Medical writers should assess if the solution builds confidence in the accuracy of the technology’s output. If the technology is not able to show how it is performing an activity that changes text or data then the perceived risk by a medical writer is high and decreases the value of the system. Medical writers also do not benefit from poorly written text, so unless the technology is able to apply best practices in medical writing then any text written by a machine will have to be re-written by the medical writer, negating the perceived benefit. Solutions that can assist the medical writer in the decision-making process have the greatest chance of adoption since they keep the medical writer in control of the final document.
Medical writers value the practitioner’s approach to authoring, applying best practices that ensure consistency in the delivery of the message. Technology, however, may require new ways of working to drive acceleration and efficiency. Medical writers don’t always author a complete document in one day and they need to pick up where they left off or even create an environment where their peers can pick up where they stopped. This must be factored into the assessment of effective or ineffective writing solutions.
Reaching your destination:
Keep in mind that any change that has an impact on the process may require upskilling of resources and training to ensure compliance. As each incremental technology solution in your roadmap builds on the last, be prepared to include a team of technical, subject matter, business process, and executive sponsors in the initial planning and incremental evaluation of your content continuum roadmap. Doing this will ensure both quick wins and effective long-term outcomes as you achieve your content automation goal.
