The Future of Workflows in AI
In recent years, Artificial Intelligence (AI) has consistently proven its transformative potential in diverse industries, with data-heavy areas such as the banking sector no exception. Generative AI (Gen AI), perhaps the most promising variety of AI, has the potential to unlock unprecedented levels of productivity and efficiency. However, it requires careful, methodical implementation.
Gen AI operates via language models such as GPT-4, leveraging data fine-tuning and prompts to generate human-like text. It learns and synthesizes vast quantities of information, which it uses to deliver insightful outputs.
But you won’t get useful insights without meaningful human input and coherent datasets. GenAI, and the LLM architecture it’s built upon, is only one element of a mix that must be calibrated for optimal results.
In financial services (FS), it’s the workflow that makes this technology integral, for instance. Workflows dictate the sequence of processes, allowing the AI to offer the most utility. Workflows ensure the right human processes are in place, the right datasets are obtained, and that the AIs are fine-tuned to a highly detail-oriented and regulated domain.
How to Position AI for Maximal Benefit
Redefining our relationship with AI is an essential step. Instead of downgrading it to just another analytic tool, or hyping it as the answer to all ills, we need to view AI as more like an assistant, consultant, or coach.
For instance, an AI assistant can perform administrative tasks like setting up meetings, while an AI consultant could analyze market trends and recommend investment strategies. As a coach, AI can help employees improve their skills by providing personalized feedback.
Furthermore, reimagining FS workflows with Gen AI can yield surprising benefits. By creating a better user experience (UX), rather than just providing access points to powerful algorithms, AI will be a game-changer, empowering human analysts and providing hitherto unimagined insights.
UX, in this context, considers the human aspects of interaction with AI. It aims to make such experiences as seamless and intuitive as possible. GenAI then becomes an assistant that never forgets anything or an analyst you can always count on to be thorough and fully up to date.
Use Cases of AI in Financial Services
It’s fair to say that FS firms aren’t jumping on the GenAI bandwagon in a hurry. There may be a good reason for this, as an article by Bain Venture Capital identifies: “Financial services institutions will not be in the early adopters of generative AI writ large. Why? For the same reason [that] financial services institutions aren’t early adopters of most forms of technology: financial services institutions are conservative by design.”
This conservatism, of course, will surely be shaken by the impressive gains GenAI is already delivering.
In terms of practical use cases, we’ve tested AI applications in FS workflows and seen promising results. In productivity, Gen AI has shown capabilities in streamlining the four broad steps for research:
- Search
- Data extraction
- Data analysis
- Final document creation.
In the test cases we investigated, AI applied to interactive search, screening, and summarization resulted in 15% savings with high accuracy. AI in an FS context has proven beneficial in scrutinizing annual reports, industry briefings, earnings summaries, and broker reports.
Data extraction, another critical step in the research process, showed a remarkable 50% savings when we employed AI. Tools like Spreadsmart have proven effective in extracting financial information and Environmental, Social, and Governance (ESG) data. Such tools provide evidence to back up the decisions our clients make; in the high-risk environments of FS, that makes a big difference.
Gen AI also excels in knowledge management, allowing accurate repository searches without the need for manual tagging. Used alongside automated document creation, it can generate tailored reports in moments, which are maximally searchable, easily shared, stored, and retrieved.
Workflows Revisited with AI in Focus
Bolstered by these successes, we’ve begun to reimagine workflows based on the principles and capabilities of Gen AI, always with the end consumer user experience in mind.
One example is the deal process. Enhancements in reporting and the creation of the “pitchbook of the future” have made data more accessible to stakeholders. Data can now be tailored very precisely to connect with a chosen human audience.
You may no longer create a single pitchbook at all. Instead, you might offer potential clients an information resource they can navigate at will, with the AI delivering exactly the data they need to feel confident in their decision to join forces.
Similarly, in credit lending, we’re exploring how Gen AI could revolutionize the creation of credit memos, minimizing errors and risk. And while AI and automation may only achieve 90% accuracy, human experts can step in to supply that final 10%.
Human Expertise will never be Redundant
Despite these technological advances the importance of creative storytelling remains. Humans cannot be removed from the loop. Human empathy and emotional intelligence complement automation by providing context and interpretation, essentially serving as the ‘zero draft’ for all automated outputs.
Prompt complexity will become ever more demanding, requiring domain expertise, logical thinking, the ability to structure an argument, and a knowledge of what data sources will best evidence the case a specialist wants to make.
Another potential application of GenAI is in the creation of a ‘Library’ where rapid consumption of data/research could be possible, thereby saving time and improving decision-making. By building a resource of ready-made materials, analyses, and documents, the process of generating new outputs is reduced.
The library becomes an increasingly valuable asset over time, potentially even a monetizable one.
GenAI’s Economic Benefits in Financial Services
As suggested in the McKinsey report, “The economic potential of Generative AI: The next productivity frontier,” Gen AI has much to offer. As the authors state in that report: “Generative AI’s impact on productivity could add trillions of dollars in value to the global economy.”
It’s a popular misconception to believe that GenAI will replace humans or result in them becoming lazy. As GenAI apps get better and better, it will simply become more possible to entrust them with mundane and repetitive tasks. This will free up human analysts to think creatively and strategically.
FS executives will be in a better position to identify and pursue more opportunities, generate a greater number of leads, and devise whole new products, tailored to each client’s needs.
Fine-tuning is the future, in other words.
Workflows of the Future, with AI Built-In
By integrating GenAI into FS workflows, we can realize this potential, driving productivity, empowering creativity, and offering improved user experiences.
The future of workflows, thus, lies in the strategic deployment of AI, ensuring that it serves as a key collaborator, revolutionizing the way we work, and the way we view work.
