3 Critical Questions About Augmented Intelligence
For anyone to be successful in solving business needs in financial services with Augmented Intelligence solutions, there are three basic questions that need to be answered:
- What is Augmented Intelligence?
- Why Augmented Intelligence?
- Why Augmented Intelligence solutions now?
Let’s start at the top.
What is Augmented Intelligence?
At CognitiveScale, we believe that the real power of AI is not about replacing what humans do, but rather augmenting it. We call this Augmented Intelligence (AI), i.e. pairing humans with machines, to help them do their jobs better—and more efficiently. Existing tools typically focus on a very specific predefined task, or a process made up of several tasks, and this is a fundamental limitation of those approaches. Augmented Intelligence helps humans to do jobs better, not just tasks. It succeeds when better can be measured in terms of increases in:
- Speed (e.g. Do the same market research in less time)
- Scope (e.g. Use more varied sources of data, across unstructured and structured sources, both internal and third-party, for richer insights)
- Size (e.g. Analyze more data)
- Scale (e.g. Handle more clients in your book of business)
- Efficiency (e.g. Do the same amount of work in less of time)
- Accuracy (e.g. Read customers, prospects, and market signals better)
These are some Key Performance Indicators (KPIs) measuring improvements in how Augmented Intelligence is helping someone to do their actual job.
Let’s think about things differently:
Augmentation is more than modeling, and more than automation, because augmentation starts with the human in the loop, not with a process, or with a historical data set.
Augmented Intelligence flips the thinking from process-centric to human-centric: When spreadsheets were invented, we had to compress our needs to fit into a spreadsheet. When process engines came out, we had to compress our needs into rigid business processes that would sometimes involve people. With predictive analytics and machine learning, it is about pattern-matching and curve-fitting—applying the outcomes of these approaches to people is typically done manually, if at all. This is inherently brittle, as it is relatively hard to tailor these approaches for individuals, especially for those individuals whose needs differ and change over time.
Why Augmented Intelligence?
Your bank likely already has developers, data scientists, and software, so what is in it for you to use Augmented Intelligence with your existing IT? That existing software is typically data and process-focused, and as such, is too brittle to start with the individual that has ever-growing, ever-evolving needs and particulars. For example, if you learn new things about a client that fall outside of the existing fields in the customer database, adding that new data can often break an existing database-driven process engine. Changing a rigid database schema can be difficult (approvals and so on), then refactoring the process engine to use that data productively is even more work. Repeating that for every needed change, and you can start to see how that approach doesn’t scale.
If you’re now more amenable to the idea that a people-first approach could best accelerate your bank’s success, the next step is to answer how you would pick an Augmented Intelligence solution provider. To use anything on the “front lines” at the bank, the solution, as well as the solution’s provider, needs to be Enterprise-grade.
At a broad level, Enterprise-grade decision criteria will usually include:
- People (e.g. Having the right team to deliver a production solution, from design to DevOps, not just model building)
- Process (e.g. A repeatable, scalable, and responsible approach)
- Products (e.g. Tools to accelerate the Augmented Intelligence process from design through to production)
- Education (e.g. Able to teach companies to do as much of this themselves as makes sense)
- Services (e.g. Best Practices tailored to the unique situation of each company)
- Support (e.g. An end-to-end partner to help get things running, and then to keep them running, especially in heavily regulated industries like FS, where new technologies are harder to implement)
- Security (e.g. New solutions need to be architected and built to be secure, as well as compliant, and explainable, to protect the best interests of both the bank and its clients)
Why Augmented Intelligence?
The Chief Financial Officer will typically ask why the bank needs to spend any money now: “Why does the bank need this, and why now? What is the compelling event?”
Essentially, it’s to drive value, measured in three major ways:
- Raise Revenue (e.g. Drive more Assets Under Management to bring in extra fees)
- Cut Costs (e.g. Avoid more compliance-related fines, use human effort more efficiently)
- Mitigate Risk (e.g. Avoid more public disclosures, reduce client churn)
If an AI vendor can’t work with you to make a real difference in one or more of these, then it’s hard to justify to anyone why the bank should pay for it. While innovation labs at many of the banks can make a relatively small investment in pioneering technologies, if it’s not Enterprise-grade, it will be hard-pressed to move anything beyond the lab. It’s the key reason to focus on getting these solutions into production, and more importantly, into the hands of the folks in the front, middle, and back-office roles that directly drive the (compliant) revenue engines for the bank. One of our favorites is driving competitive advantage, empowering banks that will partner to get a head-start over those strict “build” shops (that are effectively software companies with one customer: themselves).
That’s a snapshot of journey to driving value in Financial Services with Augmented Intelligence. In my next blog, I’ll talk about “Hitting Limits with Machine Learning..How can Augmented Intelligence Help?”