My work has that loose unpredictability I so love. Every project is a whole new world to step into—and out of, back into that “anything goes” departure lounge that is my professional life.
Last summer, an on-and-off client contacted me about some blog posts for their site. This was exciting for two reasons and daunting for one. First, this was a company that had never been late paying an invoice, whose professionalism and accountability were beyond reproach. Second, I was being asked to write about something entirely new to me, which tends to be my favorite kind of thing to jump into. The scary third thing had to do with the topics in question, because I was being asked to write about AI as used in IDP, RTP, and RTA in KYC, AML and LIBOR-related business applications. Oh, you don’t know what all those acronyms stand for? Exactly. Neither did I.
I thought back to some failed negotiations a year earlier with a different company, one that implements e-commerce IT architecture solutions. They wanted English writing for their website, but concluded that my lack of background in their niche was a decisive point against me. Not so here. This material is new to me, I admitted. We trust you’ll figure it out, they replied. I got a list of topics and each came with the name of a source I could call. I planned how to turn my inexperience into an advantage by using curiosity to guide me toward discussing complex concepts in simple, illuminating ways. It was exactly what I had offered those IT folks a while ago, only now I was getting the gig. Finally, a robust biz-tech startup was asking me to the dance.
Sometimes fulfilling work can get boring. Not drag-me-down-make-it-stop boring, but tedious and difficult. And satisfying work, even when it’s dazzling, can still require multiple dull rounds of rewriting and editing. Like when writer and client don’t see eye to eye stylistically or strategically. Or, in the case of tough tech talk, before the writer groks the deets. In this case, however, the big reveal is all about how fascinating and satisfying I have found this hardcore business writing with its soft-sell side—and how effective and effortless my communication has been with the people running the show.
That my deep dive into deep learning was an out-of-the-gate fit for what Applica needed was part luck, part chemistry, part more of that same biz savvy that makes this company one to watch. Evidently, they just don’t waste time delegating the writing to the wrong person.
Finding the process exciting is one crucial thing. Knowing that my client is happy is another. You know what else is great? Most companies might require a ghostwriter to sign an NDA, stay quiet, and just get the job done. This, what I get to do, here? Now that’s the real-real way to make work a win-win for both writer and client.
I’m providing links to some of the Applica.ai posts I’ve created so far, plus sample excerpts to give you a sense of the language and scope of the writing I get to hone on this exciting new journey.
Note: Applica.ai features content by various writers. I have written eleven blog posts to date. I am the author of the excerpts featured in this section, but Applica is the rightful copyright holder. Posted with permission.
How does intelligent document processing help a company reduce risk?
Our selective focus makes people great at tasks requiring problem solving toward a specific goal, but relatively incompetent at scanning text with a permanently alert and open mind. Applica has that inexhaustible attention-agnostic ability to not miss a thing. It never spaces out, never dozes off, never allows thoughts to wander. This means less risk of dropped data points or missed opportunities.
Harnessing AI for next-level AML
In an increasingly cash-wary world, money laundering on an international scale is now a lifeline for practically all drug cartel operators, terrorist organizations, human trafficking rings, arms dealers, art thieves, and sellers of conflict diamonds. Additionally, it involves others, whose criminality is less like an action movie plot and much more simply about cooking the books. In most cases, a matryoshka-doll structure of shell companies is involved, as well as telltale—though often obscure and complex—patterns of multiple small transactions. Fraudsters move money many times between sub-companies and they move amounts just under the threshold figure flagged by an institution for mandatory review. Though much of the time detailed analysis at the client onboarding phase can alert a bank’s risk analysis team to potential fraud line activity, needles do get lost in haystacks—even with the best people in play.
Debt collection gets a vital assist from AI
So far, debt collection has been the domain of banks and specialized collection agencies. Companies with customers in the red have generally outsourced the collecting, or they have sold off their debt at some loss to their bottom line (seventy cents on the dollar, say). It is an exciting question in these dynamic times whether the ability to deploy intelligent automation to expedite the collecting of debts will make some companies hold on to their non-paying customers’ accounts rather than make them someone else’s problem. After all, there is tremendous profit lying dormant in every debt portfolio. And it’s profit you’re owed. You just need the right tools.
Optimizing customer care with intelligent automation
Among the industries most often inundated by the proportion of customer care claims are banking and other financial services, telecom, streaming services, hospitals, medical networks, and insurance providers. However, the burden of claim processing is sector-blind, and clients who stand to benefit from intelligent document automation include not just wide-reaching service-based businesses, but also non-profits, educational institutions, and government programs. The key question is, will your company, organization or office benefit from custom-tailored, scalable measures that simplify the processing of user claims to reduce the workload on your staff by over half? And what a half it is: all the mundane, repetitive tasks are handled via Applica RTA, while humans at every skill level are deployed to do what they do best: engage with the types of situations that hold interest and allow their input to matter.
Deep learning—the key to next-generation document processing
Acquiring knowledge, problem solving, communicating, extrapolating, organizing information, translation—these are all natural modalities of the human brain. Now digital structures are replicating the behaviors of the mind contained within this sophisticated brain. In tech speak, AI is everything that fits this definition. In modern business use, AI is broad term encompassing a vast range of workflow solutions that deploy processes designed to mimic cognitive function. Among them are deep learning and machine learning, which, now you know, are the equivalent of a really fast and precise human knowledge worker who can think for herself and a fast, precise parrot. The latter is good at answering the questions you knew to ask. The former is free to address any question that may wind up relevant in the future. In fact, it might even answer some you didn’t know you had.
Before automation comes digitization
No, we don’t deploy an army of humanoid bots that handle your paperwork physically. First of all, outfitting this sci-fi platoon with limbs and eyes would be a waste of your money. Second of all, in practice it’s still too soon for this kind of the-future-is-here theater. Dig deep enough and you’ll discover that robotic arms and scanner-retinas are prohibitively expensive to produce and maintain, and it’s still a challenge to get an anthropomorphic machine not to trip on a pen lying on the floor, let alone fumble with all those sheets of paper. AI designed for scrutiny and analysis of text-based documents is purely a software phenomenon.
Rethink the contract management process with layout-aware automation
Contracts are, in the parlance of the automated text processing category, perfect examples of what’s called “unstructured” text. We may look at a contract and see block caps, bold subheadings, italics, bullet points, and appendices, but this type of editorial structure is not what is invoked by the industry definitions of “structured” or even “semi-structured” text.
Scrutinizing contracts for important information is tedious and error-prone work, which hinges on precisely the kind of needle-in-a-haystack vigilance that people lack but machines have in spades. And the greatest value does not come from expedient searches for known data points—though this would be challenging enough given the complexity of legal language and the amount of redundancy typical of so-called legalese, famous for inflating character counts considerably. Instead, the value lies in finding facts that are easy to miss, because they are buried in small print and because no one knows they are there in the first place.
AI, factoring, and good tidings all around
So what is factoring, in practical terms? It is the buying of unpaid invoices at a discount, usually of two to six percent, in order to profit from collecting on the full invoice rate. It allows a financial institution—with its sixth sense for optimizing risk and its well-oiled debt collection machine—to symbiotically support a company with day-to-day cash flow needs.
For a factor, establishing the viability of hundreds or thousands of invoices can require following just as many breadcrumb trails. From fact checking company data to verifying lien release agreements, the onboarding process requires the kind of error-proof vigilance and all-seeing exactness that humans just can’t deliver on par with robotic solutions. Plain run-of-the-mill KYC is tough enough in most instances, with financial institutions striving to leave no stone unturned as they background check a potential client. KYC in factoring, however, means you’re essentially doing wholesale KYC for all your client’s customers, too. You bet AI comes in handy.
Illustrations help make dense text more inviting, but who am I kidding, this post isn’t going to attract anyone it might scare away. Still, such is my website structure that every blog entry requires an image to appear in the intro link on the landing page. (And such is my je ne sais quoi that I need it be a photo from my own archive.)
Rubik’s cubes. Complicated things. A building under construction. More origami. A too-perfect cake. It all felt like the wrong cliché. So I went through my entire archive, bit by bit, over days, year by year, moving back in time, until I was jolted into finding… the right cliché. A bee in a patch of lavender at the Berlin Botanische Garten, June, 2007. It’s got work, it’s got pleasure, it’s got that in-the-zone quality, it’s got lust for life. And it reaches way back to a forgotten time and integrates that memory and that energy with the here-and-now.
And of course the minute I had my photo, the writing flowed.
Full list of existing Applica blog posts ghostwritten by me, updated on 2020 03 14.
https://www.applica.ai/post/understanding-the-deep-learning-generative-language-model-part-1
https://www.applica.ai/post/understanding-the-deep-learning-generative-language-model-part-2
https://www.applica.ai/post/are-you-prepared-for-omicron
https://www.applica.ai/post/gaining-the-loss-run-advantage
https://www.applica.ai/post/applica-talks-to-edward-benson-about-the-vast-world-of-text-automation
https://www.applica.ai/post/introducing-tilt
https://www.applica.ai/post/why-choose-applica
https://www.applica.ai/post/how-intelligent-document-processing-helps-reduce-risk
https://www.applica.ai/post/deep-learning-the-key-to-next- generation-document-processing
https://www.applica.ai/post/debt-collection-gets-a-vital-assist-from-ai
https://www.applica.ai/post/rethink-the-contract-management-process-with-layout-aware-automation
https://www.applica.ai/post/automated-kyc-is-better-kyc
https://www.applica.ai/post/gut-renovating-mortgage-lending-using-ai
https://www.applica.ai/post/optimizing-customer-care-with-intelligent-automation
https://www.applica.ai/post/harnessing-ai-for-next-level-aml