Sinclair Schuller February 28, 2023
“Hey <insert your favorite name here>,
One of the things I’ve been focused on lately is tightening up our messaging. Would you agree with the statement that ‘Nuvalence builds mission critical software for the world’s most ambitious organizations’?”
That’s an exact question I’ve asked many of our clients. I was surprised at what most of them honed in on: our use of the word ‘software.’ They took issue with that word via some paraphrasing of “You all focus on so much more. You focus on holistic delivery; on a solution” The reason they reacted that way is because we do much more than write code. We advise on business strategy and product strategy. We define their core software architecture and build what we architect. We also apply expert product management shrewdness in support of making sure that they launch the right product at the right time.
As of late, our clients (and all of Earth) are asking if it’s (a) too early for them to consider integrating with a Conversational AI (or ‘CAI’ – like ChatGPT), and if not, (b) how should they start thinking about integrating?
Question (a) is easy: start now. We’re entering the exponential phase of AIs adoption and a solution space experiencing exponential growth as well. Waiting will result in a “too late” scenario. My co-founder Rakesh wrote this in a recent post: “If your digital transformation strategy does not deeply consider the impact of AI, it’s dead on arrival.” Question (b) is more difficult. That’s a product strategy question.
Although our advice on the question “How should I integrate with ChatGPT” is hyper-local to each client’s domain and particular set of circumstances, categorically, there are some themes emerging:
1. Improving customer satisfaction through domain-specific, generative-question answering:
Customer satisfaction is intimately tied to customer service. To put this into perspective, just think of the penalty vs. reward spectrum established by these facts: 32% of U.S. consumers who have experienced bad customer service switched brands, and in Latin America, that figure is 49% (PwC, 2018). Meanwhile, a 5% rise in customer retention can yield a 25% increase in profit (HubSpot, 2021). In the public sector, despite the captive audience, the brand damage is likely worse for negative events given that end users can’t walk away.
Now imagine a highly knowledgeable brand, product and service expert able to help your customer with no wait times, who’s never (ok, rarely) in a bad mood, comprehends nearly every question, and is available 24×7. That’s what CAI can give you – a huge boost in “interaction to positive completion.” With generation augmented by data and knowledge specific to your company, products, services and general domain (tapping into a framework like LangChain), you’ll likely experience drastic profit and cost leverage, all while increasing brand loyalty. Still interested in maintaining a human touch? Leverage CAI as an “AI copilot,” reducing errors and mean time to issue resolution, while providing better service outcomes. Whether you’re a bank, human capital service provider, healthcare, or part of any other vertical, this is the fastest path to an MVP.
2. Improving customer outcomes by providing attentive, holistic, & successful ‘action-completion’:
Rather than just providing responses to queries, you’d likely create even more business leverage by wiring ChatGPT in a way where it can kick off hyper-specific workflows. Imagine, for example, that you’re an airline and you’re creating a CAI-based booking system. Rather than just booking your customer’s flight, you now provide centralized concierge services; based on the destination, dates, and preferences wired into the CAI, you can offer and book restaurants, hotel rooms, and anything else you decide to integrate. Kicking off workflows “to completion” enables you to move from being a “spoke” to a “hub,” disintermediating a significant chunk of the supply chain in your favor and becoming more strategic to your clients (in both B2B and B2C use cases). At Nuvalence, we know a thing or two about this: we’re actively working on an action-oriented overhaul of a large call center, leveraging AI at the core.
3. Improving quality of expert deliverables by reducing boiler-plate toil:
Not all business leverage needs to be customer-facing. CAIs can be used behind the scenes to get rid of the need to do undifferentiated boilerplate work. Leveraging CAIs will allow our world’s professionals – doctors, lawyers, and even software developers – the ability to focus their creative energy on upstream value instead of letting toil sap that creativity. But CAI developed boiler-plate isn’t your standard boiler-plate; it’s generated via prompts that allow for a “tailor fit” version of boiler-plate. Whether it’s generating a medical opinion, a specialized NDA, or creating boiler-plate code for an application to manage inventory supply and demand, deploying a bespoke CAI model can supercharge innovation at your organization.
4. Providing management and executives rapid access to organizational analytics:
Analytics has had a spotty history. On one hand, the value of analytics is clear to anyone with a pulse. On the other hand, the pipeline to go from what some executive wants to actually having it is littered with complexity, impacting ROI and time-to-value. CAI can help by wiring up your corporate data, ensuring that there are, as John Feminella put it, “…clear boundaries between the domains of interest,” and exposing analytics to management via natural language queries like “What’s our inventory turnover this month?” or “How have deadhead miles compared this month to last?”
These categories are just scratching the surface. If you’re looking for a strategy or architecture sounding board, or need help implementing, contact our AI practice area – I promise, you won’t be disappointed.