AI is changing the way nearly everyone works.
Take the researchers who used AI to read ancient scrolls for the first time, or the AI “models” on fashion runways, or the scientists using AI to run hundreds of “what if” scenarios to predict natural disasters.
AI’s role in the workplace will only increase — McKinsey & Company found that over the next three years, 92 percent of companies plan to increase their AI investments — which is why Built In found four New York City tech professionals who are using the technology to work smarter, not harder. Here’s what they had to say:
CAIS is an alternative investment platform for independent financial advisors.
What types of products or services does your engineering team build? What problem are you solving for customers?
Our users often need to sift through a long menu of alternative investment products to extract information such as performance data, sector allocation and redemption terms. Relying on our sales or customer service teams to help locate the information often slows this process down.
My team is focused on removing these friction points by building an AI-powered chat interface that integrates across our platform. It enables users to discover, compare and explore alternative investments available through the platform by simply asking natural-language questions.
Tell us about a recent project where your team used AI as a tool. What was it meant to accomplish? How did you use AI to assist?
For our product chat interface, we recently created the front-end app from scratch and AI proved useful in several ways.
One challenge was learning new technologies, such as Google’s Agent Development Kit (ADK). Using GPT-5 to ideate, explore architectural choices and understand how different components fit together was invaluable. Instead of parsing documentation and relying on trial and error, I could work through ideas interactively with the model. That back-and-forth sometimes surfaced approaches I might not have considered on my own, ultimately leading to a stronger implementation and, arguably, a more enjoyable development process, much like pair programming.
Because we were working under tight deadlines, maintaining velocity was crucial. Claude Code, an agentic coding tool, helped us move beyond boilerplate quickly and deploy a functional prototype to our infrastructure for immediate testing. It also made it easier to adapt to large architectural changes, common when working with emerging technologies, allowing us to iterate and test different strategies much faster.
What would that project have looked like if you didn’t have AI as a tool to use?
Without AI, setting up the app’s boilerplate, configuring deployments and building a proof-of-concept skeleton would have taken several extra days. Or, it would have required more developer support for areas I’m less familiar with, like our infrastructure repositories. Agentic tools helped me navigate those repositories efficiently and even ran multiple instances across domains, accelerating progress.
Beyond the setup stage, AI remained valuable. It freed time to refine the product by reducing time spent on business logic; enabled rapid iteration through frequent architectural changes common in greenfield projects; and supported fast, parallel exploration of new features and UX ideas, especially with limited product and design resources.
Overall, AI has shifted my focus away from mechanical implementation to creative problem solving. By compressing build time, it gives me more room to think critically about how to deliver better user experiences.
Gynger is an embedded financing platform built for buyers and sellers of technology that simplifies the end-to-end purchasing process by enabling businesses to pay, finance and manage all technology expenses from one platform.
What types of products or services does your engineering team build? What problem are you solving for customers?
At Gynger, we are building the only accounts receivable platform with embedded financing specifically designed for the technology sector.We solve a core B2B problem: buyers want flexible payment terms, sellers need cash upfront. Our platform lets tech vendors offer custom terms to customers while getting paid immediately.
We offer three products. Gynger Receivables is an AR automation giving finance teams receivables visibility, automated collections and cash flow insights. Gynger Capital is a financing engine providing upfront vendor payment and flexible buyer terms. Involves credit decisioning, pre approvals flows and qualification, lending limits and payment processing. Gynger Pay provides a seamless offer-to-payment flow handling invoice presentation through payment processing.
Tell us about a recent project where your team used AI as a tool. What was it meant to accomplish? How did you use AI to assist?
We use AI extensively across our product team, from writing and grooming tickets to reviewing PRs. Developers choose whatever tools provide them the most value. Personally, I lean heavily on Claude Code for planning and debugging work. There are few bug tickets I tackle without first asking Claude what the issue might be.
AI coding tools evolve rapidly and we maintain an open mind about leveraging them responsibly while shipping production-ready code faster.
You ask about a specific project, but the reality is AI is embedded in all of them at some level. It’s not a separate proof of concept or experimental consideration, it’s simply part of our daily development process. With the advent of model context protocol agents and third parties opening up their capabilities to models like ChatGPT and Claude, the leverage available to developers and product teams is only expanding.
What would that project have looked like if you didn’t have AI as a tool to use?
Since AI is part of our daily workflow, the impact is cumulative rather than project specific.
Without AI, I’d spend significantly more time context switching. Debugging our credit underwriting flow means tracking down logs and tracing data through services. With AI tools, I describe symptoms, paste code and get a hypothesis in seconds, collapsing 20 minutes of investigation into two.
Code reviews would be slower. AI catches obvious issues, missing error handling, race conditions, unclear naming, before I push. Without it, those surfaces in PR comments require another round trip.
Documentation takes less time. Translating vague bug reports into structured tickets with repro steps and acceptance criteria used to take 15 to 20 minutes. Now it’s more like five minutes. The bigger shift isn’t speed, it’s cognitive load. AI handles pattern matching and boilerplate, letting me focus on architecture and business logic that requires human judgment.
You still need strong fundamentals to know when AI is wrong. But with them, AI becomes a force multiplier for shipping quality code faster.
Trumid is a fast-growing fintech bringing leading-edge technology and product design to corporate bond trading.
What types of products or services does your engineering team build? What problem are you solving for customers?
Trumid combines deep expertise in corporate bond trading with agile technology to deliver innovative liquidity solutions for the fixed income market. One of our trading protocols, Swarms, reimagines how bond traders discover and execute opportunities — powered by high-performance interfaces designed to make the experience fast, intuitive and seamless.
Blending cutting-edge design with deep market insight, Swarms tackles one of bond trading’s biggest challenges, liquidity — finding reliable buyers and sellers in a fragmented market.
Traditionally, traders source liquidity across venues, potentially revealing intent and driving up costs. Swarms flips that model, bringing participants together in synchronized, anonymous sessions that create fresh liquidity opportunities throughout the day.
We design lightning-fast interfaces for high-stakes environments where milliseconds and precision matter. Our goal is to eliminate friction and empower users to operate at market speed — efficiently and securely.
Tell us about a recent project where your team used AI as a tool. What was it meant to accomplish? How did you use AI to assist?
AI has become a core collaborator across our teams, powering everything from idea generation to production features. Two projects show its impact: We built Chef, an internal Slackbot designed and refined by Trumid Product Analyst Henry Hobin, to cut through the noise of our bond platform and deliver actionable insights directly to sales, with no dashboards required. AI coding assistants help turn rough ideas into prototypes in minutes, letting us iterate quickly on what works. That speed helps us push beyond the typical “good enough” nature of internal tools, fine-tuning alerts for real impact.
We've also used AI to automate manual trade workflows reducing errors and saving time. Together, these innovations show how AI accelerates creativity and helps us deliver smarter, faster tools.
What would that project have looked like if you didn’t have AI as a tool to use?
Without AI, projects like Chef and other recent initiatives might never have happened. Parsing messy chat logs or experimenting quickly with new ideas would’ve required complex, brittle code that took weeks to build and maintain — work that often outweighed the value.
AI changed that equation. It handles unstructured language effortlessly, allowing us to prototype in hours, not weeks. Instead of debating what’s feasible, we test and learn by running small experiments, keeping what works and discarding what doesn’t.
That shift has redefined how we think about innovation. We’re no longer constrained by what’s “easy to code;” we’re guided by the experience we want to create. AI has made Trumid’s development process faster, more experimental and more human, enabling us to ship more, learn faster and turn once “nice-to-have” ideas into real, valuable tools.
Agentio enables marketers to buy creator led ads as easily and scalably as Meta or Google ads — automating sourcing, pricing, contracting, delivery and measurement — so work that took months now happens in minutes.
What types of products or services does your engineering team build? What problem are you solving for customers?
Agentio is an AI-powered platform for creator marketing that plans, executes and optimizes creator-led advertising for brands. Built on foundation models and autonomous workflows, Agentio turns a brand’s objectives and context into end-to-end campaign actions — strategy generation, creator and content sourcing and approvals, activation and continuous performance optimization — while learning from results to compound efficiency and outcomes over time.
One example of our AI capabilities is our content review system. Traditionally, when a creator submits a video draft, brands need to manually watch it and verify the creator followed the brief — checking talking points, product placement, messaging and dozens of other requirements. This review process was a major bottleneck, often taking days per video. Our AI-powered content review system uses cutting edge vision models to automatically watch video drafts and evaluate them against the brand’s brief in real-time. This helps brands significantly speed up their review process, transforming what was once the slowest, most manual part of creator sponsorships into an automated, scalable operation.
Tell us about a recent project where your team used AI as a tool. What was it meant to accomplish? How did you use AI to assist?
One of the most exciting recent projects is making data access truly self-serve across the company. As we’ve grown, the volume and complexity of data questions from different teams has grown. Things like campaign performance breakdowns, creator comparisons and revenue attribution used to always go through engineering or data science. That created friction and delays.
To solve that, we prototyped a system that adds MCP functionality directly to our API. That lets reasoning-capable models act as an interface between people and our data. Instead of writing SQL or waiting for someone technical to pull numbers, anyone can now ask questions in plain language, like “Which creator campaigns had the highest view through rate last quarter?” and get back real, structured results right away.
The impact has been huge. Non-technical teammates can move faster and explore data on their own and engineers spend more time building instead of responding to ad-hoc requests. It’s opened new ways to use the powerful analysis capabilities of AI. By pairing our domain-specific data with reasoning models, we’re giving everyone at the company a smarter, faster way to make decisions.
What would that project have looked like if you didn’t have AI as a tool to use?
If we had built this without AI, it would have looked completely different. The only real option would have been to build many custom dashboards and canned reports for every team. That would mean months of engineering and data work up front and a constant backlog of new requests whenever someone needed a slightly different view. It also would have locked people into whatever questions we had anticipated ahead of time, instead of letting them explore data freely.
With AI in the loop, we didn’t have to solve every edge case ourselves. The reasoning layer handles the complexity and interpretation that would otherwise require custom code or human involvement. That meant we could prototype something valuable in weeks vs quarters and it continues to improve over time as models get more capable.
AI is changing how we think about building software. We’re starting to design systems around collaboration between people and models rather than static interfaces. That lets us move faster, build more flexible tools and give each team at Agentio more leverage without adding headcount. It’s becoming a core part of how we work, not just something we layer on after the fact.
