 |
|
|
|||||
In September 2023, we thought we had done something revolutionary. Helios AI became the first company in our industry to launch a generative AI agent. We called her Cersi. She was designed to help food companies understand the climate risks threatening their agricultural supply chains. She was powerful, intuitive, years ahead of the curveand almost completely ignored. At the time, ChatGPT had just exploded onto the scene, and the hype around AI was deafening. Headlines promised that AI would transform every corner of business. Venture capital poured into the sector. But hype doesnt always translate into real-world useespecially in industries that arent built to adopt change quickly. Food procurement, where billion-dollar decisions hinge on weather patterns and multiyear contracts, not to mention generational relationships and personal rolodexes, is one of those industries. In a space where legacy companies tout decades of experience, newness isnt always a boon. We assumed that if we built something technically advanced, adoption would follow. We were wrong. Why our first AI agent failed Cersi was conceived as a conversational assistant. Type a question about your supply chain into a chat box, and she would pull from Helioss massive dataset to provide a rich, insightful answer. It sounded futuristic, and on technical merit, it worked. But the glaring problem was, it didnt fit the way our customers actually worked. First, they didnt want to chat. Procurement executives, commodity traders, and risk managers wanted structured, decision-ready insights. They wanted something they could paste into a slide deck for the CFO, or drop into an email to their sourcing team. A conversational AI, however clever or time-saving, wasnt the format they needed. Second, most users kept asking the same five questions again and again. That told us something important: their needs werent open-ended. They wanted repeatable analysis, standardized for their business, not a new way to brainstorm with an algorithm. Finally, and perhaps most critically, Cersi sometimes produced answers that were technically correct but felt superficial. In an industry where credibility and precision matter, close enough wasnt good enough. What we learned The biggest lesson was simple but humbling: AI itself isnt the productthe outcome is. In other words, customers dont care how elegant your models are. They care if your product saves them time, reduces their risk, or helps them make a better decision in a high-stakes environment. We had fallen into the classic founder trap of building something because we could, not because our customers had asked for it. But we had still built something exceptional and groundbreaking in its industry, beyond what the standard benchmarks in the field were capable of. So a few months after Cersis underwhelming debut, we reimagined her role. Instead of a front-end chatbot, she became a behind-the-scenes analyst. Rather than flowing conversation, she was rebuilt to generate thousands of custom agricultural reports every montheach tailored to a customers commodities, sourcing regions, and climate risks. These reports would land directly in a customers inbox or workflow, in the format they need, where they can actually use them. In shifting Cersi from a face to a function,” adoption skyrocketed. The AI didnt become less powerful, it just became better integrated. In making our AI less visible, it became much more useful. As the saying goes, good design should be invisible. Building on that success, last month, almost two years after Cersis flop, we launched Helios Horizon, the first multi-agent platform in our industry. Its designed to handle complex, interconnected tasks that a single agent couldnt. Instead of one assistant, Horizon uses a coordinated set of AI agents that monitor risks, flag disruptions, and deliver analysis specific to each customers supply chain. This level of advanced AI wouldve been hard to imagine back in 2023, but wed taken our lessons from Cersi to heart.The next wave of AI adoption will look different from the hype cycle of 2023. Enterprises arent asking whether AI is possible anymore. Theyre asking if its practical, trustworthy, and built to fit their workflows. And those are harder questions to answer. 3 takeaways for Founders 1. Great AI isnt enough The technology has to map directly to real workflows and be right-sized for the industry it serves. In 2023, most of our customers had barely used a chatbot, and they werent ready to experiment with one in their jobs. Today, after nearly two years of ChatGPT in the mainstream, familiarity is higher. We no longer have to teach people what a natural-language interface isbut we still have to prove why it matters in their world. 2. Users dont want AIthey want its benefits Our customers dont wake up excited to use AI. They wake up trying to secure coffee from Brazil or wheat from Kansas before climate shocks, trade restrictions, or shipping delays throw their budgets into chaos. What matters most is the outcome: did the system save them hours of manual analysis? Did it prevent a costly mistake? Thats why one of Horizons most popular features is simple: it shows customers how many hours weve saved them. Time is a currency they value even more than insights. 3. The best AI isnt visible The future of AI in the enterprise isnt necessarily chatbots or flashy dashboards. Often, the most valuable AI disappears into the background, quietly doing the work and surfacing results at the right time, in the right format. Cersi failing taught us that quieter AI can be more powerful than any buzzy avatar, co-pilot, or assistant. Horizon was built with that in mind. For founders building in this space, the lesson we learned is clear: resist the temptation to build AI for the hype cycle. In fact, do the very oppositebuild your AI not to be the flashy new thing, but to be so good that it becomes invisible.
Category:
E-Commerce
As a learning designer at Zapier, I used to spend my days helping my teammates learn: I built and led trainings, created enablement resources, and helped folks better understand how their work contributed to company strategy. Now, I sit inside our HR team as an AI automation engineer. But the through line is the same: I still help my teammates (and now customers, too!) do their best work. What is an AI automation engineer? AI automation engineer sounds like a vague title, so here’s the job, plainly: I embed with a team (HR, in my case), spot opportunities to enhance the team’s work, and build AI-powered workflows that jump on those opportunities. The goal is to create measurable improvements that free my teammates up for creativity, strategy, and connection. I think we’ll be seeing this title pop up more and more as time goes on. For example, instead of hiring a new content writer, content marketing teams might look for AI automation engineers with a strong eye for content. Instead of a new junior coder, engineering teams might look for an AI automation engineer with a technical background. Why the AI automation engineer role matters Lots of teams see AI’s potential but get stuck turning ideas into action. The gap is less about technology and more about translation: understanding how a real process works today, where it fails, what data is safe to use, and what “better” even looks like. AI automation engineers close that gap. We prototype fast using tools like Zapier, ChatGPT, Airtable, and Cursor, then we harden those prototypes into reliable internal tools. In HR, that looks like: Reducing the back-and-forth between recruiting, interviewers, and candidates Auto-summarizing interview debriefs so we can make decisions faster Keeping people data in sync across tools, with the right guardrails for privacy and compliance Giving folks self-serve answers to policy questions without losing the human touch And it is not just about helping my own team. A big part of my role is building repeatable HR workflows that we can share with our customers. When I design something for Zapiers people team, like an interview debrief summarizer or a self-serve policy bot, I’m also thinking about how it could work for other HR teams out in the world. Sometimes it even goes the other way: a customer use case inspires a workflow that we bring back inside Zapier. What I actually do week to week as an AI automation engineer AI automation engineer is a new type of role, and I’m even newer to it myself, but here’s a glimpse into how I spend my days. Triage workflows: I map how work really happens, quantify the cost of the current process, then rank opportunities. I’m looking for spots where I can have a big impact. Prototype quickly: Once I know where I want to help, I build small, testable versions using Zapier and other AI tools. Embed with the team: I sit with the people doing the work. We try the prototype in the real flow and adjust prompts and logic. We document what to trust and when to escalate to a human. Scale the AI automation: Once the workflow proves itself, I add error handling, retries, observability, and access controls. I create a runbook so the team can own it, not me. Teach and enable: I host short workshops, write playbooks, and pair with team leads so they can spot the next opportunity themselves (and know when not to use AI). Measure outcomes: I track hours saved, error rate reduction, cycle-time improvements, adoption, and the business outcome (e.g., faster time-to-hire, better candidate experience). Partner with sales: When we see an HR workflow that really moves the needle, I package it into a demo or playbook that our sales team can share with prospects. Sometimes I join those conversations directly to explain the workflow in plain language: why it works, what problems it solves, and what business results it drives. Share feedback upstream: Because I’m building on top of Zapier and AI tools every day, I run into edge cases, missing features, or things that could be smoother. I funnel that feedback back to our product team, often with concrete examples from both our people team and customers. It means the next version of Zapier is more aligned with how real HR teams actually work. For some examples of what I’ve built, read about my favorite agents or take a look at our HR AI automation playbook. What does it take to be an AI automation engineer? I don’t come from a technical background, but I’m a tinkerer, and I think that’s what makes me suited for this role. I’m comfortable building with no-code tools and love to ship solutions. Skills I’ve picked up along the way are prompt engineering, responsible AI practices, and understanding how to pick the right AI tools for the job. One of the most important parts of the role, though, is something I have a lot of experience with: enablement. I need to make sure the folks I’m building for understand how to make the most of these systems. One important thing to note: My focus is squarely on HR. That’s where I build, prototype, and enable. While I love seeing how AI automation engineers show up in marketing, IT, or engineering, my role is all about HR use cases. I help our people team work smarter, and I help our customers run stronger HR operations. But I’m also proof that you don’t need to be a software engineer to become an AI automation engineer. Here are some other folks from the Zapier community who I’d argue are AI automation engineers, each from a different background. Remote’s Marcus Saito (head of IT) used AI to auto-resolve 27.5% of IT tickets. This saved his team more than 2,200 days and $500,000 in hiring costs. Vendasta’s Jacob Sirrs (marketing operations specialist) used AI to automate sales workflows, save more than 282 workdays a year, and reclaim $1 million in revenue. ActiveCampaign’s Tabitha Jordan (manager of product education) implemented AI-powered lead enrichment to give the sales teams time to focus on high-value activities. Moving from learning & development into this roleas an AI automation engineer for HR hasn’t changed my mission. I still help people work better. If you’re AI-curious, start with the smallest annoying task you do every week. Fix that. Measure it. Then fix the next one.
Category:
E-Commerce
Long Beach Airport had a trailer problem. Long Beach’s quaint municipal airport originally opened in 1924 when airplanes flew using propellersand the art deco terminal hadn’t undergone a full-scale renovation since. Instead, it adapted to the increased spatial demands of late 20th and early 21st century air travel, like increased security screening and modern baggage handling, in a rather temporary way: trailers. “It was known as the trailer park airport,” says Michael Bohn, a partner at Studio One Eleven, a Long Beach-based architecture and design firm. “It just became a hodgepodge. You went down these crazy aisles, and through different trailers. They had vending machines for snacks. It was probably one of the worst experiences you could have.” In 2012, the city decided to do something about that. It launched a multiphase, $185 million renovation project. Two new concourse buildings were added, making it more feasible for the airport to handle passengers for major airlines like Southwest and JetBlue. Concessions were expanded. A new welcome gateway was added. It was all intended to reset the airport in the public’s mind, moving it away from its jumbled past to becoming a more seamless gateway for traveler opting against the nearby behemoth of Los Angeles International Airport. But the trailers were still making up key parts of the airport’s operations. [Photo: courtesy Studio One Eleven] “Trailer park airport” no more Studio One Eleven stepped in to rethink the space around the main terminal building and to do away with the trailers once and for all. The firm led the historic renovation and seismic upgrading of the terminal building, designed in a streamlined style and adorned with WPA-era artwork. The project also included a large-scale enhancement of the terminal’s public realm, much of which had been taken over by trailers and other ad hoc building annexes and airport infrastructure. [Photo: courtesy Studio One Eleven] “We said, ‘what if you could pull this stuff away and create a negative space instead of all this clutter?'” says Kirk Keller, principal landscape architect at Studio One Eleven. The designers moved IT equipment into the basement, and relocated the baggage handling infrastructure behind the scenes. “It was really trying to carve away space for people.” [Photo: courtesy Studio One Eleven] That opened up new space for a more open terminal experience and, rarely for an airport, outdoor terrace space once passengers make their way through security. “We look at the space between buildings as being just as important as the architecture itself,” says Bohn. Their design interventions have gotten rid of the trailer park problem, and helped turn Long Beach Airport into one of the most beloved airports in the United States. A recent Washington Post ranking of the top 50 airports placed it as the second best in the nation, behind only Portland’s elegant new mass timber terminal. [Photo: courtesy Studio One Eleven] New outdoor space Outdoor space became a key focus for the project. Once holding the overflow services and equipment that created the airport’s trailer problem, space that exists between the historic terminal and the two new concourse buildings became ripe for reinvention. “It was almost just interstitial space between these two concourses. It served no purpose,” says Bohn. Studio One Eleven reframed the space as a central plaza. [Photo: courtesy Studio One Eleven] Set between the cruise ship-esque facade of the historic terminal and the modern facilities leading into the secure section of the airport, the plaza has become a unique public space in the city, where people can greet arriving friends and family, access one of the airport’s local concessionaires, or simply catch views of airplanes taking off and landing. [Photo: courtesy Studio One Eleven] To keep it as open as possible, the designers used the region’s iconic palm trees as both landscaping and lighting infrastructure, while also webbing the space with an overhead catenary wire system to hold additional exterior lights. Keller says they’re meant to evoke the flight paths of airplanes and seabirds from Long Beach’s coastal environment. Long Beach-based Studio One Eleven was tuned into these local influences. The designers also knew that one of the airport’s biggest strengths was its relatively modest size. “We were just respecting that Long Beach doesn’t want to try to compete with LAX or Portland, or San Francisco,” says Bohn. “It’s got its charm, and we just wanted to build on that.”
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||