Startups that succeed in the agentic AI space are betting on vertical specialization, digital labor and new kinds of software primitives. Rather than broad platforms, these companies are zeroing in on deep domain challenges and embedding AI agents where judgment, context and autonomy matter most. Instead of retrofitting yesterday’s SaaS models, HOAi focuses on a labor-intensive, highly contextual domain: Homeowner association management. That clarity of focus enables the company to design agentic systems with three core components: cognitive reasoning engines, seamless integration with existing workflows and a flexible orchestration layer for agents. By targeting labor spend rather than IT budgets, startups such as HOAi create new categories of digital workers that operate alongside humans. This shift enables access to budgets that are 10–20 times larger than traditional enterprise IT, according to Haoyu Zha, founder and chief executive officer of HOAi. To distill the lessons from HOAi and similar innovators, here are five keys to building a successful agentic AI startup, according to Zha: Go vertical in nuanced markets: Specialized agents can capture untapped value in industries with unique operational needs. Follow the labor spend, not the IT: Labor budgets are significantly larger than IT budgets and far less saturated. Empower decisions over tasks: Build agents that enhance human judgment, not just automation. Decision intelligence is the new strategic edge. Rethink software: Go agentic: Don’t retrofit software-as-a-service blueprints. Design end-to-end workflows with autonomous, context-aware agents from the ground up. Visibility fuels viability: In a crowded market, discovery matters. Build brand awareness early or risk being invisible, regardless of how advanced your tech is.
Keys to creating an agentic AI business- Vertical specialization, chasing the labor spend as against enterprise IT budgets, focusing on cognitive reasoning and human judgement and designing end-to-end workflows
Startups that succeed in the agentic AI space are betting on vertical specialization, digital labor and new kinds of software primitives. Rather than broad platforms, these companies are zeroing in on deep domain challenges and embedding AI agents where judgment, context and autonomy matter most. Instead of retrofitting yesterday’s SaaS models, HOAi focuses on a labor-intensive, highly contextual domain: Homeowner association management. That clarity of focus enables the company to design agentic systems with three core components: cognitive reasoning engines, seamless integration with existing workflows and a flexible orchestration layer for agents. By targeting labor spend rather than IT budgets, startups such as HOAi create new categories of digital workers that operate alongside humans. This shift enables access to budgets that are 10–20 times larger than traditional enterprise IT, according to Haoyu Zha, founder and chief executive officer of HOAi. To distill the lessons from HOAi and similar innovators, here are five keys to building a successful agentic AI startup, according to Zha: Go vertical in nuanced markets: Specialized agents can capture untapped value in industries with unique operational needs. Follow the labor spend, not the IT: Labor budgets are significantly larger than IT budgets and far less saturated. Empower decisions over tasks: Build agents that enhance human judgment, not just automation. Decision intelligence is the new strategic edge. Rethink software: Go agentic: Don’t retrofit software-as-a-service blueprints. Design end-to-end workflows with autonomous, context-aware agents from the ground up. Visibility fuels viability: In a crowded market, discovery matters. Build brand awareness early or risk being invisible, regardless of how advanced your tech is.
Citi Appoints Wyatt Crowell as Head of North America Citi Commercial Bank; previously at HSBC as Head of U.S. Commercial Banking
Citi appointed Wyatt Crowell as Head of North America Citi Commercial Bank (NAM CCB). Wyatt will assume his role in May 2025, and will report to the Head of CCB, Tasnim Ghiawadwala and Citibank N.A. Chief Executive Officer (CEO) and Head of NAM, Sunil Garg. Based in New York, Wyatt joins Citi from HSBC where he has been the Head of U.S. Commercial Banking since 2015. Prior to joining HSBC, Wyatt was at Barclays and served as the Co-Head of U.K. and Ireland Corporate Banking Coverage as well as Head of Global Multinational Corporates. He started his career at J.P. Morgan in investment banking, holding various roles in M&A and mid-corporate investment banking. In his more than three decades in the banking industry, Wyatt has accumulated extensive knowledge of multiple client segments and has a deep understanding of banking products and solutions. He is a proven leader that drives growth with a focus on client experience and operational excellence. Wyatt will focus on the execution of the strategy and accelerate CCB’s growth in North America. He will propel the business through growing client relationships and delivering innovative client-centric solutions.Wyatt has an MBA with honors from Columbia University and a BA in Economics from Colgate University. Wyatt is on the Board of Directors for Eve’s Fund, a small non-profit organization promoting hope and wellness for young Native Americans living on and near the Navajo Nation in Arizona, New Mexico, and Utah.
UWM sets ambitious goal to reach $280 billion in mortgage production by 2028; more than double the $138 billion the lender originated in 2024
Mat Ishbia, CEO of United Wholesale Mortgage (UWM), the top U.S. mortgage lender, unveiled the company’s ambitious goal to reach $280 billion in mortgage production by 2028. That is more than double the $137.8 billion UWM originated in 2024, a year in which the wholesale lender posted a 28% year-over-year increase despite heightened competition and persistently high mortgage rates. The target also far exceeds the combined 2024 production of the year’s second- and third- largest lenders. Pennymac, second in the ranking and best known for its correspondent business, originated $115 billion in 2024, up 16.5% year over year. Rocket Mortgage, which does the bulk of its business as a direct-to-consumer lender, ranked third with $95.8 billion, a 25.9% increase from 2023. Ishbia has reiterated that UWM plans to achieve its growth organically, rather than through mergers and acquisitions like some of its peers. The lender has been steadily increasing its operational capacity and is positioning itself to benefit from refinancing opportunities when rates decline. UWM is also in hiring mode, with a long-term focus on growth. After peaking at 8,000 employees in 2021 and dipping to 6,000 in 2022, its workforce grew to 6,700 in 2023 and surged to 9,100 in 2024. Ishbia announced plans to promote 4,500 employees—about half of UWM’s current workforce—within the next three years. UWM is also investing heavily in technology, product innovation, and operational processes to better support its broker partners. Ishbia set a goal for the broker channel to reach a 33% market share within three years, up from its current 27.8%. In response to the competitive dynamics in the market, UWM has decided to bring its mortgage servicing operations in-house, following the termination of its relationship with Mr. Cooper after the latter’s $9.4 billion deal to sell to Rocket. UWM aims to achieve a 90% “perfect service” score from brokers and borrowers in three years.
Chime files for IPO; on back of active members growth of 82% since the first quarter of 2022 to 8.6 million, 67% of whom are using Chime as their primary account relationship
Chime has filed a registration statement on Form S-1 with the U.S. Securities and Exchange Commission relating to the proposed initial public offering of its Class A common stock. Chime has applied to list its Class A common stock on the Nasdaq Global Select Market under the ticker symbol “CHYM.” The number of shares to be offered and the price range for the proposed offering have not yet been determined. Morgan Stanley, Goldman Sachs & Co. LLC and J.P. Morgan will act as lead book-running managers for the proposed offering. Barclays will act as an additional book-running manager. In its Form S-1, Chime said it addresses “the most critical financial needs of everyday Americans,” including spending, liquidity, credit building, savings, community, support and safety. The company added that it has 8.6 million active members, up 82% since the first quarter of 2022, and that 67% of these active members use Chime as their primary account relationship. It noted that it is a technology company, not a bank, and that its bank partners provide its FDIC-insured accounts. “Looking ahead, with less than 5% adoption in our core target market, we see an enormous opportunity to grow for years to come,” Chime Co-Founder and CEO Chris Britt and Co-Founder Ryan King wrote in the Form S-1.
Goldman Sachs to roll out AI Assistant to all its workforce to serve as backbone of other such tools- Banker Copilot, Legend AI Query, Legend Copilot and Translate AI
Goldman Sachs has been building out its generative AI toolkit. The firm aims to release one of its tools, an AI assistant, to most staff this year. Here’s a look at five such tools — the promise of what they can do, plus who’s using them and how.
- GS AI Assistant. What it does: Goldman Sachs’ in-house version of ChatGPT. Think of the GS AI Assistant as a sidekick for Goldman employees. It uses a chat interface, similar to that of ChatGPT, but can pull its responses from the bank’s confidential data repository. Right now, it’s available to about 10,000 workers; the firm is intending to get it in the hands of the rest of the bank’s workforce of over 46,000 by the end of the year. It can perform a variety of functions, helping executives draft presentations and plan off-site meetings, or serving as a “personal tutor” for quant strategists.
- Banker Copilot. What it does: Streamlines some aspects of investment bankers’ jobs. Members of Goldman’s investment banking division are also set to get an AI boost with the bank’s so-called banker copilot, which makes access to high-level, protected data about matters like deal-making available to eligible users. Only a small group numbering in the dozens has access to it right now since it’s in the early stages of development. But the promise of what the AI assistant could represent for the banking business is hard to deny. Solomon himself has acknowledged the potential for AI to automate large chunks of tedious processes, like drafting S-1 regulatory disclosures for initial public offerings, for instance.
- Legend AI Query: What it does: A search tool that uses AI to navigate the bank’s vast repository of data. Goldman uses Legend, an open-source data management and governance platform. Accessing Goldman’s vast vault of knowledge used to require users to know what they were looking for — as well as where they were looking — ahead of time, using a tool called “Legend Query.” Think of this process as being like perusing dozens of stacks in a library, but without a librarian to help. Enter Legend AI Query, a query tool that saves time by tapping artificial intelligence to serve as that librarian.
- Legend Copilot. What it does: A fast-tracked way to upload data onto Legend, and keep the system organized. Legend Copilot, which launched in October, is a tool primarily designed for use by data engineers to maintain Legend’s infrastructure, and keep its information streams organized for others to access. Legend draws on data that originates in other databases, but still needs to be routed into the centralized Legend system.
- Translate AI: What it does: In-house language translation to and from English. As a global bank, Goldman Sachs has clients worldwide. Sometimes, those clients have a preferred language that’s not English. To reach clients in their non-English speaking language, the bank would historically outsource some of this translation work, but turnaround times could stretch into days.
Vanguard is deploying agentic AI that let users pull data from databases in natural language and conduct data lineage checks, which avoids a costly undertaking
Ryan Swann, Vanguard’s chief data analytics officer says the mutual fund giant is using data and AI not only to gain insight, but also to build enterprise-wide agility and accelerate value creation for itself and its clients. Data and analytics have a seat at the C-suite table at Vanguard, a shift that Swann said enables the company to embed intelligence into decision-making at all levels. “It is allowing us to bring insights to all of our executives about what’s happening with our clients. “It allows us to inform and influence the strategy and what we should focus on next. It also allows us to accelerate our business strategy. We’re a digital organization; we don’t have bricks and mortar, so the majority of our clients interact with us through data, through transactions, through clicking around on our website, maybe calling our call center, which is just another type of unstructured data. That data allows us to understand what our clients need and how we should respond,” the executive said. Swann’s office oversees the entire data lifecycle — from engineering and management to advanced analytics, machine learning and behavioral science. Central to Swann’s approach is a hub-and-spoke model that bridges the technical and business sides of the company. To foster cross-functional collaboration, Swann said sitting together and sharing the same objectives and key results (OKRs) matter. For example, Swann said an AI model could look through the data to coach the sales team into performing better. It could recommend the three best actions to take, the next three people to talk to, and the three topics they should talk to them about. By A/B testing these recommendations, the team tracked the performance improvements and worked with finance to quantify the gains. These and other initiatives have brought value to Vanguard. “Last year alone, we were up over $300 million of incremental value” across revenue generation, cost efficiencies, cost avoidance and risk reduction, he said. As for agentic AI, Vanguard is deploying agents that let users do things like pull data from databases in natural language and conduct data lineage checks, which avoids a costly undertaking. Swann said Vanguard has data processes going back decades, and the company must now create lineage to trace where the data comes from and where it goes.
Fed’s latest Diary of Consumer Payment Choice reveals consumers made an average of 11 payments per month with a mobile phone in 2024, up from four payments in 2018; cash remains a key backup payment method
Federal Reserve Financial Services today issued the 2025 Diary of Consumer Payment Choice (Diary), an annual survey measuring the evolving role of cash in the U.S. economy. Findings from this nationally representative survey showed that amid the increasing digitalization of payments, consumers continue to use cash and keep it handy. Cash ranked third as a top payment instrument among consumers, a position it has held for the past five years. In 2024, it accounted for 14% of consumer payments by number, while credit and debit cards accounted for 35% and 30% of payments, respectively. Overall, U.S. consumers made an average of 48 payments per month, continuing an upward trend that began in 2021. In 2024, this growth in the overall number of payments was driven by increased credit card usage, remote payments and payments made with mobile phones. The survey also revealed generational and demographic trends in payments. Households earning less than $25,000 per year and adults 55 and older relied more on cash than other cohorts. In contrast, adults aged 18 to 24 were more likely to pay with a mobile phone, using their phones for 45% of all payments. Other key findings included:
- S. consumers made an average of 11 payments per month with a mobile phone in 2024, up from four payments in 2018.
- Cash remains a key backup payment method for U.S. consumers. Of all cash payments in 2024, nearly two-thirds were made by consumers who prefer other payment methods, such as debit or credit cards.
- Nearly 80% of U.S. consumers have held cash in their pockets, purses or wallets for at least one day of the month for each Diary survey conducted since 2018. Though the value of these holdings has decreased since 2022, it remained elevated in 2024 compared to pre-pandemic levels.
- More than 90% of U.S. consumers intend to use cash as either a means of payment or store of value in the future.
MCP’s ability to let enterprises custom-configure servers for autonomous AI agents, its directionality to access information on client side and interoperability is spurring adoption in AI development workflows
Anthropic released Model Context Protocol (MCP) seems to have become the winning protocol choice for the AI industry. Despite the number of companies announcing MCP servers, MCP is technically not a standard. But many see MCP as one of the main protocols, if not the potential winner, for the agentic ecosystem. A lot of MCP’s attractiveness comes from streamlining how models interact with data and tools. Before MCP, developers pointed models and agents to data with APIs. However, APIs are imperfect connectors, especially for agents that access data to complete tasks automatically. Unlike APIs, organizations can configure their MCP servers with custom instructions laying out what agents can or cannot access. The server can “ask” an agent for its identity and determine if it can tap information on the MCP client side. Companies have more of a say on what outside agents can access on their end, giving MCP more directionality from the enterprise. Sagar Batchu, co-founder and CEO of API tooling company Speakeasy, said MCP transforms the work interface and API to a chat interface. He said MCP makes it so Speakeasy and its customers don’t need to rewrite or manually maintain APIs constantly. Yaniv Even Haim, chief technology officer of website builder Wix, told that MCP aligns with the company’s goals because it believes MCP can act as a “bridge” for its AI development workflows. “Wix chose the MCP model in particular because it aligns with the industry’s shift toward LLM-powered development, where context-rich, intelligent interfaces are key,” Haim said. For many companies, MCP will be one of many protocols they support as their customers decide which interoperability and agent communication methods to use. The growing adoption of MCP, for the varied reasons many companies have, proves that demand for standards is only growing.
Adoption of developer-focused AI tools surging at the expense of , freelance platforms such as Fiverr and Upwork; AI writing tools, crowdsourcing and search are fading fast
A new report released by the publicly traded market research and intelligence firm SimilarWeb—covering global web traffic patterns for AI-related platforms for 12 weeks through May 9, 2025—offers a helpful look for enterprises and interested users into the current landscape of generative AI usage online. Using proprietary analytics based on site visits, the report tracks trends across sectors including general-purpose AI tools, coding assistants, content generators, and more. Here are five key findings from the report: 1) Usage of developer AI and coding tools is rising fast: Developer-focused AI tools are surging in adoption, with traffic to the category up 75% over the past 12 weeks. That growth includes Lovable, which exploded with a jaw-dropping +17,600% spike, and Cursor, which grew steadily month over month. 2) We all know DeepSeek had a moment earlier this year — but so did Grok — and now both have fallen back into low plateaus: Grok traffic skyrocketed more than 1,000,000% in March—driven by its branding as an uncensored yet powerfully intelligent platform and Elon Musk association—before falling more than 5,200% by early May. DeepSeek saw a similar arc, peaking at +17,701% growth before crashing -41%. The takeaway: virality can’t replace retention, especially compared to AI leader OpenAI and also legacy tech brand Google. 3) AI writing tools are fading fast: Category traffic fell 11% overall, with platforms like Wordtune (-35%), Jasper (-19%), and Rytr (-23%) all trending downward. Only Originality.ai bucked the trend with steady traffic gains, likely due to its focus on AI detection rather than generation. This plateau suggests content saturation and possibly growing skepticism over quality or usefulness. 4) AI image generators and design tools show extreme volatility: Design-focused AI remains a mixed bag. While overall category usage dipped slightly (-6% over the 12-week window), some platforms made outsized gains. The erratic pattern may reflect a crowded landscape of tools that offer similar functionality but compete on novelty or aesthetics. 5) AI is eating up legacy tech such as crowdsourcing and search: Freelance platforms such as Fiverr (-17%) and Upwork (-19%) are losing traffic, possibly as users turn to AI tools for tasks like design, writing, and code. Search engines such as Yahoo (-12%) and Bing (-14%) continue a multi-quarter drop in visits, while consumer EdTech companies like Chegg (-62%) and CourseHero (-68%) are in free fall. The signs point to early-stage AI disruption beginning to erode the utility of some legacy platforms. It also offers a hint to enterprises that either leverage or create such services — the time may be coming to reduce dependency on them, either from a revenue generation, marketing, or overall business perspective.