Head to head: Meta: Muse Spark 1.3 vs Z.ai: GLM 5.3 Prime

Meta: Muse Spark 1.3: # Startup Spotlight: LedgerLark Wants to Clean Up Post-Trade Operations For many small investment firms, the trade itself is the easy part. What follows — matching confirmations, checking custodian records, updating internal ledgers — can absorb hours of staff time when something does not line up. LedgerLark is building AI agents to take on that reconciliation work, focused on smaller firms that cannot staff a large back office. ## An operations problem Co-founder Mina Batra spent seven years in operations at a regional brokerage before starting LedgerLark with co-founder Owen Carr, who previously built reconciliation systems at a large bank. Batra traces her interest in the problem to a specific incident: a six-person team spent three days resolving a trade mismatch caused by a copied account number. The experience raised a question that shapes the company — how much operations time goes to tracking down small data errors across systems that do not quite agree? ## What the product does LedgerLark’s agents compare three sources that often have to agree for a trade to settle cleanly: broker-dealer confirmations, custodian records, and a firm’s internal ledgers. When the sources differ, the system flags the exception and proposes a correction, such as updating a fee, date, or account field, for a person to review and approve. The agents do not execute trades or move money, a boundary the company presents as a control for firms wary of automation near client assets. ## Early customers and funding LedgerLark reports five paying customers, with annual contracts ranging from $36,000 to $110,000, according to the company. As an example of early use, the company says one customer reduced its unresolved daily exceptions from 70 to 18 after adopting the product. That figure has not been independently verified and reflects one firm’s experience over an unspecified period, so it is difficult to know how representative it may be. The company has raised a $3.5 million seed round led by Motive Row. For now, the team appears focused on learning from a small group of users rather than expanding quickly, which seems practical given the complexity of post-trade data and the variation between firms. ## Open questions and competition LedgerLark is entering an area with existing options. Competitors include established reconciliation vendors and software built internally by banks and larger firms, which means potential customers may already have a workflow, even if imperfect. For smaller firms, the alternative is often manual review in spreadsheets, which can be hard to displace because staff are used to it and wary of changing controls. Another question is sales timing. Financial institutions may require lengthy security and procurement reviews before adopting new tools, especially those that touch ledger data. That can stretch early deals and make revenue uneven, even if interest is present. With five customers, LedgerLark has limited evidence so far about how quickly it can move through those reviews or how its agents perform during higher volumes or volatile markets. The usefulness of human-in-the-loop design will also be worth watching. Proposing corrections for approval keeps people involved, but firms will need to see whether review workloads stay manageable and whether error explanations are clear enough for compliance teams. If LedgerLark can show consistent results across more firms — fewer lingering exceptions without adding review burden — it could find a steady role among smaller managers. The next useful evidence will be retention, broader performance data, and whether security reviews become routine rather than exceptional.

Z.ai: GLM 5.3 Prime: # Startup Spotlight: LedgerLark Points AI Agents at Post-Trade Drudgery *Small firms, big reconciliation problems, and agents that never touch the money.* The

Meta: Muse Spark 1.3: # QueueZero Launches Proxy to Turn Meeting Talk Into Agent Action SAN FRANCISCO — QueueZero on Wednesday launched Proxy, an AI meeting assistant designed to attend virtual meetings, generate notes and assign follow-up work to software agents. The launch, held at the company's developer event in San Francisco, focused on reducing post-meeting administrative work. QueueZero said Proxy listens to meeting conversation, summarizes decisions and creates structured tasks that can be routed to humans or to connected third-party agents for execution. A live demonstration briefly shifted attention from the planned script. During a question-and-answer session, a customer in the audience said, “Someone should tell Evan to stop talking and send us the pricing sheet.” Proxy, which was transcribing the session, interpreted the remark as an instruction. It created two tasks assigned to QueueZero CEO Evan Wu: “Stop talking” and “Send pricing sheet.” Wu laughed, completed the second task and deleted the first. He then sent the pricing information to attendees and continued the presentation. The audience applauded, and the company moved on to a technical overview of the product's integrations. QueueZero executives said the incident illustrated both the capabilities and the controls of the system. According to the company, Proxy requires confirmation before an external message is sent. Internal tasks, by contrast, can be created automatically without additional approval. Administrators can disable automatic task creation, the company said. That setting requires managers or IT staff to review proposed tasks before they are assigned or acted upon. Proxy costs $30 per user per month. It enters public beta next week, QueueZero said. The pricing places Proxy in line with other paid enterprise productivity tools that combine transcription with workflow automation. The company said early enterprise testers have used Proxy to draft project updates, file support tickets and trigger code review workflows. QueueZero plans to add broader integration with customer relationship management and project tracking tools during the beta period. Analysts said meeting assistants are becoming a competitive segment as vendors seek to connect transcription with workplace automation. QueueZero is positioning Proxy as a bridge between conversation and execution, rather than as a passive note-taking tool. The company did not disclose customer numbers for the private test. It said general availability will follow the public beta, pending feedback on accuracy and permission controls.

Z.ai: GLM 5.3 Prime: # QueueZero's AI Meeting Assistant Assigns Its CEO Two Tasks During Launch Demo *The new assistant, Proxy, took an audience remark at face value — and put the company's chief executive on the task list.* QueueZero launched an AI meeting assistant called Proxy on Wednesday. The product is designed to attend meetings, produce notes, and assign follow-up work to software agents — and by the end of the launch presentation, it had demonstrated that final capability on an unintended target: QueueZero's chief executive. The moment came during the public demonstration, when a customer in the audience remarked, "Someone should tell Evan to stop talking and send us the pricing sheet." Proxy interpreted the remark as an instruction. It created two tasks, both assigned to CEO Evan Wu: "Stop talking" and "Send pricing sheet." Wu laughed, completed the second task, and deleted the first. The software, in its defense, had followed the instruction to the letter. QueueZero says Proxy requires confirmation before an external message is sent, meaning the assistant cannot dispatch messages to customers or other outside parties without a human approving them first. Internal tasks — like the two assigned to Wu — are governed by looser rules: they can be created automatically, with no approval step. Administrators can disable automatic task creation for organizations that prefer human review of every task, though the default, as Wednesday's demonstration showed, errs on the side of action. The episode was an unplanned illustration of both Proxy's promise and its potential pitfalls. Meeting assistants are designed to listen for actionable statements — deadlines, assignments, commitments — and act on them. Meetings, however, are also full of jokes, asides, and remarks addressed to no one in particular, the kind of speech humans filter out without thinking. On Wednesday, the result was a pricing sheet and a good-natured laugh. Aimed at different targets, the same literal-mindedness could create real problems, which is why the confirmation step for external messages and the administrator controls exist. In normal use, Proxy attends a meeting, produces notes, and — the feature at the center of Wednesday's moment — assigns follow-up work to software agents, converting conversation into delegated action items. The demonstration suggested those assignments are not limited to software: humans, including the company's CEO, can land on the receiving end as well. Proxy costs $30 per user per month and enters public beta next week, when a wider set of customers will find out what its task engine makes of their meetings.

Meta: Muse Spark 1.3: # Duplicate Reimbursements Cost KiteBridge Customers $2.8 Million After Failed Update Ninety-one companies were left out of pocket Tuesday after expense-management software provider KiteBridge sent employees thousands of duplicate reimbursements worth a combined $2.8 million, an error the company took hours to publicly acknowledge. The erroneous payouts, which totaled 13,400 separate transactions, were funded by employers, not taken from employees. Workers who were owed a reimbursement simply received it again, while their companies absorbed the extra cost. KiteBridge, whose platform handles expense reports and automated payouts for corporate clients, said the problem began Tuesday morning with the release of version 4.2 of its software at 8:45 a.m. That update coincided with a timeout from the company's payment processor. When a payment processor times out, it means KiteBridge sent instructions to move money but did not receive immediate confirmation that the transfer succeeded. To avoid leaving employees unpaid, KiteBridge's software includes a fallback system that retries the payment. In this case, the fallback backfired. Instead of recognizing the retries as follow-ups to a payment already in progress, the system treated each retry as a brand-new reimbursement request. The result was a cascade of duplicates issued through its automated payments system for about an hour. The failure is striking because KiteBridge had previously said it was protected against exactly this type of error. According to its April changelog for version 4.0, the company had introduced "exactly-once payment protection," a safeguard designed to assign each payment a unique identifier so that even if a request is sent multiple times, money moves only once. For general users, it works like a stamp that tells the system: this bill has already been paid, do not pay it again. That protection did not function during Tuesday's incident. The first customer complaint arrived less than half an hour after the deployment, at 9:12 a.m. KiteBridge suspended all automated payments at 10:03 a.m., halting the spread after roughly 75 minutes of faulty operation. Its public communication lagged behind its internal response. The company's incident page initially described the issue only as delayed reimbursements, suggesting employees would be paid late rather than paid extra. It was six hours after automated payments were suspended before the page was updated to disclose that duplicate payments had been sent. By Wednesday evening, the company said $2.3 million of the $2.8 million had been recovered, with the remainder still pending. It said no money had been withdrawn from employees' accounts to cover the error. Chief Executive Lena Ortiz apologized for the incident and said KiteBridge would publish a full technical report explaining why the safeguard failed and what changes would prevent a recurrence. For corporate finance teams, the episode underscores the high stakes of automation in payroll and expenses, where a brief software fault can translate within minutes into millions of dollars in errant transfers that take days to unwind.

Z.ai: GLM 5.3 Prime: # KiteBridge Software Update Triggers $2.8 Million in Duplicate Reimbursements Across 91 Companies *A routine release went wrong in an expensive way: thousands of duplicate payments, funded by employers, and a recovery still incomplete a day later.* A software update released Tuesday by KiteBridge, a maker of expense-management software that companies use to reimburse employees for travel, meals and other costs, generated 13,400 duplicate payments totaling $2.8 million. The failure affected 91 corporate customers, and it was the employers — not the employees — who covered the cost. The failure stemmed from a payment-processor timeout that KiteBridge's fallback system mishandled: when a payment appeared to stall, the backup mechanism resubmitted it, treating each retry as a brand-new reimbursement rather than a repeat of the original. Automated payments were suspended within hours, and by Wednesday evening $2.3 million of the duplicated payments had been recovered, with roughly $500,000 still pending. ## How the failure unfolded KiteBridge deployed version 4.2 of its software at 8:45 a.m. Tuesday. The first customer complaint arrived 27 minutes later, at 9:12 a.m., and by 10:03 a.m. the company had suspended its automated payment system — roughly an hour and 18 minutes after the update went live. The 13,400 duplicate payments had been generated within that window. The mechanics of the failure are a well-known hazard in payment engineering. When expense software submits a reimbursement, it passes the request to a payment processor and waits for confirmation that the money has moved. On Tuesday, that confirmation did not arrive within the expected time limit — what engineers call a timeout — and KiteBridge's fallback system stepped in to try again. The problem was that it resubmitted the payment without recognizing it as a retry. Each attempt was recorded as a fresh reimbursement request, so employees were paid again for expenses that had already been covered. ## A safeguard that wasn't That kind of double-payment bug is precisely what payment systems are designed to avoid. Robust designs tag every transaction with a unique identifier so that a command submitted twice still executes only once — a guarantee engineers call "exactly-once" processing. That term has particular significance at KiteBridge. An April changelog for the software stated that version 4.0 had introduced "exactly-once payment protection," a feature whose name describes exactly the kind of failure that occurred on Tuesday. The company has not yet explained how the glitch related to that protection — whether the mechanism was bypassed, disabled or simply failed — and the answer is likely to feature prominently in the technical report the company has promised. ## Disclosure and recovery KiteBridge's public incident page initially characterized the problem as delayed reimbursements. Only six hours after the company suspended automated payments — around 4 p.m. Tuesday — did the page disclose that duplicate payments had been made. Customers relying on the page in the interim would have understood the incident as a delay, not a duplication. For employees, the failure carried no direct financial harm: no workers had money withdrawn from their accounts, and the duplicate payments were funded by their employers. The affected companies bore the immediate cost, and the clawback remains incomplete. Chief executive Lena Ortiz apologized for the incident and said KiteBridge would publish a technical report. As of Wednesday evening, $2.3 million of the $2.8 million had been recovered; the rest remained pending.

Meta: Muse Spark 1.3: # Arcwell Launches Relay, an AI Assistant for Support Teams Arcwell has launched Relay, an artificial intelligence platform that examines customer support conversations and suggests answers to agents. At its core, Relay reads an unfolding exchange and surfaces a draft reply drawn from what the company already knows. In addition to suggested answers, the platform identifies customers who may cancel and creates summaries for agents to keep with the case. The pitch is straightforward: cut down the time agents spend hunting through old tickets and help articles. “Support agents spend too much time searching for information that their companies already have,” CEO Priya Shah said. Relay was tested by 12 companies over six months. Arcwell says agents using it closed tickets 14% faster. The company did not release customer-retention results, so it remains unclear whether flagging at-risk customers translated into fewer cancellations. Relay costs $65 per agent each month, positioning it as a per-seat addition to the existing support stack rather than a wholesale replacement. That positioning reflects how most support teams are encountering AI — not as a single sweeping overhaul, but as a series of smaller interventions in daily workflow. The value, when it lands, tends to show up in minutes saved per ticket and cleaner handoffs between shifts, rather than dramatic transformation. Access is the trade-off. To work, Relay requires access to support transcripts and internal help documents. Arcwell says customer data is encrypted and is not used to train shared models. Companies can choose a 30-day or 90-day retention period, a control that will likely draw close attention from privacy and security reviewers. For Arcwell, the question now is adoption. Faster resolution is an appealing promise for teams under pressure to do more without adding headcount, but leaders will weigh it against data-governance concerns and the still-unproven impact on retention. Relay will succeed if agents trust its suggestions enough to use them — and if customers never notice the help behind the reply.

Z.ai: GLM 5.3 Prime: #

Meta: Muse Spark 1.3: # The Last Cohort Leaves Foundry Hall MILWAUKEE — Foundry Hall will close Oct. 31, bringing an end to seven years as a nonprofit startup accelerator for early-stage companies in the city. Opened in 2018 in a converted warehouse on the edge of the Third Ward, the organization supported 212 startups that now collectively employ 1,140 people, according to its final impact report. Six staff members will lose their jobs with the closure, and 12 startups currently working out of its shared space will have to find new homes by the end of the month. Executive director Camille Ross said the board voted last week after months of trying to replace its largest source of funding. Blank River Bank, which supplied 38% of Foundry Hall's annual budget, declined to renew its sponsorship for next year. The bank said the move reflected a shift in its community investment focus, not dissatisfaction with the accelerator's work. Ross said the organization has enough cash to operate through October but not to responsibly commit to another full program cycle. Fundraising conversations with other donors stalled over the summer as costs rose. “We kept waiting for the replacement check that never came.” For founders who came up through the building, the loss is measured less in dollars than in proximity. Dev Malik, whose logistics software company was part of the 2019 cohort, recalled arriving for his first investor meeting after months of pitching from the back of a coffee shop. That sense of having a place to be taken seriously, he said, was what Foundry Hall offered to teams with little more than an idea. City officials said they are trying to soften the landing for the dozen teams being displaced. Deputy Mayor Susan Cole's office is in discussions with nearby building owners and universities about temporary workspace. Still, Ross acknowledged there is no plan for a direct successor. Unlike a startup, she said, an accelerator cannot simply pause and restart when money returns; mentors scatter, leases end, momentum fades. In its final weeks, Foundry Hall plans to keep the workshop calendar full and host an open house for alumni. The mood inside, staff members said, is subdued rather than bitter — pride in what was built, tempered by exhaustion. After Oct. 31, the sign will come down. What remains, Ross hopes, is less visible but harder to erase: more than a thousand jobs that trace some part of their origin to a crowded room where founders learned how to ask for money, and sometimes got it.

Z.ai: GLM 5.3 Prime: #

Meta: Muse Spark 1.3: C

Z.ai: GLM 5.3 Prime: **C** It accurately captures the core news event (the product pull tied to the data exposure), includes the specific, verifiable details ($12/month AI assistant, 4,860 accounts, workspace titles, 11 hours), and doesn't overstate the facts. - **A** is wrong: message contents were *not* sent, and "massive privacy disaster" overstates the known facts. - **B** buries the news entirely — it omits the data exposure, the reason the product was pulled, and uses promotional filler. - **D** is wrong: only titles were exposed, not "everything they had written," and there's no evidence of hackers or malicious access.

Meta: Muse Spark 1.3: # Cloudnote Left Customer Prompt Excerpts Exposed for Four Hours *Company says debugging page showing snippets from 1,842 sessions was taken down within minutes, but investigation into access is ongoing.* Cloudnote left a debugging page accessible without authentication for four hours on Monday, exposing short excerpts of customer prompts before taking it down 19 minutes after being alerted by a researcher, the company said. Logs reviewed by the company show the page contained up to 200 characters from prompts in 1,842 user sessions, a small fraction of Cloudnote's 3.2 million registered users. The company said full documents, passwords and payment information are stored in separate systems and were not included on the exposed page. The exposure matters because prompts can include sensitive personal or business information, and even a 200-character snippet can reveal substantial context depending on its contents. The incident also highlights the risk of internal debugging tools being left accessible on the public internet. What remains unclear is whether anyone other than the researcher viewed or copied the data. Cloudnote says it has found no evidence of additional access so far, but emphasized its investigation is not complete and it could not yet rule it out. The company did not say how the page became public or whether affected users will be notified individually.

Z.ai: GLM 5.3 Prime: # Cloudnote Says Unprotected Debug Page Exposed Customer Prompt Excerpts for Four Hours *Excerpts from 1,842 user sessions were reachable without a login for four hours on Monday. Cloudnote says it has found no evidence that anyone besides the researcher who flagged the page accessed it — but its investigation is not yet complete.* An internal Cloudnote debugging page that displayed excerpts of customer prompts was accessible online without authentication for four hours on Monday, the company said. Cloudnote, which has 3.2 million registered users, removed the page 19 minutes after a security researcher reported the exposure. Logs show the page contained excerpts from 1,842 user sessions, with each excerpt spanning up to 200 characters of a prompt. Full documents, passwords and payment information were stored in separate systems and did not appear on the page, the company said. Cloudnote said it has found no evidence that anyone other than the researcher accessed the page — a finding that reflects what the company knows so far, not confirmation that no one else viewed the excerpts. The investigation is not complete, and Cloudnote has not said what the affected prompts contained or whether its review could expand the tally of exposed sessions. The logged sessions represent a small fraction of Cloudnote's 3.2 million registered accounts, and the most sensitive categories of customer data were not involved. Still, the episode shows how internal debugging tools left unprotected, even briefly, can expose user content — and it leaves questions that Cloudnote's unfinished investigation will have to answer.

Meta: Muse Spark 1.3: # HelioDock Launches Night Shift for Remote Monitoring of Warehouse Robot Fleets *Company points to 17% drop in unplanned stoppages during 14-month beta as it raises $24 million and sets September release* On Thursday, HelioDock launched Night Shift, software that lets operators monitor fleets of its autonomous warehouse robots from a remote location. The release establishes off-site supervision as a distinct product for the warehouse-robot maker. Paired with a $24 million Series B financing led by North Coast Ventures, which brings HelioDock's total funding to $39 million, the debut links an expansion of its software line to new capital. Night Shift is scheduled to become generally available September 15 and will cost $90 per robot per month. Night Shift was evaluated for 14 months in a private beta before Thursday's launch. Nine logistics companies participated, according to the company, operating a combined 310 robots while using the software in warehouse settings. HelioDock says those beta customers experienced 17% fewer unplanned robot stoppages. The figure has not been independently verified. The company attributed the change to use of Night Shift during the beta period. "A warehouse manager should not have to drive across town at two in the morning because one robot is confused by a pallet," CEO Mara Velez said. The statement frames remote monitoring as a response to off-hours interruptions. Rather than traveling to a facility to inspect a stopped machine in person, an operator using Night Shift could supervise the fleet from elsewhere. That capability depends on a sustained connection. Beta customer Aaron Pike of Fulcrum Logistics said the system worked well in facilities with strong connectivity but became less useful during rural LTE outages. When connectivity fails, robots return to their locally configured behavior, according to HelioDock. Remote control becomes unavailable in those circumstances, leaving each robot to operate under its on-site configuration until oversight from afar can be re-established. The distinction matters for where remote supervision is most applicable. In warehouses with reliable networks, operators retain the ability to watch fleets without being present. In locations subject to LTE interruptions, particularly rural sites described by Pike, that remote visibility narrows and intervention must wait on local behavior. Pike did not describe robot failures during those outages, only a reduced usefulness for the remote system itself. His account suggests performance of Night Shift tracked closely with network conditions across Fulcrum Logistics facilities. The per-robot pricing ties the cost of the software directly to fleet size. At $90 per robot per month, an operator with dozens or hundreds of machines would pay on a recurring basis for continued remote access after general availability begins. HelioDock is entering an area where fleet-management alternatives already exist. Competing fleet-management products are available from Relay Robotics and YardOS. The Series B financing adds $24 million led by North Coast Ventures. With total funding now at $39 million, according to the company, the round provides capital as HelioDock moves Night Shift from private testing with nine companies and 310 robots to broader sale next month.

Z.ai: GLM 5.3 Prime: #

Meta: Muse Spark 1.3: 6

Z.ai: GLM 5.3 Prime: **6** Every line in the draft violates at least one rule: 1. `#Flood software moves upstream` — missing space after `#` (headline format) 2. `The company is expanding its pilot.` — dek is not italicized 3. `##Details` — missing space after `##` (section heading format) 4. `Read [the announcement](www.example.com/news).` — link is not an absolute HTTPS URL 5. `* The first pilot begins Monday.` — list item begins with `*` instead of a hyphen 6. `<div>The second pilot begins Friday.</div>` — raw HTML

Meta: Muse Spark 1.3: # The Price of Good Enough Is Falling If you run customer support, you know the pitch: AI drafts answers so good your agents just approve and move on. A new benchmark of five language models on real support work suggests that future is getting cheaper. The top-scoring model didn't win outright. A rival tied it on quality while costing less than half as much and answering faster. That tie matters more than any leaderboard. But like any lab test, it needs context. ## How it was tested Researchers took 200 anonymized tickets from one software company and had each model — North, Vale, Ember, Kite and Harbor — draft one answer per ticket. Three experienced support managers graded each answer blind, without knowing which model wrote it. Scores combined factual correctness, usefulness and tone. Costs were estimated per 1,000 tickets, the number that decides whether a pilot can scale. Blind grading by experts is a solid step up from demos. It reduces brand bias, and the three-part score reflects real work, where a correct but cold or vague answer still fails. Still, 200 tickets is a snapshot. Even expert judgment involves subjectivity, and the narrow setup limits how far results travel. ## Tied on quality, not on cost Two models separated from the rest, both averaging in the mid-80s. North was numerically first, Vale a fraction behind. Ignore that fraction. The difference was not statistically significant, meaning it could easily be noise from this sample of tickets and graders. For quality, treat them as tied. For operations, treat them as different. North cost an estimated $1.90 per 1,000 tickets; Vale cost $0.74, a saving of more than 60%. Vale was also 18% faster. That is the clearest finding: at this level, you did not have to pay more for better drafts. At a million tickets, the modeled gap exceeds $1,000, before counting the benefit of faster replies for wait times and agent throughput. Workload fit adds nuance. North was strongest on billing disputes, where policy precision and empathy count. Vale led on technical troubleshooting, where clear step-by-step reasoning counts. The best choice depends on what fills your queue. ## Going cheaper costs quality Further savings came with a real drop. The other three models averaged in the 70s, even as estimated costs fell to $0.31 per 1,000 and below, to just $0.08 at the cheapest end. So there is a tradeoff curve, just not at the top. Once you leave the leaders, you pay less and get noticeably weaker drafts. Small gaps among those lower models should not drive decisions from this sample alone. One failure hit everyone: tickets that depended on an image attachment, like a screenshot of an error. Text-only models all performed poorly there. If your customers often say “see attached,” these scores overstate what you will see. ## What this doesn’t prove Everything here is English-language tickets from a single software company. Products, jargon, customers and languages will shift results. Crucially, researchers tested only draft replies. They did not test agentic actions, tool use or multilingual support — no looking up orders, issuing refunds, or working across languages. In production, a slightly weaker writer that can safely use tools may beat a stronger writer that cannot. Tone expectations also vary by industry. Use this as a first filter, not a buying decision. Shortlist the efficient frontrunner — here, Vale for most teams — then pilot on your own tickets, especially hard cases with images, other languages and tool-dependent steps. Track resolution, escalation and satisfaction, not just grader scores. For buyers, the question shifts from which model writes best in a lab to which efficient model holds up on your tickets, at your scale, with your guardrails. Top-tier draft quality no longer demanded top-tier price. That is worth testing in your own queue.

Z.ai: GLM 5.3 Prime: #

Meta: Muse Spark 1.3: # Raincheck Raises $8 Million to Help Cities Get Ahead of Floods *The Austin startup maps drainage systems and pairs them with rainfall forecasts to guide crews before storms hit* Raincheck, a maker of flood-risk software for local governments, announced today that it has raised $8 million in seed funding. The Austin company maps drainage systems and combines those maps with rainfall forecasts. Cities use the software to decide where to send maintenance crews before a storm. Details on its approach are available in [Raincheck's forecasting methodology](https://example.com/raincheck-methodology). The seed round was led by Basin Ventures, with participation from existing investor Long Mile. The round brings Raincheck's total raised to $10.5 million. CEO Ana Flores said: > “Our customers usually know which streets flood. The harder problem is deciding which blocked drain, broken pump or unfinished repair should receive attention before the rain arrives.” Raincheck currently works with seven cities in Texas and Louisiana. The company declined to identify them because some contracts have not been publicly announced. Its forecasts are recommendations. Emergency managers retain control over deployments and public warnings. Other flood-planning vendors include HydroMap and municipal engineering contractors. Raincheck says the new funding will help it hire eight engineers and expand into Florida.

Z.ai: GLM 5.3 Prime: # Raincheck raises $8 million to help cities get ahead of flooding *The Austin company's software combines drainage maps with rainfall forecasts to show local governments where to send maintenance crews before storms arrive.* Raincheck, an Austin company that makes flood-risk software for local governments, announced today that it has raised $8 million in seed funding. The company maps drainage systems and combines those maps with rainfall forecasts, an approach detailed in [its methodology](https://example.com/raincheck-methodology). Cities use the software to decide where to send maintenance crews before a storm. Basin Ventures led the round, with existing investor Long Mile also participating. Raincheck has raised $10.5 million in total. CEO Ana Flores said in a statement: > Our customers usually know which streets flood. The harder problem is deciding which blocked drain, broken pump or unfinished repair should receive attention before the rain arrives. Raincheck currently works with seven cities in Texas and Louisiana. The company declined to identify them because some contracts have not been publicly announced. Its forecasts are recommendations; emergency managers retain control over deployments and public warnings. Other flood-planning vendors include HydroMap and municipal engineering contractors. Raincheck says the new funding will help it hire eight engineers and expand into Florida.

Meta: Muse Spark 1.3: # Raincheck raises $8 million for flood-risk software *The Austin startup helps cities decide which drainage problems to address before storms.* Raincheck has raised an $8 million seed round led by Basin Ventures. ## Expansion plans The company plans to hire eight engineers.

Z.ai: GLM 5.3 Prime: # Raincheck raises $8 million for flood-risk software *The Austin startup helps cities decide which drainage problems to address before storms.* Raincheck has raised an $8 million seed round led by Basin Ventures. ## Expansion plans The company plans to hire eight engineers.