The Second User Changes the Product

BRIJ Medical

A surgeon can see a device work in minutes. Getting it into the hospital can take considerably longer, a timetable Tim Gleeson has had to learn while building BRIJ Medical's commercial reach.

In a direct interview with Medsider founder Scott Nelson, the CEO described expecting a six-month sales cycle and encountering something closer to 13 to 15 months. That distance between a convincing demonstration and a purchasing decision gives BRIJ's new $5M financing a useful frame.

Announced November 3, the round was led by BRIJ Medical's board, with most participation from existing investors. The company started with a $3M target and increased it as demand grew. John Grotting, board chair, described a larger commitment to the company's next phase of growth.

The people already inside the business are financing a wider commercial assignment. BRIJ-SEAL launched in July, and BRIJ-Ag, a silver-based antimicrobial dressing, is planned for later this quarter. The proceeds support commercial expansion and upcoming launches. BRIJ lists offices in Marietta, Georgia, and Solana Beach, California.

Founder and CMO Felmont (Monte) Eaves III, MD, came at the problem through plastic surgery. Emory's account describes roughly 40 prototypes before the design for Brijjit took shape, with manufacturers involved along the way. A device that looks simple still has to survive the work of making it at scale.

Brijjit offloads tension around an incision. BRIJ-SEAL covers and protects it, and the products can be used together. BRIJ describes its 14-day bacterial barrier through in-vitro testing, a specific evidence claim that should stay attached to its testing context.

For a surgeon, those are complementary jobs during recovery. For Gleeson's team, they create more to explain, demonstrate and support across plastic, orthopedic and spine procedures. A broader portfolio expands the conversation with a practice while increasing the commercial work behind it.

BRIJ has also been building the purchasing route. Its July announcement brought BRIJ-SEAL into Premier's pre-negotiated pricing and terms, available at members' discretion. Access gets a product into consideration; a hospital still chooses whether and how to use it.

That makes the board's larger commitment especially practical. Gleeson is taking an expanded incision-management portfolio through the conversations, evaluations and purchasing decisions that sit between a surgeon's interest and the next order.

Deeper Analysis 👇 https://devcuration.com/articles/brij-medical-raises-5m-incision-management

Lazarus AI

Alex Panait is selling enterprises the right to change their minds about AI. The job for Lazarus AI is to make that freedom useful after the purchase, when somebody still has to support the system.

The Boston company announced $43M in funding on November 3, with QBE, Mucker Capital, Motivate Ventures, State Farm Ventures, Wave Function Ventures, Indelible Solutions, Joyance Partners, AllegisNL Capital, One Way Ventures, AngelList, Techstars and Plug and Play among the participants. A lead investor and round stage were not disclosed.

Lazarus traces its founding to Ariel Elizarov and Anish Patel. Alex Panait now leads the company as CEO, taking a business rooted in difficult document work into a wider assignment across enterprise intelligence, cybersecurity and legacy modernization.

Those jobs put different people on the receiving end of an AI result. An insurance team needs to find the page behind a finding. A security engineer needs enough evidence to reproduce a flaw. A modernization team needs to recover the business rules inside old code before deciding what can change.

AIS, Clearwing and Constructor address those respective tasks. Lazarus says customers can run its platform on premises, in sovereign environments or in their own cloud, choose models and approve model changes. Clearwing supplies findings and proposed fixes while engineers retain the decision about applying them.

There is a commercial choice inside that architecture. A buyer can pilot a workload with Lazarus, co-build, operate the system with support, or work through a partner. Giving customers those routes makes the handoff part of the product: who keeps it running, who checks the evidence and who helps when the next model arrives.

The absence of a per-token charge changes the conversation about volume. Hardware and operating costs still belong in the calculation, and the company's reported 78% cost reduction comes from a modeled high-volume scenario. A customer's workload has to earn its own economics.

QBE's relationship already reaches into underwriting and claims. AllegisNL's earlier investment connected Lazarus with a fund supported by Nippon Life Group, giving the partnership strategy a history beyond this announcement.

The money will support partnerships, R&D and new products. As Lazarus gives more institutions the option to run AI themselves, its own team inherits the continuing work of making that option practical for the people who have to operate it.

Deeper Analysis 👇 https://devcuration.com/articles/lazarus-ai-raises-43m-for-independent-enterprise-ai

Fractile

Fractile's careers page has room for a compiler engineer in London or Bristol and a foundry engineer in Taipei. That is a revealing distance for an AI company to cover before a customer gets a faster answer.

Walter Goodwin, founder and CEO, is building chips, software and systems for AI inference. The assignment spans the machinery that runs a model and the software that makes that machinery useful, with long reasoning tasks putting pressure on both.

Its financing record has grown substantially. Nine Companies House returns filed on October 15 and 28 show Series C preferred-share allotments with about £577.4M in recorded consideration, calculated from the individual share counts and amounts paid. The allotments run from September 9 through October 23.

That figure describes the share issuances. The returns alone do not reconcile how much fresh cash arrived during this financing, and the primary sources reviewed leave the Series C investors, valuation and use of proceeds undisclosed. Three preferred classes carry different prices, so flattening the record into one uncomplicated check would lose useful information.

Fractile had already announced $220M in May. Accel says it co-led that Series B; Founders Fund, Factorial Funds and other backers also appear in the company's investment history. Those relationships belong to their documented rounds, rather than an invented Series C syndicate.

Goodwin's argument for specialist hardware starts with time. A model working through a difficult problem can generate a long sequence of intermediate output. Faster inference could let the user attempt more of that work within a useful deadline, instead of simply receiving the same short answer sooner.

Fractile is pursuing that ambition across the stack. Its open roles cover silicon verification, compilers, device software, manufacturing quality and yield engineering. Each discipline has a say in whether an architectural idea becomes a system somebody can actually use.

Software compatibility and repeatable manufacturing travel into the purchasing decision alongside speed, because the buyer has to keep the whole system running.

For an infrastructure buyer, those connections matter. A faster chip still needs software that can carry the workload and production processes that can deliver it reliably. The compiler team in Britain and the foundry work in Taipei sit inside the same commercial promise, even when they appear in different departments on the hiring page.

Deeper Analysis 👇 https://devcuration.com/articles/fractile-series-c-share-allotments-577m

Pliant

The name on the Pliant account still matters when an AI agent does the work. The agent inherits that person's permissions, and the person stays accountable.

That is a consequential detail inside a payment company's AI pitch. A finance team can ask an assistant to investigate spending or help organize accounting data, while the authority attached to the work stays inside the company's existing roles.

Malte Rau, co-founder and CEO, and Fabian Terner, co-founder and CPO, built Pliant in Berlin in 2020. With CTO Alex Korotkikh, the team is opening a platform spanning commercial cards, payment accounts, FX and spend management to another kind of user.

Pliant announced an $80M Series C on November 4, co-led by Seaya Nova and Speedinvest, with Illuminate Financial, Motive Partners and Molten Ventures participating. The money supports US and Canadian expansion, preparation for further markets and the company's AI roadmap. Its $40M Series B in April 2025 had already financed the US push; Coastal became its sponsor bank this July.

There is a useful boundary in the documentation. The standard Model Context Protocol connector can help search transactions, change accounting data and manage card controls. It does not move money. Payment execution belongs to the next stretch of the roadmap, with permissions and confirmations already part of the design.

Pliant has an existing business to carry into that work. It reported passing $100M in annual recurring revenue in July, and its latest announcement puts the team above 300 people across 8 offices. Those are company-reported measures, with ARR describing a run rate rather than audited annual revenue.

Its customer can also be somebody else's customer. Banks, fintechs and software platforms embed Pliant's infrastructure into their own products. Giving those partners an agent interface means extending a financial relationship that already has a brand, a credit decision and an accounting trail attached.

The travel business offers a concrete example. A new S4BT/Trevium hotel wallet uses Pliant virtual cards so the hotel and traveler draw against the same authorized balance. Convenience reaches the traveler's phone while finance keeps the reconciliation process it already uses.

That is the work traveling with this round into North America and the AI roadmap. When an assistant prepares the next finance action, the person approving it will still need to recognize the account, the permission and the records that follow.

Deeper Analysis 👇 https://devcuration.com/articles/pliant-raises-80m-series-c-for-ai-payment-infrastructure

Path Fertility

A man can finish a fertility workup with an unexplained result. The couple still has to decide what to do next, with the calendar and the treatment bill continuing to move.

Andy Olson and Kristin Brogaard, PhD, built Path Fertility around the information missing from that conversation. The Salt Lake City company's SpermQT examines sperm DNA methylation, looking beyond the visible characteristics measured in a conventional semen analysis.

Now NICHD, part of the NIH, is funding an extension into some of the hardest cases. Path Fertility, the commercial brand of Inherent Biosciences, announced a 2-year, $2M Direct to Phase II SBIR grant on October 26 to develop and clinically validate an epigenetic diagnostic for non-obstructive azoospermia, or NOA.

Brogaard, the co-founder and CSO, is the award's principal investigator. Olson, the co-founder and CEO, has a company to build around a delicate proposition: better information can matter enormously before a couple commits to another procedure.

The funded work centers on epigenetic changes in MSH5, a DNA-repair gene. Path plans a multisite study, a semen-based reflex test and laboratory validation, with the assay intended for its NOA Guide platform. The award also includes research in 500 colorectal-cancer patients. That cancer-risk work is an investigation, with its clinical meaning still to be established.

SpermQT already gives the team a commercial starting point. A January 2026 peer-reviewed study analyzed 537 couples across 10 US clinics. Among 202 couples with IUI outcomes, pregnancy rates differed by SpermQT category. Those retrospective findings support further investigation; they cannot prescribe an individual couple's treatment or be transferred wholesale to men with NOA.

For a diagnostic company, the ordering physician is part of the product's journey. A result has to arrive in time, make biological sense and help a clinician explain the choices ahead. Laboratory sophistication earns its place in that conversation through evidence people can actually use.

Path already offers physician-ordered testing with at-home or clinic collection. Bringing a new assay into that workflow means the science also has a practical route to the people ordering it.

The SBIR award pays for research without selling an equity stake. Its value to Path lies in the specific work it funds: carrying an epigenetic signal through more clinical sites and toward a report a reproductive urologist can discuss with a patient facing surgery.

Deeper Analysis 👇 https://devcuration.com/articles/path-fertility-2m-nih-sbir-grant

Shackleton AI

The software that helped Ernest run its own construction jobs is now being sold to other builders. Michael Ragheb and the Shackleton AI team are taking an operating advantage they developed inside a contractor and making it somebody else's infrastructure.

That changes the assignment. A tool built around your own people can rely on habits nobody bothered to document. Sell it to another general contractor and those habits have to become a product that works through somebody else's bids, budgets, subcontractors and approvals.

Shackleton AI announced a $5M Seed round on November 4, led by Battery Ventures. Brick & Mortar Ventures, Trust Ventures, Correlation Ventures, Motivate Partners, Thayer Investment Partners, Act One Ventures, Brainstorm Ventures and Silence VC participated. The money will expand the workflows its AI agents handle and accelerate go-to-market.

CEO Michael Ragheb, CPO Hassan Yasin and CFO Olexandr Kaltsev bring the contractor's view into that work. The founding team's account of growing Ernest to $35M in revenue includes an uncomfortable discovery: buying more specialized software kept giving project managers more information to move by hand.

Shackleton puts the project record together before asking an agent to act on it. Bids, budgets, requests for information, submittals, change orders and field reports sit in a shared model. Agents draft RFIs, follow up on overdue submittals and flag budget variances before a payment request reaches approval. People keep the review and decision work.

There is a useful commercial detail in the integration list. QuickBooks and Sage can stay. A construction company can change how its jobs are coordinated while retaining the accounting system its finance team already knows. The rollout still has to bring field updates and office records into agreement, a demanding introduction for any new software supplier.

Shackleton says it manages more than $1B in project value across more than 15 states, serving midsize contractors and brands including Red Lobster, Dog Haus and Pepper Lunch. Those brands have another reason to care about coordination: each location brings another project whose progress has to remain intelligible to the owner.

The team learned the problem while it could walk onto its own jobs. With this round, more of its learning will come from a builder it doesn't employ, a ledger it didn't choose and an owner waiting for an answer about tomorrow's opening.

Deeper Analysis 👇 https://devcuration.com/articles/shackleton-ai-raises-5m-in-seed-funding

fal

“Can we try another version?” is a creative decision with a compute bill attached. fal has built a business around making that request easier to say, then carrying the infrastructure work each new attempt creates.

Co-founders Burkay Gur, CEO, and Gorkem Yurtseven, CTO, are now putting $1.65B across Series E and F behind that assignment. Sequoia Capital led the E, which closed earlier this year. Bessemer Venture Partners leads the F, with continued investment from NVIDIA. Two rounds disclosed together, supporting a company whose customers increasingly need generation inside the workday.

fal runs image, video and audio models for developers and creative teams. The appeal is straightforward: reach models through an API and spend more of the team's attention on the application, campaign or experience being built. Someone still has to make the hardware, model and serving software cooperate each time a user asks for a result.

That is where the founders' infrastructure backgrounds become useful. Burkay Gur came through Coinbase; Gorkem Yurtseven through Amazon. Building for the moment a customer presses generate means taking responsibility for the machinery behind a very human habit: looking at the result and wanting to change something.

The creative loop can keep opening new work. A different camera movement, another product image, a revised sequence. Each attempt has a duration, a cost and a place in somebody's deadline. An API supplier becomes part of the customer's ability to keep experimenting without taking over GPU operations themselves.

Bessemer describes fal expanding into workflow tools, custom infrastructure and post-training alongside model labs. That brings the company closer to decisions about what an output should look like, as well as how quickly it arrives. The relationship with the lab and the relationship with the developer begin to meet inside the same service.

The pricing page makes the buyer's job fairly clear. Model charges vary by output unit, while custom compute has its own rates. The useful calculation follows the acceptable asset through all its attempts. A cheap generation that needs several revisions can leave a very different bill from the first price on the screen.

fal says the funding supports inference capacity, its team and research into 2027. For Burkay Gur and Gorkem Yurtseven, that means building enough room for customers to keep asking for another take, while the people making the work keep discovering what they actually want.

Deeper Analysis 👇 https://devcuration.com/articles/fal-series-e-f-1-65b-generative-media

Ambiki

Ambiki lets a pediatric therapist type a note, dictate it, ask AI to organize their observations, or record the session for Tenalog to process. Keeping all 4 choices says something useful about selling software to people whose attention already belongs to a child.

The practice still needs the note finished. Then it needs the claim submitted, the payment matched and the patient's balance corrected when an insurer changes the answer. The software purchase follows that work well beyond the treatment room.

Founder Kevin Dias arrived at this problem from inside Sidekick Therapy Partners in Tennessee. Ambiki's history begins with software built there in 2019, followed by its January 2022 spinout and a January 2023 EMR launch. The administrative work came before the commercial product.

Green Hills Partners is putting additional equity behind that product and the people supporting it. Ambiki's own announcement is dated October 15, 2026; the wire release followed October 23. The amount and valuation remain undisclosed. This is a returning investor, with an existing relationship extending into the company's operating roots.

David R. “Dave” Hanson, Green Hills Partners' co-founder and CEO, is backing development across Reconsil, managed billing, data migration, implementation and clinical tools including Tenalog. The Nashville family investment platform also lists Sidekick among its investments. There is a fairly direct route here from owning an interest in therapy operations to funding the software around them.

Reconsil handles the part of the business where a payment correction can change several other numbers. Ambiki describes a ledger that recalculates patient balances and aging reports as transactions are adjusted, alongside claim submission and remittance matching. A polished note needs somewhere reliable to send its financial consequences.

The service work matters to the sale. Moving records into a new EMR and helping a practice through billing gives Ambiki responsibilities that continue after the demo. A growing clinic is deciding how much of that transition it can manage while keeping appointments on the calendar.

Tenalog brings session recordings into documentation, while AI Scribe works from a clinician's dictated observations. The clinician remains responsible for review. Ambiki is funding both the tools that help produce the note and the team that helps the practice carry its work through to the account.

Deeper Analysis 👇 https://devcuration.com/articles/ambiki-green-hills-partners-pediatric-therapy-investment

Finpilot

A pension fund's investment history can outlast several generations of software. Finpilot is selling a way to carry that history into AI workflows while keeping the systems the team already trusts.

That changes the assignment for an AI vendor. An investment team has manager letters, portfolio records, meeting notes and its own history of decisions. Getting those materials into one useful view means understanding which references belong together and which disagreements need a person.

CEO and co-founder Lakshay Chauhan and co-founder John Alberg, Finpilot's executive chairman, are building for that office. The Seattle company announced a $10M Series A on November 4, led by Lobby Capital with Madrona participating. The proceeds will support product expansion and team growth.

Madrona's original investment write-up in 2024 described a financial analyst copilot for searching sources and producing analysis. Finpilot's current proposition reaches into the recurring work around an institutional portfolio: monitoring managers, preparing committee materials and carrying approved changes into the record. There is a longer customer relationship inside that ambition.

The product's details make the sale interesting. A fund can have different names across documents and systems. Finpilot describes linking those references while keeping their origins visible, with uncertain matches left for review. An office gets to preserve its definitions and still ask a question across its information. That makes knowledge of the customer's working habits part of the product work.

Its agents also have specified places to stop. A manager update can be prepared against approved sources, then held for investment-team review when a material change appears. The software needs to know who can approve the next step before it starts doing more of the work.

Finpilot names Seattle Children's, TIFF and the Wyoming Retirement System among institutions using its platform. Those are company-reported customer relationships, each with an investment office whose own context matters. A useful report has to reach the committee in a form that team recognizes and can examine.

The hiring plan reaches across product, engineering, client success, marketing and sales. For Finpilot, the next institution brings another set of records, permissions and accumulated decisions for the team to understand before an agent can carry the office's work forward.

Deeper Analysis 👇 https://devcuration.com/articles/finpilot-raises-10m-series-a-investment-office-ai

Emvena Therapeutics

The first company a new biotech fund builds tells you where it is willing to do the difficult work. For the Penn-BioNTech Innovative Therapeutics Seed Fund, that work begins with the placenta.

The PxB Fund has launched and financed Emvena Therapeutics, a Philadelphia company developing placenta-targeted mRNA medicines. The November 4 Seed announcement leaves the investment amount undisclosed. The fund itself is $50M, formed by the University of Pennsylvania, BioNTech and Osage University Partners.

Michael J. Mitchell says the fund brought the team together in a matter of months. That is an unusually useful detail in a financing release: the investor helped assemble the organization that will have to carry the science forward.

The founding team keeps the laboratory close. Michael J. Mitchell is a Penn bioengineering professor and scientific advisor. Kelsey L. Swingle, now an assistant professor at Rice University, helped develop the technology during her Penn doctorate. Hannah Safford did her doctoral work in the same lab and is Emvena's Director of R&D. All 3 are scientific co-founders.

They are joined by co-founder Allison August, an obstetrician-gynecologist and Senior Clinical Advisor with experience at Comanche Biopharma and Moderna. Co-founder Young T. Kwon is CEO and Executive Chair, bringing prior leadership experience at Momenta Pharmaceuticals and Alchemab Therapeutics. PxB General Partner Anna Turetsky sits on the board.

That mix matters in preeclampsia. Clinicians can face a punishing decision when a mother's condition worsens: delivering the baby early may become necessary while the pregnancy still needs time. Emvena is developing an approach aimed at the placental biology involved in the disease.

Its lipid nanoparticles carry mRNA instructions. The delivery vehicle is part of the therapeutic idea. Foundational research published in Nature showed delivery of VEGF mRNA, encoding a protein involved in blood-vessel growth, reduced disease severity in mouse models. Human safety and efficacy still require research.

The Seed financing supports advancement toward IND-enabling studies and expansion into other maternal and fetal indications. No clinical launch date has been announced.

For the researchers, company formation now gives the discovery a team with different responsibilities: preserve what made the delivery science promising, develop the medicine, and keep the clinical needs of mother and baby inside the same development conversation.

Deeper Analysis 👇 https://devcuration.com/articles/emvena-therapeutics-secures-pxb-seed-for-preeclampsia-research

Nooks

Nooks has put its own sales engine in front of a second customer: the institution financing its growth. General Catalyst's Customer Value Fund is committing $500M to help the company acquire more customers, with repayment tied to the customer value that spending creates.

That makes the sales pitch unusually personal. Nooks runs its own go-to-market on the software it sells, so the economics behind this financing are also part of the product's argument to a revenue leader.

Co-founder and CEO Dan Lee, co-founder and CTO Nikhil Cheerla, and co-founder and CPO Rohan Suri started Nooks in 2020 as a virtual office. Sales teams pulled them toward a more specific problem: helping people get through to buyers and carry those conversations forward.

The company now reports more than 2,000 customers, approximately 3x year-over-year growth and 135% net revenue retention. More than half its customers use multiple products. Those are company-reported figures, and General Catalyst is backing the customer economics with non-dilutive growth capital rather than buying an equity stake in this transaction.

Kleiner Perkins led its $43M Series B in 2024. This commitment comes with a different claim on the business. There is no new valuation attached to the announcement. The repayment terms were not disclosed, so treating the commitment as a conventional $500M equity round would tell the wrong story about what Dan Lee chose to finance.

Alongside the commitment, Nooks introduced its Revenue Agent Platform and Node, a proactive assistant beginning with deal management. A promised follow-up, a quiet buying-committee member, a deal losing momentum: Node carries that context between conversations and brings the rep a proposed next move. Reps approve sends.

That is a longer relationship to earn than helping someone make another call. Nooks is asking account executives, account managers and solutions engineers to let software remember the work passing between them. Node's general availability is planned for early 2027; early access is open now.

The capital will support new products, enterprise and international reach, and hiring. Nooks says it has more than 100 open roles. The founders are carrying a business built around outbound into the meetings, promises and renewals that decide whether an account stays.

A buyer returning to a conversation is still waiting on whatever the rep promised last time. That small piece of memory now sits inside the much larger business Dan Lee and the Nooks team are financing.

Deeper Analysis 👇 https://devcuration.com/articles/nooks-500m-general-catalyst-growth-capital-node

Phaze

Helping an AI agent can mean joining the computer it is already using. Phaze lets a technician take over the same remote session, resolve what stopped the agent, and hand control back. The employee becomes part of how the work continues.

Benjy Boxer, CEO, and Christopher Dickson, CTO, are building that relationship into Phaze after previously building Parsec. Their new company announced $8.75M in Seed funding on November 4, led by Andreessen Horowitz, with Indie, NextView, Makers and Lerer Hippeau participating.

The founders know the human customer rather well. An artist turning a complex 3D model or an engineer working in CAD notices when the machine hesitates. Moving that workstation somewhere else still leaves the software with an obligation to the person moving the mouse. Phaze is carrying that experience into a product that also lets agents operate remote desktops and send terminal commands.

There is a useful commercial connection here. A company can keep a powerful machine under its own management, give an employee access when needed, and let an agent work on another computer while that employee keeps using their own. Access becomes something IT can assign across a group of machines, instead of a permanent relationship between one workstation and one desk.

Phaze's current AI page labels its MCP server experimental. Its support-bot example shows an agent asking for help when it gets stuck or reaches a Windows permission prompt. Those interruptions give the technician an actual job inside the workflow, and put some substance behind the familiar promise of human oversight.

The company describes enterprise authentication, directory sync, role controls and audit logs around that access. Its public company pricing is $0.20 per hour, capped at $20 per month per user. Charging for connection time gives the vendor a reason to care whether the software gets used after installation, a more interesting relationship than counting every name in the directory.

Martin Casado at a16z is backing a team the firm already knew through Parsec. His investment thesis follows agents into the operating systems and applications where work already happens, including workloads that depend on particular hardware. Phaze says the funding will expand engineering, advance the product and support customer growth.

That leaves Boxer and Dickson serving two users of the same machine: the person who can feel a delayed response, and the agent that may need that person to come back into the session.

Deeper Analysis 👇 https://devcuration.com/articles/phaze-raises-8-75m-seed-for-human-and-ai-remote-access