The round prices a decision-automation platform the company will not put a number on, and it lands on the one function most alt-lenders still run by hand: the credit desk that turns a tape into a yes or a no.

What happened. Taktile, a credit-decision automation company founded in 2020 in Berlin by machine-learning engineers Maik Taro Wehmeyer and Maximilian Eber and schooled in the Y Combinator class of that summer, raised a $110 million Series C led by Growth Equity at Goldman Sachs Alternatives, with Index Ventures, Tiger Global and Y Combinator participating. The round was announced on June 24, 2026 and disclosed first to Fortune.¹ It brings the company's total funding to about $184 million and follows a $54 million Series B in February 2025, a fast step up that the company says it will spend on global expansion across the United States, Europe and Latin America.²

What is being funded. The product is what Taktile calls an agentic decision platform: software that runs the multi-step credit and risk decisions a lender would otherwise staff, from customer approval to fraud and anti-money-laundering screening to business-loan underwriting. The company reports that it has reached 95% automation in B2B underwriting for customers, cut anti-money-laundering false positives by 75%, and lets institutions process five times more small-business loans without adding staff.² Chief executive Maik Taro Wehmeyer frames 2026 as the year AI moves into financial services in production rather than in pilots.¹

Why an alt-lending desk should care, and the limit of it. The function Taktile automates, turning a borrower's data into an underwriting decision in seconds, is the exact cost center a merchant cash advance, factoring, equipment, or revenue-based lender lives on. That said, the named customers in the public record are neobanks and spend-management fintechs such as Mercury, Monzo, Faire and Pleo, not MCA or factoring shops.³ So the honest read is not that the alt-lending mid-market is already running on this. It is that the tooling layer your bank and fintech competitors are buying just took a nine-figure check from the most credible name in the room.

The part to keep in view. Wehmeyer declined to disclose the valuation.¹ For an operator, the price tag is not the missing number that matters. The missing number that matters is the loss curve: a 95% automation rate is a throughput claim, not an accuracy claim, and no automated underwriting stack on the market has a published performance record through a full credit downturn. The capital is real and the speed is real. The cycle test is still unwritten.

_{Sources
1 Fortune | Exclusive: Taktile Raises $110M Led by Goldman Sachs to Automate High-Stakes Banking and Insurance Decisions
2 PYMNTS | Taktile Raises $110 Million to Automate High-Stakes Banking and Insurance Decisions
3 Business Wire | Taktile Secures $110M in Goldman Sachs-Led Series C to Power AI Transformation in Financial Institutions
4 Finextra | AI-Native Loan Origination Platform Lama AI Lands $20M Series A
5 SiliconANGLE | Lama AI Raises Series A to Accelerate Automated Loan Originations
6 Finance Magnates / FF News | Lama AI Secures Series A to Scale Agentic AI for Community Banks
7 BNN Bloomberg | Float Financial Raises CAD $85M Series C to Build Canada's Intelligent Financial Operating System
8 The Globe and Mail | Toronto Fintech Float Hits $548 Million Value With New Funding
9 Bloomberg | Private Credit's Big Arbitrage Trade Gains Backing From Advisers
10 Yahoo Finance | Private Credit's Big Arbitrage Trade Gains Backing From Advisers
11 American Banker | Is Private Credit a $2 Trillion Insurance Timebomb?
12 PR Newswire | Commercial Credit Group Completes CCGR Trust 2026-1 Term ABS Issuance of $517.9 Million
13 Equipment Finance News | Commercial Credit Group Closes $517.9M ABS Deal}

Because the thing being automated is your most expensive judgment, not a back-office chore. An alternative lender's edge has rarely been its cost of capital, which a bank usually beats. The edge is speed and a willingness to underwrite files a bank will not touch, fast enough that the merchant takes the offer in front of them. Taktile's pitch goes straight at the first half of that edge: it says a lender can run 95% of business-loan underwriting decisions through software and clear five times the small-business volume without hiring.²

If that holds at production scale, the turnaround gap that lets a non-bank lender win a deal a bank is still committee-reviewing starts to close from the bank side. The competitive question is no longer whether you can decide faster than an incumbent. It is whether you can decide faster than an incumbent that has bought the same automation layer.

The platform sits between a lender's data sources and its decision, orchestrating the steps a credit or risk analyst would otherwise run: pull the applicant data, run the models and policy rules, screen for fraud and anti-money-laundering flags, and return an approve, decline or refer.¹ The three figures it leads with are a 95% automation rate in B2B underwriting, a 75% cut in anti-money-laundering false positives, and a five-times increase in small-business loan throughput at constant headcount.²

Read those as vendor-reported outcomes, because that is what they are. They describe how much of the work runs without a human and how much manual review noise drops, which are real operational wins. They do not describe whether the automated decisions were the right decisions, and the company has not published default or loss data tied to the automated book. There is a parallel risk upstream of the model itself: an automated decision is only as sound as the borrower and entity data it ingests, and a stack fed by stale or batched records will reproduce those errors at machine speed, on every file, rather than catching them the way a human reviewer occasionally would. Goldman's Jade Mandel, who led the round, cites faster product launches, sharper risk outcomes and operational efficiency as the appeal.¹ Two of those three are throughput claims. The risk-outcome claim is the one without a public number behind it.

The customers in the public record skew toward digital banks and spend-management platforms: Mercury, Monzo, Faire and Pleo are the names attached to the company.³ Those are businesses that approve cards, accounts and short-line credit at high volume, which is a natural fit for decision automation. None of the disclosed marquee customers is a merchant cash advance funder, an invoice factor, or an equipment lessor.

That gap is the honest boundary on this story. It would be a stretch to tell you the MCA mid-market is already underwriting on Taktile. The defensible claim is narrower and still matters: the same engine that lets a neobank clear card and account decisions at scale is being sold, and funded, as general-purpose financial decisioning, and the segments next to yours are the early adopters. When the tooling that compresses underwriting cost becomes a procurement decision rather than a build, the lenders who move first reset the cost baseline everyone else is measured against.

Wehmeyer explicitly declined to give the valuation on the $110 million round.¹ A private company is under no obligation to, and a withheld number is often just negotiating posture. For a lender reading this as market intelligence, the valuation is the least useful figure anyway. What the undisclosed price does is remove the one external check, a marked-up or marked-down round, that would tell you how the market is pricing automated-underwriting risk versus automated-underwriting growth. Absent that, you are left to weigh the capital and the named lead, which point to conviction, against the absence of any loss-performance disclosure, which points to an unproven core.

It is a wave, and it is worth sourcing rather than asserting. The same week Taktile closed, Lama AI raised a $12 million Series A led by EJF Ventures that lifted its total funding above $20 million, for AI-native loan origination already live at community banks including SouthState and Colony.⁴⁵ That round puts AI decisioning inside the community-bank stack, the exact institutions that compete with non-bank lenders for small-business credit.⁶ In the same stretch, Goldman Sachs Alternatives also backed Float Financial's CAD $85 million round at a roughly $548 million valuation, a working-capital and spend platform for businesses.⁷⁸ Read together, the signal is that institutional capital is funding the automation of small-business credit decisioning from several directions at once, not betting on a single company.

Three moves, each with one job.

First, benchmark your own decision economics before a competitor sets them for you. Put a real number on your fully loaded cost and median time to underwrite a file today. You cannot tell whether a 95%-automation pitch is a threat or noise until you know what your manual desk actually costs per decision and how long it takes.

Second, pilot automation on the reversible decisions, not the underwriting itself. The lowest-risk entry is the work where a wrong call is cheap to unwind: document intake, fraud and anti-money-laundering pre-screening, and routing. Capture the throughput gain there and keep your loss exposure unchanged while you learn the failure modes.

Third, keep a human on the adverse-action and exception path, and demand loss data before you automate the credit call. The credit decision is where an automated error becomes a charge-off and a fair-lending question at the same time. Before any vendor underwrites your book, require performance and loss data through a stressed vintage, not just an automation rate, and keep human review on declines and edge cases so the speed gain never costs you the audit trail.

A registered entity and a finished storefront do not prove the seller owns the submitted tax ID, and that gap surfaces at 1099 time as finance cleanup.

Cobalt's TIN/EIN Verification API validates the seller's business name and tax ID against IRS records at onboarding, returning an IRS code your platform can route.

Schedule a FREE demo call

Financial advisers are now formally recommending an arbitrage that points straight at the funding base under private credit: redeem non-traded business development companies at full net asset value, then buy publicly listed BDCs trading at a median discount of about 25% to their own book, an 83% price-to-book against a long-run average near 95%.⁹ The trade is gaining traction because the non-traded vehicles are gating: Morgan Stanley capped its fund after redemption requests hit 11.6% of a quarter, Apollo capped after a 17% request, and the listed BDC index is down about 11% so far in 2026 even as leveraged loans sit slightly positive.¹⁰ The operator signal for any lender whose warehouse line traces back to a non-traded BDC: that capital pool is facing redemption pressure and visible repricing, so stress-test your line availability for the third quarter now and ask your provider directly where its own funding sits, rather than assuming a multi-year facility is insulated from a retail redemption cycle.

US life insurers have put nearly a third of their $5.6 trillion in assets into private credit while moving close to $2 trillion of policyholder liabilities into offshore and captive reinsurers, roughly $1.3 trillion offshore and $600 billion in domestic captives by one forensic estimate, even as the minimum collateral to back such deals has fallen from $250 million in 2017 to as little as $2 million in some states.¹¹ Regulators are visibly uneasy: Treasury Secretary Scott Bessent convened an emergency meeting with state insurance commissioners in May, Fed Vice Chair Michelle Bowman said in early June that it is unclear what the bank lending behind nonbanks actually funds, and the NAIC's new Actuarial Guideline 55 has drawn initial cash-flow-testing filings from about 80 insurers.¹¹ The operator signal is that the capital chain funding direct lending now runs through lightly capitalized insurance structures regulators admit they cannot fully see, so a lender relying on insurance-linked or private-credit funding should know which insurer balance sheet ultimately backstops its facility and treat that concentration as a real counterparty question, not an abstraction.

Independent equipment lender Commercial Credit Group priced CCGR Trust 2026-1, a $517.93 million term ABS backed by loans and leases on construction, fleet transportation, manufacturing and waste-industry equipment, on June 22, its 21st securitization in the program.¹² Wells Fargo Securities ran the deal as structuring agent and lead bookrunner across five tranches, with the senior $342.56 million class rated AAA by Fitch and Aaa by Moody's, a clean print for a non-bank lender to small and mid-sized businesses.¹³ The operator signal is a useful counterpoint to the BDC stress above: capital markets remain fully open for seasoned, well-documented equipment paper from an established issuer, so an equipment or asset-based lender weighing a term-ABS execution should read this as confirmation that the window is open for clean collateral, while the redemption pressure on private-credit vehicles is a separate and more crowded funding lane.