Aiolum Browse stories
All stories
AI agents for carrier-bill audit, customer invoicing and collections at freight brokers and 3PLs · United States

Lighthouz AI: Product, team and how it was built

Small teamFundedB2BLogistics

Lighthouz AI (AI agents for carrier-bill audit, customer invoicing and collections at freight brokers and 3PLs, United States): Used by 40+ freight brokers and 3PLs, September 2026. Evidence grade E: no revenue figure was checked. Team: 2 married co-founders at the start in 2023; 8 people on the Y Combinator profile, November 2026. Launched March 2025; still running.Added 8 November 2026

Outcome
Used by 40+ freight brokers and 3PLs, September 2026
Evidence
E no revenue figure was checked
Launch
March 2025
Status
Still running
Team
2 married co-founders at the start in 2023; 8 people on the Y Combinator profile, November 2026
Outside funding
Y Combinator, Summer 2024 batch; no amount stated in the sources
Main channel
Cold outreach
No revenue figure

What is not verified

No revenue figure was found. The sources read were the company's site, its two Y Combinator launch posts, its Y Combinator profile and archived copies of the site; none gives revenue, prices or a funding amount, and no payment-processor profile, filing or press interview with figures was found. The headline, 40+ freight brokers and 3PLs, is the company's own statement of 7 September 2026. A reader could look for a funding announcement, a TrustMRR or payment-processor profile, or the Transportation Intermediaries Association's award records. This story is grade E: the business is real and its product, price and customers are described from public pages, but no revenue figure was checked. Aiolum shows it as an idea worth studying, not as proof that it makes money.

Lighthouz AI sells AI agents to freight brokers, companies that take a shipper's load and hire a trucking company (the carrier) to move it, and to 3PLs, third-party logistics providers. The agents work in accounts payable (AP, checking and approving the bills that carriers send), in accounts receivable (AR, creating and sending invoices to shippers) and in collections. Their stated goal is a lower DSO, days sales outstanding: the average number of days between an invoice and the customer's payment.

The founders are Srijan Kumar, a former computer science professor at Georgia Tech and research scientist at Google, and Sonali Pattnaik, a former lead AI engineer at Halliburton, Progressive Insurance and American Family; they met at IIT and are married. The Y Combinator profile says the company was founded in 2023 in Atlanta and joined the Summer 2024 batch. Archived copies of the site show an AI evaluation platform in April 2024, then "Pam, the AI Procurement Assistant" until February 2025, and AR and AP agents for freight from March 2025. In a Y Combinator launch post of 7 September 2026 the company said it "powers finance operations of 40+ freight brokers and 3PLs" and named four of them.

This page has no revenue figure (grade E): the company publishes no prices and no revenue, and its customer count is its own statement. The Y Combinator profile lists 8 people in November 2026. The page also shows failed companies that promised to automate back-office or freight work.

Lighthouz AI: Used by 40+ freight brokers and 3PLs, September 2026. AI agents for carrier-bill audit, customer invoicing and collections at freight brokers and 3PLs, United States, team of 2 to 10, evidence grade E.
The story in one card. Share it with the address of this page.
01

How much revenue does Lighthouz AI make?

Each row shows the evidence grade and the source numbers from the list at the end of this page. The grades are explained in How Aiolum verifies.

Two founders, two earlier AI products, and a freight back-office product that the company says 40+ brokers and 3PLs use; no price and no revenue published
FactValueGradeHow we knowSources
Customers, 7 September 202640+ freight brokers and 3PLsCCompany's Y Combinator launch post: "Lighthouz already powers finance operations of 40+ freight brokers and 3PLs, including Hoplite Logistics, Southern Reins Logistics, ADL, and NAD Logistics." The company's own count; no customer list was published.[1]
Customers, 22 December 2025no count givenCEarlier launch post: "We are already working with leading freight brokers and 3PLs in the US and Canada."[2]
Company founded2023, in Atlanta, GeorgiaCY Combinator company profile, read 8 November 2026: year founded 2023, location Atlanta, GA, batch Summer 2024, status Active.[3]
Founding team2 co-founders: Srijan Kumar (CEO) and Sonali Pattnaik (COO)CY Combinator profile and the company's About page. The launch post says they met during undergraduate studies at IIT and have been married for more than 5 years. Archived pages of 2024 and 2025 call Pattnaik the CTO.[3, 5, 2, 9]
Team size, November 20268CY Combinator profile, team size field, read 8 November 2026.[3]
FundingY Combinator, Summer 2024 batch; amount not statedCThe About page calls the company "a YC S24-backed freight back-office automation platform". No source read for this page gives an amount or a later round.[5, 3]
Earlier productsAI evaluation platform (April 2024); AI procurement assistant "Pam" (September 2024 to February 2025)CInternet Archive copies of lighthouz.ai. 20 April 2024 description: "a platform for generative AI product evals and enhancement of performance and safety". 9 September 2024, 5 January 2025 and 13 February 2025: "The AI Procurement Assistant".[8, 9]
Freight AP and AR product first seen15 March 2025CInternet Archive copy of 15 March 2025: AR and AP agents that "Verify carrier Proof of Delivery with AI" and "Automate freight rate verification". The copy of 13 February 2025 still shows the procurement assistant.[10, 9]
Pricesnot publishedCNo pricing page; the home page offers a demo and a call with the CEO.[4]
Share of carrier bills with no human step70% to 85%CCompany claim on the home page: "70%–85% of bills clear with zero human touch." The Y Combinator profile also says brokers invoice "36% faster", are paid "5 days sooner" and cut back-office costs "by 40%". These are customer results stated by the company, not revenue.[4, 3]
Case-study customer resultsDSO from 37.60 to 32.78 days; 120-140 to 450-550 invoices a week with the same teamCCompany case study of Southern Reins Logistics by Sonali Pattnaik, dated 7 July 2026. The customer's own outcomes as the company reports them.[6]
Industry awardTIA Technovator Award, 2025 (and 2026 in the September 2026 post)CHome page: "TIA 2025 Technovator Award Winner". Launch post of 7 September 2026: "has won the TIA 2025 and 2026 Technovator Awards." Not checked with the awarding association.[4, 1]
Time from the freight product to the headline18 monthscalcCalculated: March 2025 (first archived copy of the freight product) to September 2026.[10, 1]

How to read this table: a grade C value is what the company, its founders, its Y Combinator pages or its own case study say. A "calc" value (grade X) was computed on this page. No value is grade A or B, and no revenue figure was found.

Lighthouz AI prices in November 2026: no public price list (grade C)
PlanPer month, billed yearly (US$)Per month, billed monthly (US$)Limit
Collections: Email Collections Agent, Portal Collections Agent, work queue, AR risk dashboardprice not published
Accounts receivable: Shipper Invoicing Agentprice not published
Accounts payable: bill processing and audit, rate verification, payment status replies, disputesprice not published

The site has no pricing page. The home page offers "Book a demo" and "Book a call with the CEO". The rows below are the three modules named on the home page, read on 8 November 2026. The customer count and the four customer names come from the company's launch post of 7 September 2026, and one customer, Southern Reins Logistics, is the subject of a case study on the company's site.

02

When did Lighthouz AI launch, and how did it grow?

From an AI evaluation platform in 2024 to freight accounting agents with 40+ customers, as the company counts them, in September 2026. Each entry carries its grade.

  1. 2023CCompany founded in Atlanta, according to its Y Combinator profile. [3]
  2. 20 April 2024CThe site describes Lighthouz as a platform for evaluating generative AI products for performance and safety. [8]
  3. Summer 2024 to February 2025CY Combinator Summer 2024 batch. The site sells Pam, an AI procurement assistant that enters quotes and invoices into ERP systems and creates purchase orders. [3, 9]
  4. 15 March 2025CThe site now offers AR and AP agents with freight tasks: checking carrier proof of delivery and freight rates. [10]
  5. 22 December 2025CY Combinator launch post for AP and AR automation for freight brokers and 3PLs, already in use at brokers in the US and Canada. [2]
  6. 7 July 2026CCase study of Southern Reins Logistics, a freight brokerage founded in Alpharetta, Georgia, and twice on the Inc. 5000 list, on carrier-bill audit. [6]
  7. 7 September 2026CSecond launch post, for Lucie, a collections agent. The company says 40+ freight brokers and 3PLs use Lighthouz. [1]
03

Why did Lighthouz AI work?

Four causes, in the order they mattered. Each one is a reusable lesson.

01

One industry's paperwork

The product is built for freight documents only: carrier bills, proof-of-delivery papers, bills of lading and lumper receipts, with multi-stop loads, rotated scans and handwritten notes. The launch post calls this "freight-native". A general invoice tool does not know what a detention charge or a fuel surcharge on a truckload bill should be.

02

Agents that act, people for exceptions

The agents do the work, not only suggest it: they approve clean carrier bills, file disputes, submit invoices to customer portals and send reminders. A person sees only the exceptions and a ranked queue of accounts. The company says this cuts the work that grows with the number of loads.

03

Two pivots to a buyer with a clear cost

The founders tried AI evaluation and then procurement before freight. In freight brokerage the cost of slow invoicing is visible in days of unpaid invoices, and the head of accounting at the case-study customer said the next hire would not be in accounts payable.

04

Hands-on onboarding

The case study describes in-person demonstrations, on-site support and a three-to-six-month period in which the system was tuned to the customer's rules before staff trusted it.

How it was built

These details come from the company's launch posts, its Y Combinator profile and its own website. They are grade C. [2, 4, 1, 7, 3]

  • The agents read carrier bills, proof-of-delivery papers, bills of lading, lumper receipts and invoices from email and uploads, and extract rates, accessorial charges, remit-to details, signatures and handwritten notes.
  • They compare each carrier bill with the rate confirmation or quote, flag duplicates, overbilling, missing papers and invalid accessorials, and post the checked data into the broker's TMS (transportation management system) with an audit trail.
  • The home page lists connections to TMS products such as McLeod, Turvo, MercuryGate, Revenova and Tai, and to QuickBooks, NetSuite and Sage Intacct for accounting.
  • The collections agents send reminder emails on a cadence set per customer and check customer AP portals such as Cass, Uber Freight and E2Open for invoice status.
  • The marketing website is built with Lovable: its page source names Lovable as author, and its sitemap lists pages on a lovable.app address.
  • No source names the language model behind the agents. The founder's Y Combinator biography says the agents are "powered by dynamic knowledge graphs".

Tools named in the sources

Each one opens its page in the Tools tab, with every story that uses it.

04

Who tried this before Lighthouz AI, and why did they fail?

Three companies in our database of failed startups promised to automate back-office or freight work and failed. The values below are from that database and are not independently verified.

Died 2012 to 2023Burned US$852M

Olive AI

The promise

An AI workforce to automate hospital administration: bots for prior authorizations and billing.

Why it died

The database says it sold autonomous AI but delivered automation that needed many engineers per customer, and costs outran revenue.

Unit Economics
Died 2015 to 2018Burned US$290K

Shipitwise

The promise

Automated freight quotes and logistics for small companies without a shipping department.

Why it died

It depended on unreliable third-party logistics partners and spent its money on complex integrations.

Unit Economics
Died 2012 to 2015Burned US$500K

Keychain Logistics

The promise

Matching truck drivers with shippers to reduce empty miles.

Why it died

Better-funded freight platforms took the market, and its technology could not keep up before the money ran out.

Competition

The failed startups come from our database of 1,749 dead startups, retold here in our own words. Its entries are AI-assisted summaries of public sources, so these cards carry no evidence grade.

What changed

Olive AI promised AI workers for a whole administrative department, and the database says the work behind them was custom engineering for each customer. Shipitwise and Keychain Logistics tried to change how freight is bought and moved, against large competitors and many integrations.

Lighthouz does not try to change how freight moves. It takes one narrow job inside the broker, reading and checking freight paperwork, including scanned and handwritten documents as the launch post says, and it connects to the TMS that the broker already uses. The risk Olive AI shows is still present: the case study describes months of tuning and on-site support per customer.

05

Can one person copy Lighthouz AI now?

Only with industry knowledge and AI engineering.

The founders are AI researchers and engineers who also spent two pivots finding their buyer. A solo founder could copy the method, one industry's documents and one back-office job, but would need deep knowledge of that industry's billing rules and integrations with its main software.

Where it is still open: Back offices with heavy, irregular paperwork in other trades: customs brokers, freight forwarders, construction subcontractors, equipment rental and medical billing.

  1. 1Pick one industry whose back office still runs on PDFs, scans and email, and learn its billing rules in detail.
  2. 2Automate one job end to end, approving the clean cases and sending only exceptions to a person.
  3. 3Connect to the industry's main system of record, so customers do not change software.
  4. 4Win one customer, measure the before and after in their own numbers, and publish the case study.

What could go wrong for Lighthouz AI?

The risks are the mirror image of the causes above.

What is not verified

No revenue figure was found. The sources read were the company's site, its two Y Combinator launch posts, its Y Combinator profile and archived copies of the site; none gives revenue, prices or a funding amount, and no payment-processor profile, filing or press interview with figures was found. The headline, 40+ freight brokers and 3PLs, is the company's own statement of 7 September 2026. A reader could look for a funding announcement, a TrustMRR or payment-processor profile, or the Transportation Intermediaries Association's award records.

Results are the company's own figures

The 70% to 85% no-touch rate, the faster invoicing and the lower back-office cost are company claims. The Southern Reins case study gives three different growth figures for the customer: 50% in its title, 47% in 2025 in its results, and 40% in 2025 in a quoted remark. The home page says "SOC 2 Type II Compliant"; no audit report is linked from the pages read.

AI emails and portal logins

The Email Collections Agent sends payment reminders "trained on your SOP, your tone" and replies to basic customer emails. The sources do not say whether these emails tell the recipient that an AI wrote them. The Portal Collections Agent logs into customer payment portals; the sources do not say how those portals' terms treat automated access.

Competition and model providers

Other AI back-office stories on this site are Cleavr, which chases late B2B invoices for any industry, and 5U AI, which sells AI workers for freight forwarders; Lighthouz is narrower than Cleavr and works for brokers, not forwarders. No source names the language model provider, so how much the product depends on one provider's prices or terms is not known.

A young product from a team of eight

The freight product is about 18 months old at the headline date, after an AI evaluation platform and a procurement assistant. The Y Combinator profile lists 8 people, more than the two founders, and the case study describes months of tuning per customer.

Sources

The grade next to each source is the best grade that the source can support. "Read in full" means that the text of the page was downloaded and the numbers were checked in it.

  1. C Y Combinator Launches, 7 September 2026: Lighthouz AI: AI to automate collections for freight brokers and 3PLs
    Read in full (the page returned its JSON record, 8 November 2026). Company post. Source for the 40+ customers, the four named customers, the collections agents and the TIA awards.
  2. C Y Combinator Launches, 22 December 2025: Lighthouz AI: AI to automate AP and AR for freight brokers and 3PLs
    Read in full (JSON record, 8 November 2026). Founders' post. Source for the founders, the workflow, the 24-hour invoicing claim and the US and Canada customers.
  3. C Y Combinator company profile: Lighthouz AI
    Read the company data embedded in the page, 8 November 2026: year founded, team size, location, batch, founder biographies and description.
  4. C Lighthouz AI home page, read 8 November 2026
    Read in full, and the page source. Source for the modules, integrations, demo offer, testimonials, award and the Lovable author tag.
  5. C Lighthouz AI: About
    Read in full, 8 November 2026. Source for the founding team and the YC S24 backing.
  6. C Lighthouz AI, 7 July 2026: Case Study: How Southern Reins Logistics Saved 40% Back-Office Cost with Lighthouz AI While Growing 50% in Revenue
    Read in full, 8 November 2026. Company case study by Sonali Pattnaik. Source for the customer's results and the onboarding period.
  7. C Lighthouz AI: AI for Collections module page
    Read in full, 8 November 2026. Source for the portal list and the collections agents.
  8. C Internet Archive copy of lighthouz.ai, 20 April 2024
    Read the page description (the page body loads by script). Source for the AI evaluation product.
  9. C Internet Archive copy of lighthouz.ai, 13 February 2025
    Read in full; the copies of 9 September 2024 and 5 January 2025 show the same procurement assistant. Source for Pam and the founders' earlier titles.
  10. C Internet Archive copy of lighthouz.ai, 15 March 2025
    Read in full. First archived copy with the freight AP and AR agents.
  11. C Database of failed startups, local export
    Local export. AI-written summaries, not independently verified. Source for the failed predecessors.

Run this company? Send corrections or a verified figure to aiolum@livyden.com

Back to all stories