Corvera: Product, team and how it was built
Corvera (AI sales agent for consumer-goods brands that sell through distributors and retailers, United States): More than a dozen brands signed, May 2026. Evidence grade E: no revenue figure was checked. Team: 3 co-founders and 1 founding engineer at launch in January 2026; Y Combinator lists 4 people in November 2026. Launched 21 January 2026; still running.Added 6 November 2026
- Outcome
- More than a dozen brands signed, May 2026
- Evidence
- E no revenue figure was checked
- Launch
- 21 January 2026
- Status
- Still running
- Team
- 3 co-founders and 1 founding engineer at launch in January 2026; Y Combinator lists 4 people in November 2026
- Outside funding
- US$6.2M in total by the company's account, including a US$4.2M seed round announced in May 2026
- Main channel
- Cold outreach
What is not verified
No revenue figure was found. The headline, more than a dozen brands signed within five weeks of the first customer, is the CEO's own statement in the seed-round post of 5 May 2026, and no later count or customer list was published. Looked for: a payment-processor profile, a founder interview with figures, a filing. A reader could look for a later interview or a company register entry. The total of US$6.2M raised is the company's own text on its Y Combinator page; the seed post names US$4.2M. 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.
Corvera sells to CPG brands (consumer packaged goods: food, drinks and household products sold in shops). Such a brand gets depletion data, the cases each distributor sold on to stores, from many distributors and retailer portals at once. Corvera joins those feeds into one context layer, a single cleaned and access-controlled copy of the data, ranks the stores worth a call, and drafts the follow-up. It also serves the data over MCP (Model Context Protocol, an open standard that lets AI assistants read outside data), so staff can ask Claude or ChatGPT about their own accounts.
Christopher Kong, who ran the tempeh brand Better Nature for seven years, started it in 2025 with Dirk Breeuwer, a former data and AI lead at Google, and Matthew Collins, a former product head at Rosemark. It launched in Y Combinator's Winter 2026 batch on 21 January 2026 as "AI agents for CPG brands" with one founding engineer. On 5 May 2026 the company stated that it had raised a US$4.2M seed round and had signed more than a dozen brands within five weeks of its first customer.
No revenue figure was found in any source, so this page is grade E. The headline is the company's own customer count. The customer results on this page (queries run, opportunities found) are from Corvera's own case studies and describe customers, not Corvera's income.

How much revenue does Corvera 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.
| Fact | Value | Grade | How we know | Sources |
|---|---|---|---|---|
| Brands signed, May 2026 | more than a dozen, within five weeks of the first customer | C | The CEO's seed-round post of 5 May 2026. No customer list or later count is given. | [4] |
| Founded | 2025 | C | Y Combinator company page, read 6 November 2026. | [1] |
| Public launch | 21 January 2026 | C | Y Combinator launch post, then titled "AI agents for CPG brands". | [2] |
| Team at launch | 3 co-founders and 1 founding engineer | C | The launch post names Christopher Kong (CEO), Dirk Breeuwer (CTO), Matthew Collins (CPO) and a founding engineer who leads AI engineering. | [2] |
| Team, November 2026 | 4 people | C | Y Combinator lists a team size of 4. The company's team page shows three people and not Matthew Collins. | [1, 7] |
| Seed round, May 2026 | US$4.2M | C | The company's post of 5 May 2026: led by 6 Degrees Capital, with 20VC, Rebel Fund, Duke Capital Partners, Multimodal Ventures and Y Combinator. | [4] |
| Total raised | US$6.2M | C | The company's text on its Y Combinator page. It does not say where the US$2M above the seed round came from (calculated: US$6.2M minus US$4.2M). | [1, 4] |
| Prices | not published | C | Demo only. Target customers: US brands with more than US$20M in retail revenue, according to the Y Combinator page. | [1, 3] |
| Largest case study, August 2026 | 12 users at one brand, 5,000 queries in 40 days | C | Corvera's case study of the drinks brand CLEAN Cause, 18 August 2026. A customer's usage, not Corvera's revenue. | [5] |
| Second case study, August 2026 | 1,100+ queries in 22 active days | C | Corvera's case study of Wild, a deodorant brand that the study calls a Unilever brand, 26 August 2026. | [6] |
| How customers query the data | in Corvera, or through Claude and other MCP tools | C | FAQ and both case studies. CLEAN Cause and Wild staff asked their questions "via Claude" and "through Claude". | [5, 6, 8] |
| Time from launch to the headline | 4 months | calc | Calculated: January 2026 to May 2026. | [2, 4] |
How to read this table: a grade C value is a statement by the company on its own site or on its Y Combinator pages. No value has been checked by a payment processor, a filing or a journalist. A "calc" value was computed on this page.
| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Contract after a demo | no prices published |
The site has no pricing page. Every page leads to "Schedule demo", and the seed-round post offers a thirty-minute walkthrough. The Y Combinator page says the target customers are US brands with more than US$20M in retail revenue, and that Corvera sells to vice presidents of sales and chief executives, which suggests contracts priced per customer.
When did Corvera launch, and how did it grow?
From a launch post about back-office agents to an AI sales agent, in ten months. Each entry carries its grade.
- 2025CFounded by Christopher Kong, Dirk Breeuwer and Matthew Collins. [1]
- 21 January 2026CLaunch post in Y Combinator's Winter 2026 batch: "the AI agent workforce for CPG back-office", for sales orders, inventory and demand forecasts. It asks for introductions to brands with more than US$50M in revenue. [2]
- 5 May 2026CSeed round of US$4.2M announced. The product is now described as a context layer served to AI tools over MCP. More than a dozen brands signed within five weeks of the first customer. [4]
- 18 and 26 August 2026CCase studies of CLEAN Cause (US drinks) and Wild (UK deodorant) published on the company site. [5, 6]
- November 2026CThe site and the Y Combinator page call the product "the AI sales agent for CPG brands". Y Combinator lists 4 people and US$6.2M raised. [1, 3]
Why did Corvera work?
Four causes, in the order they mattered. Each one is a reusable lesson.
Domain experience
The CEO ran a consumer brand for seven years, sold it into about 5,000 stores, and writes that he has "sat in the seat our customers sit in". The product starts from the exact pain he knew: distributor reports that disagree on store and product names.
Name matching first
Before any AI answer is useful, the same store must have one identity in every feed. In the CLEAN Cause study, nearly 8,000 stores were matched across feeds, so a store that changes distributor no longer looks like a lost customer.
Use the customer's AI assistant
Instead of building its own chat window only, Corvera serves the cleaned data over MCP, and customers ask Claude. The model provider does the language work; Corvera supplies the data and the business terms.
How it was built
These details come from the company's own pages and Y Combinator posts. They are grade C. [2, 3, 4, 5, 6, 8]
- Connectors for distributor feeds, retailer portals, syndicated data and ERP systems; the seed-round post names NetSuite, Cin7, Unleashed, Shopify, Amazon, Snowflake, Celigo and Tableau.
- A mapping engine that matches each retailer's product codes and store names to one master list.
- An MCP server that lets Claude, ChatGPT and other MCP tools read the governed data, with access control and an audit log.
- A daily ranked action list and scheduled email reports, with drafts to reps, brokers or buyers that a person approves.
- The company website is a Next.js application, as its page source shows.
Tools named in the sources
Each one opens its page in the Tools tab, with every story that uses it.
Who tried this before Corvera, and why did they fail?
Three companies in our database of failed startups promised to turn scattered sales or distribution data into actions, and failed. The values below are from that database and are not independently verified.
Index
The promiseSales intelligence that joins scattered data sources into real-time insights for sales teams.
Why it diedSalesforce already sat in its customers' systems, and Index could not show a clear, quick return that justified switching.
Slik
The promiseA platform that collected and analysed sales data to rank which leads to work first.
Why it diedIt could not set itself apart from LinkedIn Sales Navigator and ZoomInfo, and keeping its data fresh was costly.
ShopX
The promiseA B2B marketplace connecting consumer-goods brands directly to India's small neighbourhood stores, past the distributors.
Why it diedThin margins on consumer goods met a high cost of signing each store, and brands protected their distributors.
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
Index and Slik sold general sales intelligence and lost to large platforms with more data. Corvera picks one narrow industry, consumer-goods brands selling through US distributors, where the data is the brand's own but scattered across formats that general tools do not read.
ShopX tried to replace the distributors. Corvera works with them: it reads their reports and drafts the follow-ups to them. Language models now do the plain-English answers that once needed an analyst, which is why a four-person team could take on the work.
Can one person copy Corvera now?
Hard for one person.
The method needs deep knowledge of one industry's data, many data connectors, and enterprise sales to brands with US$20M or more in revenue, according to the company's own target; Corvera raised venture money to do it. The idea is copyable in a smaller niche: pick an industry whose suppliers send messy reports in many formats, match the names, and serve the result to the AI assistant the customer already uses.
Where it is still open: Other industries with distributor or partner reports in many formats: wine and spirits, pharmacy suppliers, building materials, spare parts.
- 1Choose an industry you have worked in, where small companies receive data from many partners in different formats.
- 2Build the name and code matching first, and show the customer one clean list.
- 3Serve the cleaned data over MCP, so the customer's own AI assistant can answer questions.
- 4Publish case studies with usage counts, and sell by demo.
What could go wrong for Corvera?
The risks are the mirror image of the causes above.
What is not verified
No revenue figure was found. The headline, more than a dozen brands signed within five weeks of the first customer, is the CEO's own statement in the seed-round post of 5 May 2026, and no later count or customer list was published. Looked for: a payment-processor profile, a founder interview with figures, a filing. A reader could look for a later interview or a company register entry. The total of US$6.2M raised is the company's own text on its Y Combinator page; the seed post names US$4.2M.
Customer results are not revenue, and come from the company
The CLEAN Cause study reports US$250K of "opportunities surfaced", a 7x return and two analyst salaries saved, and the Wild study a £3.4m to £6.6m sales opportunity. These are customer outcomes and estimates, written by Corvera's CEO with quotations from named customer staff. The January launch post promised profits "up to 40%" higher with no source for the figure.
AI that emails and calls
The footer of both case studies says "Corvera emails the buyer, calls the store, and follows up with the distributor". The home page shows drafts "ready to approve" by a person. No page says whether calls made by the agent tell the store that it is an AI.
Changing product and team
The product was described as back-office agents for orders and inventory in January 2026, as a data context layer in May 2026, and as an AI sales agent in November 2026. Matthew Collins is a named co-founder in the launch post and on the Y Combinator page, but not on the company's team page or in the seed post. The seed post says the money goes to hiring more engineers.
Competition and model providers
The site already has Fathom AI, which scores accounts for field sales reps; Corvera competes in the same task for brands, and with the brokers that the Y Combinator page says take 5% to 12% of a market. Customers ask their questions through Claude, according to both case studies, so the plain-English answers depend on Anthropic and other providers that support MCP.
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.
- C Y Combinator company page: Corvera, the AI sales agent for CPG brandsRead in full, from the pipeline's download of 6 November 2026. Company-written text. Source for the founding year, team size, total raised and target customers.
- C Y Combinator launch post, 21 January 2026: Corvera, AI agents for CPG brandsRead in full. Source for the launch date, the first product description and the founding team.
- C Corvera home page, read 6 November 2026Read in full. Source for the current product and the drafts that a person approves.
- C Corvera news, 5 May 2026: We've Raised $4.2M to Build the Context Layer for AI-Native CPG BrandsRead in full. The CEO's post. Source for the seed round, the brand count and the integrations.
- C Corvera case study, 18 August 2026: CLEAN Cause runs and enhances its daily sales decisions on CorveraRead in full. Written by the CEO. Source for the usage figures and the store matching.
- C Corvera case study, 26 August 2026: Wild's category team runs on CorveraRead in full. Written by the CEO. Source for the second customer's usage.
- C Corvera team page, read 6 November 2026Read in full. Shows three people.
- C Corvera FAQ, read 6 November 2026Read in full. Source for MCP and the plain-English questions.
- C Database of failed startups, local exportLocal 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