Within (YC Summer 2018): teardown vs XMarts

In one line: Within watches how staff at very big companies really work, builds a "Work Brain" from it, and ranks which jobs to hand to AI [site] [yc]. Threat: 1, because it has the same idea but sells to a different buyer (Fortune 500, through Accenture) [site] [press]. Expect a YC partner to name it anyway, because its YC line uses the words "Company Brain" [yc].

1. Fact file

2. Their proposition, reverse engineered

3. Positioning

• Category they claim: Work Brain / Company Brain for AI transformation [site] [yc].
• Category a YC partner would put them in: process discovery for enterprises, with an agent upsell (inference).
• What they want to be known for: the context layer every enterprise needs before AI agents can work.
• Where they sit on the map: L1 3, L2 1, L3 1. Buyer: enterprise, business.

4. Three layers vs ours

5. Same and different

6. Wedge overlap

7. Whitespace (where they stop)

• Layer gap: they map and recommend; the work is done by someone else [site] [press]. To own it we need one customer workflow run by our agents for 4+ weeks.
• Buyer gap: no product or case study for firms under about 1,000 people (inference from customers and price) [site]. To own it we need SME or scale-up buyers saying they have the pain.
• Number gap: no money result in public case studies [site]. To own it we need one before/after number a finance person would sign.
• Price gap: six-figure contracts (inference) and consultants [jobs]. To own it we need a price a 50-person firm can pay, tied to the result.
• Speed gap: 20 to 30 days to first value [jobs]. To own it we need first value in under 14 days.

8. For the YC application

• Proven today: Marton's logs show his system already runs agents that do and check work: about 5,600 jobs, each checked by a different AI, with about 160 failure records that each added an automatic check. Within's public results count processes mapped, not work done [site].
• Draft answer to "how are you different from Within?": DRAFT. "Within maps how big companies work. Our system already runs agents that do the work and check it, every day, on our own business." Founders say it their own way.
• Future claim (needs proof): "We do the work for smaller firms and own one number." Needs one customer workflow run by our agents for 4+ weeks, with a before/after number.
• What we understand that they don't: a company brain learns faster from checked results than from watching screens. needs proof on a customer; Marton's logs show it on his own work (offer feedback lines fell from 44 to 18 as rules were added).
• Copyability (M3): medium. They already generate skills and run light agents [site], so the tech is within reach. Going to small firms would fight their enterprise sales model and the Accenture channel [jobs] [press] (inference). Not hard, so M3 can not score above 3 against Within alone.
• Why now (M4): None found in this teardown. Do not use "agents can now do the work": that is general AI progress.
• Do not claim: "we do the work and own the number" (no customer run yet). "Company brain" as our headline (their YC line uses it).

9. Ideas worth taking

1. Tell staff what they do well while you learn from them [yc]. It makes capture feel like help.
2. Hand back a ranked list with hours on each line. Easy for a buyer to act on.
3. Proof as a count and a time. Use the same shape once we have a real number.

10. Unknowns

• Whether they sell in the UK or Europe. Look at their careers and customer pages for UK offices.
• Real prices. Ask a buyer who has seen a quote.
• Whether Companion stays installed after onboarding (continuous) or is used in sprints only.

Sources

• within.ai
• within.ai/platform
• YC company page
• Klarity is now Within
• The Within Manifesto
• Accenture invests in Within, Sep 2026
• Klarity $70M Series B, Jun 2024
• Within job ad: Senior AI Value Delivery Strategy Lead [jobs]
• Capturing current state, user docs [docs]
• Deeper sources: ../../research/within-study-2026-10-19.md