Dailybot (YC Summer 2021): teardown vs XMarts

In one line: Dailybot is a chat tool for team check-ins and async standups that, since April 2026, also shows what each team's AI agents are doing and where they are stuck [site] [blog]. Threat: 3 for W2 only, because it sells the same "see what your agents did" promise to the same engineering leader, it is live in Slack and Teams, and it names large customers [site]. For the full loop (all three layers) the threat is low: it watches work, it does not do it [docs].

1. Fact file

2. Their proposition, reverse engineered

3. Positioning

• Category they claim: "coordination layer for people-agent teams" [blog]. Some pages say "orchestration", others "coordination" (inference from search results, wording is still moving).
• Category a YC partner would put them in: async standup tool adding agent observability (inference).
• What they want to be known for: the place where the manager sees all work, human and agent.
• Where they sit on the map: layers L1/L2/L3 = 1/2/1, buyer = sme, business (with developer-style tools: CLI, API).

4. Three layers vs ours

5. Same and different

6. Wedge overlap

7. Whitespace (where they stop)

• Layer gap: they watch and route; the agents and the checks of finished work belong to someone else [docs]. To own it we need a customer workflow where our agents do the work and a different checker confirms it landed, running for 4+ weeks.
• Number gap: their proof counts teams, users and logos, not money or time saved [site]. To own it we need one before/after number for a team running agents (for example hours of agent work that actually merged, and cost per merged change).
• Buyer gap: none found. They already serve small engineering teams, which is our target [site].
• Price gap: none. At $2.40 to $5 a user they are cheap [site]. A new seller cannot win on price.
• Speed gap: none. Free sign-up and install in chat [site].

8. For the YC application

• Proven today: Marton's logs show his system has run about 5,600 agent jobs, with a heartbeat that reports what is stuck, what silently stopped and what it costs, and about 160 failure records that each added an automatic check. Dailybot's agents report their own status through an API [docs].
• Draft answer to "how are you different from Dailybot?": DRAFT. "Dailybot shows you what your agents say they did. Our system checks what actually landed and what it cost, because we run 24/7 agents ourselves and had to." Founders say it their own way.
• Future claim (needs proof): "Teams trust their agents' work because we check it independently and show the cost per finished change." Needs one outside team running agents for 4 weeks with a before/after number (for example share of agent jobs that merged).
• What we understand that they don't: an agent saying "done" is not the same as work landing; the real failures are silent and cost money. proven on our own fleet (Marton's logs show 42% of jobs ended wont-merge and about 15% reached main); needs proof on any customer.
• Copyability (M3): easy. They already ship fleet health, heartbeat alerts and blocker routing [docs], and Dailybot 3 is their whole new story [blog]. Adding cost tracking or a "did it merge?" check through GitHub is weeks of work and fits their buyer (inference). Not hard, so M3 can not score above 3 against Dailybot.
• Why now (M4): None found in this teardown. Dailybot 3 (April 2026) shows the agent-visibility need is real, but it helps them, not us. Do not use "teams now run many agents": that is general AI progress.
• Do not claim: "no one shows what agents do" (Dailybot does). "Our observability is better" (no outside user has run it). "Agents report their own status unreliably" as a fact about them (inference only: no source shows how often it fails). Any cost or merge number as a customer result.

9. Ideas worth taking

1. Credit agent work to the human who delegated it [docs]. It keeps the human the hero and answers "who is responsible?".
2. Ship an install line the agent can read (Agent: install .../skill.md) [site]. The agent installs the tool itself.
3. "Less dashboard, more briefing" [blog]: one short summary instead of another screen to check.

10. Unknowns

• Real size today: the YC page says 11, careers says 15, their stated customer counts are marketing claims. Look at the "Connections" or "Customers" pages and case studies for named, dated results.
• Funding: only third-party trackers show about $500k (CB Insights, Extruct), and Dealroom shows none. Check YC's deal terms (standard $125k) and any later filing.
• Founding year: 2017 (About page), 2021 (YC), early 2020 (2021 article). Not important for the pitch.
• UK or Europe presence: none seen. 2021 article lists customers in Australia, Japan, Indonesia, Mexico, Brazil and Colombia [press].
• Whether they track agent cost or whether "done" is checked against a repo. Read the Advanced-plan agent analytics docs or try the free tier.
• Revenue and growth since 2021: nothing public. A 2021 article expected about $380k that year (they claim).

Sources

• dailybot.com home [site]
• Pricing [site]
• Enterprise [site]
• About [site]
• Careers [jobs]
• YC company page [yc]
• Dailybot 3: coordination layer for people-agent teams, 16 Apr 2026 [blog]
• Dailybot blog index [blog]
• Changelog, Jul 2025 to Jul 2026 [docs]
• Agents (academy) [docs]
• Dailybot 3 launch changelog, 7 Apr 2026 [docs]
• Forbes Colombia: el bot que desarrollaron en Pereira, 22 Oct 2021 [press]
• CB Insights: Dailybot financials and Extruct funding page, seen in search results only, not opened [press]