Shepherd
AI work memory that indexes Slack, Notion, GitHub and more for agents.
Shepherd is a three-person entrant chasing a company-memory niche that Glean already serves at scale and that Apple is building into macOS itself.
Shepherd is a tool that stores and finds company information so employees can search what they need instead of hunting through emails and files. The problem is that Glean already does this for big companies and has hundreds of millions in revenue, while Microsoft, Google, and Apple are adding similar features to their own products for free. Shepherd might survive if it becomes the go-to memory tool for technical teams writing code (where it could lock in engineers), but it has to prove that works before Apple releases its new memory search in autumn 2026, and before Rowboat (a copycat launched the same time) or Glean's newer versions take over that same space.
Rowboat announces named customers or funding before Shepherd does. Would show the same-batch clone winning the identical niche first.
Apple ships macOS 27 full release (autumn 2026) with system memory search. Would show the native OS absorbing Shepherd's core function for free.
- 3 employees at YC S26 launch
- Product at v0.1.158, early release stage
- Rowboat, same-batch rival, launched July 21, 2026
- No structural moat identified against Glean or platform incumbents bundling equivalent memory features.
- Same-batch clone Rowboat launched an identical product shape one month before Demo Day.
- Apple macOS 27 full release in autumn 2026 could obsolete standalone general-purpose version.
- Zero independently verified revenue, funding, or customer data as of July 2026.
- Heavy name confusion with unrelated Shepherd Insurance, YourShepherd legal, and Korso's Shepherd tool.
| Rival | Direction | Evidence |
|---|---|---|
| Glean | losing to | Glean at ~200M ARR (vendor-claimed, doubled 2025) and shipping agent architecture content through mid-2026 |
| Rowboat | contesting | Same-batch YC S26 launch on 2026-07-21 with near-identical AI coworker with memory positioning |
| Apple (macOS 27 Siri AI) | losing to | System-wide personal-context search and cross-app actions in public beta June-July 2026, full release autumn 2026 |
| Microsoft 365 Copilot / Google Gemini Enterprise | losing to | Bundled AI search and assistant features included free in core suites as of 2026-07 |
| Cerenovus | contesting | Same YC S26 batch targeting overlapping company knowledge graph niche as of 2026-07-25 |
| Notion AI / Slack AI | losing to | Free bundled summarization and drafting features compressing point-solution addressable market, 2026-05 |
- Is Shepherd (askshepherd.ai) profitable in 2026?
- No revenue, ARR, or profitability data has been independently reported for Shepherd as of July 2026. The company is a 3-person, pre-traction YC S26 startup in early release on macOS, with its site showing only illustrative mock UI content, not verified financials.
- Could AI replace Shepherd's product?
- Shepherd's product is itself an AI ingestion and memory layer, so the bigger risk is incumbents like Glean or Microsoft bundling equivalent agentic memory features. Apple's macOS 27 Siri AI, in public beta June-July 2026, ships system-wide personal-context search that overlaps directly with Shepherd's function.
- Who is Shepherd's biggest competitor?
- Glean is Shepherd's biggest competitor, an enterprise AI search incumbent that vendor-cited reporting says roughly doubled ARR to about $200M in 2025. Same-batch YC rival Rowboat launched a near-identical 'AI coworker with memory' product on 2026-07-21.
- Who owns Shepherd (askshepherd.ai)?
- Shepherd is privately held and founder-led, backed by Y Combinator's Summer 2026 batch. Founder/CEO Philip Meng and CPO Ishan Ramrakhiani lead the 3-person San Francisco team; no outside investors beyond YC were disclosed as of 2026-07-25.
- Does Shepherd have any real customers or paying users?
- No independently verified customer or revenue data exists for Shepherd as of 2026-07-25. Its marketing site shows an illustrative mocked sales pipeline example ('Acme, verbal yes, $2k/mo') described as vendor-claimed UI demo content, not confirmed traction.
- How is Shepherd different from Rowboat?
- Both are YC S26 'AI work memory' startups with nearly identical positioning. Rowboat launched publicly on 2026-07-21 with founders from a Coinbase-acquired startup; Shepherd, per its own site, differentiates via deep ingestion of coding tools like Claude Code, Codex, and Cursor, though this wedge is unvalidated.
- 2026-07-25YC directory lists Shepherd as active with 3 employees in San Francisco
- 2026-07-21Same-batch rival Rowboat launches near-identical AI coworker with memory product
- 2026-07-20Apple macOS 27 Golden Gate public beta ships Siri AI with system-wide memory search
- 2026-07-08Unrelated shepherd-agents GitHub project published, noted as separate entity
- 2026-07-01DevRev round-up lists Microsoft 365 Copilot and Gemini Enterprise as bundled alternatives
- 2026-05-27GoSearch round-up lists GoSearch, Guru, Coworker.ai as Glean-alternative competitors
- 2026-05-13Notion AI and Slack AI cited as bundling free summarization compressing point-solution market
- 2026-04-23Vendor source cites Glean valuation at 4.6 billion dollars
- 2026-04-06Vendor source cites Glean valuation at 7.2 billion dollars
- 2026-03-24Unrelated Shepherd Insurance raises 42 million dollar Series B, distinct company
- Shepherd: Give your company the knowledge it deserves
- Shepherd, Know everything happening at your company
- Complete Analysis
- 10 Best Glean Alternatives in 2026 (Ranked) (2026-04-06)
- 9 Best Glean Alternatives for Product Teams (2026) (2026-04-23)
- 7 Core Components of an AI Agent Architecture Explained
- Launch YC: Rowboat, Your AI coworker, with memory (2026-07-21)
- Search Startups funded by Y Combinator (YC) 2026 (2026-07-25)
- macOS 27 Golden Gate Guide (2026-07-20)
- 4 Best Glean Alternatives for AI Enterprise Search in 2026 (2026-05-13)
- From Enterprise Search to Autonomous Action (2026-07-01)
- Insurance tech startup Shepherd raises $42M to underwrite the physical layer of AI (2026-03-24)