foundr.companyby Perea
← foundr.team·Soon

foundr.team — market insights

MECE market analysis. Numbers are point-in-time (May 2026) — sources linked so you can re-verify. TAM > SAM > SOM are nested slices, not aspirational forecasts.

TAMTotal addressable

~$13B / yr (2025) enterprise AI agents + copilots; ~$40B agentic AI 2026 midpoint

Proxy

AI agent platforms + workforce automation + agent observability (the three layers foundr.team unifies — orchestration, governance, retro).

Calc

CB Insights $13B enterprise agent revenue 2025 (150% YoY) + Information Matters $40B agentic AI 2026 midpoint (anchored to Anthropic's $30B run-rate). MarketsandMarkets pegs AI Agents at $7.84B (2025) → $52.62B (2030) at 46.3% CAGR.

Sources
SAMServiceable addressable

~$2.4B / yr

Proxy

Solo founders + micro-SMEs running 3–10 MCP-speaking agents — the indie tier Agentforce/Copilot Studio ignore and CrewAI Enterprise overshoots.

Calc

~2M global indie/solopreneur AI buyers (Sintra alone has 40K+ paying in 100 countries, n8n 230K active users half indie, YC W25 22% voice/agent) × ~$100/mo blended ARPU = $2.4B addressable annual spend. Cross-checks: Sintra €12M ARR + Lindy $5.1M + n8n $40M (half indie) underwrite $1–3B SAM.

Sources
SOMServiceable obtainable (3–5 yr)

$24M ARR by year 5 (~1% of SAM)

Proxy

Free quote-tweet viral funnel + 18% Solo conversion + 4% Pro upgrade + small-team seats.

Assumptions
  • 80K free quote-tweet users (3 slots, $0)
  • 18% convert to Solo $19/mo (14.4K × $228/yr = $3.3M)
  • 4% convert to Pro $49/mo (3.2K × $588/yr = $1.9M)
  • 250 small-team seats × $2,400/yr = $0.6M
  • Second cohort growing 3× → $18–24M ARR year 5
Analog precedent

Sintra hit €12M ARR in 14 months on a near-identical solo-founder ICP. Lindy reached $5.1M ARR in 18 months bootstrapped. n8n grew 5× ARR to $40M in 12 months.

Sources

The top 3 incumbents

Who controls the market — and why they can't pivot.

Each incumbent's vulnerabilities tagged by kind: technical, business model, regulatory / channel, cultural.

#1CrewAI

$24.5M raised, Series A Oct 2024, ~$100M valuation, 450M agentic workflows/month, 60% of Fortune 500, 4K signups/week

  • Tech debt

    Python framework-locked; every agent is a Python class. MCP-as-tools is bolted on, not the spine. No native budget caps — cost governance lives outside (Helicone/Langfuse, now both acquired and in turbulence).

  • Business model misalignment

    Enterprise sales motion (homepage is "Request a Demo," not "Sign up"). No indie tier. Pricing gap from open-source free → enterprise "$99k+/yr" leaves solo founders unserved.

  • Regulatory / channel dependency

    Insight Partners–led round = enterprise GTM lock-in. Can't credibly run a viral quote-tweet free tier without channel conflict.

  • Cultural / incentive trap

    "Loved by AI builders, trusted by AI leaders" — they answer to CIOs now, not solo founders. 4K signups/week is monetized as enterprise pipeline.

#2Lindy

$5.1M ARR 2024, $49.9M raised, $49.99–$199.99/mo, 400K professionals

  • Tech debt

    Inbox-and-meetings shaped — built around email/calendar/phone, not a generalised MCP roster. The agent IS Lindy; you can't bring your own Claude Code / Cursor / custom MCP agents under one roof.

  • Business model misalignment

    No free tier — 7-day trial only. $49.99 entry floor excludes "I'm running 3 agents and want a dashboard" indie. "AI executive assistant" frame is single-agent personhood, not roster.

  • Regulatory / channel dependency

    Heavy on Google/Outlook OAuth — inbox vendor risk if Gmail tightens agent access.

  • Cultural / incentive trap

    "The AI executive assistant" — singular. Lindy is a teammate, not a team. Re-positioning to "your AI team" cannibalizes the single-Lindy mental model.

#3Salesforce Agentforce

$540M ARR Q3 FY26 (5.4× in 9 months), 18,500 deals, 9,500 paid, 12% of Salesforce's 150K customer base

  • Tech debt

    Agentforce only reasons over Salesforce Data Cloud — the agent's "world" is CRM. MCP support nascent; agents are first-class inside Salesforce walled garden, second-class everywhere else.

  • Business model misalignment

    Flex Credits start at enterprise-volume floors. No founder is buying Agentforce for $19/mo. Salesforce hired 2,000 sales reps to sell Agentforce — that's not self-serve.

  • Regulatory / channel dependency

    Locked to the Salesforce account model + Data Cloud. Pricing model mid-transition from seat → usage; reps are conflicted.

  • Cultural / incentive trap

    "Agentic enterprise" — Benioff's frame. Solo founders don't have an enterprise. Structurally incapable of shipping $19/mo indie without cannibalizing $120k Data Cloud.

Strategic moves (12 mo)

Ranked by leverage. Top of the list ships first.

Leverage is encoded in position — no fake score. #1 is the highest-leverage move we can make in the next quarter.

  1. 01

    Lead with the MCP-roster wedge, not "AI agents"

    Position foundr.team as "the roster your MCP clients can already see" rather than another agent platform. Every Claude Desktop / Cursor user is one `mcp add` away from your agents — incumbents (CrewAI, Lindy, Sintra) require a UI rewire.

  2. 02

    Make "hard $/wk cap, refused pre-flight" the headline feature

    Not soft alerts. A `402 Payment Required` that fires BEFORE the model call. The $6k Reddit story, the $1,800 Max-subscriber API leak all share the same root cause: post-hoc accounting.

  3. 03

    Ship the weekly retro as a product surface, not a feature

    Sunday email + dashboard tile: what each agent did, what it cost, where it looped, what to delete. Anchor to a named ritual ("Sunday Retro") so it becomes a habit, not a settings page.

  4. 04

    Flat $/mo, never per-seat — turn it into a moat narrative

    Per-seat is collapsing (Bessemer: 21% → 15% in 12 months). Build the pricing page around "your agent fleet costs the same at 3 or 300 agent-runs/day" and call out CrewAI by name.

  5. 05

    Open the roster schema, then host a registry

    Q2-Q3

    Publish a foundr.team agent manifest format and let anyone publish agents to a public registry (think `homebrew tap` for AI teammates). Sintra's 12 closed personas are the foil.

  6. 06

    Outcome-token billing for the Pro tier

    Q3-Q4

    Layer an optional "pay only for completed retros / shipped PRs / closed tickets" SKU on top of flat $49. Mirrors Intercom Fin ($0.99/resolution → $100M ARR in 24 months).

  7. 07

    Ship "import from Claude Code" as the onboarding path

    Auto-discover the MCP servers + skills + agents in a user's `~/.claude` config and reify them as a roster on first login. Zero-typing onboarding is the only way to beat Sintra's "polished demo."

Economic moats

What we can hold — and what we can't.

Honest split. We refuse to call cost-leadership or distribution a moat unless it actually defends.

Real (defensible)

  1. 01

    MCP-native roster surface

    Every roster member is addressable from any MCP client (Claude Desktop, Cursor, Claude Code, future entrants). CrewAI/Lindy/Sintra are closed UIs — they cannot retrofit MCP-native multi-client addressing without rebuilding their orchestrator.

  2. 02

    Hard pre-flight budget caps + atomic reserve-commit pool

    Documented as missing from Helicone, Langfuse, LangSmith, OpenAI/Anthropic dashboards, and every incumbent roster. Implementing correctly (atomic reservation, race-free across N agents) is non-trivial engineering.

  3. 03

    Flat-fee SKU as positioning identity

    Once you brand as "the flat-fee roster," every move to per-seat is a public retreat. CrewAI/Salesforce can't credibly mirror without cannibalising existing per-seat ARR.

Not real (incumbents can match)

  1. 01

    The agent personas themselves

    Sintra, Lindy, Arahi all ship named personas (Cassie, Penn, Maya). Copying a roster of named characters is a weekend's work.

  2. 02

    Budget caps as a feature

    AgentBudget, Magicrails, Circuit Breaker, Agent Fuse, Shekel, baar-core all ship some version. Caps alone are commodity; the integration with roster + retro + flat pricing is what compounds.

  3. 03

    Multi-agent orchestration

    50+ open-source orchestration frameworks already. If you market on it, you lose to whichever framework fits the user's IDE.

Switching costs in our favor

  • Retros accrete: 12 weeks of retro history is a personalised performance record that doesn't export
  • Budget history + per-agent ROI data lives in your DB — incumbents starting from zero have no baseline
  • MCP server URL is wired into every client config — uninstalling means re-onboarding agents in N IDEs

Switching costs against us

  • Agent prompts/skills are markdown files — trivially portable to any other roster
  • MCP itself is an open standard — competitors can match the entry surface in a sprint
  • Anthropic could ship "Claude Roster" natively and erase our entry advantage overnight

Power-user pain

5 unaddressed pains, real voices.

Each pain has ≥3 independent quotes from Reddit / HN / GitHub / X. If an incumbent could fix it, they would have already.

Pain A

Overnight runaway bills with no pre-flight cap

  • “I left a loop running longer than I realized... I forgot about it, went to bed, and woke up $80 lighter. No alerts. No cap. No warning. Just a bill.”

    — godnick, dev.to (Apr 2026)

  • “I accidentally burned ~$6,000 of Claude usage overnight with one command.”

    — Reddit user via DevToolPicks (May 2026)

  • “$313.23 burned in 8.5h on a single retry-stuck item... NO per-invocation cost echo. NO spend cap flag.”

    — jwtrading1000, Claude Code Issue #57719

Why incumbents
can't fix

Anthropic, OpenAI, CrewAI, Lindy, Salesforce all account for spend AFTER the response arrives. Implementing pre-flight reserve-commit requires re-architecting billing as an execution constraint — and incumbents profit from the burn.

Coverage

✅ Shipped foundr.team budget caps are pre-flight, atomic, per-agent, refusing the call entirely when reservation > remaining.

Pain B

Per-seat pricing punishes adding agents

  • “If adding a new team member to the workspace costs $15/month for the AI agent, there's a subtle pressure to restrict access.”

    — Helenmireille, Medium (Mar 2026)

  • “Per-seat pricing for an AI agent is pricing based on the thing that least correlates with cost. Either lazy or deliberately exploitative.”

    — Helenmireille, Medium (Mar 2026)

  • “Pure per-seat fell from 21% to 15% of SaaS companies between 2025 and 2026, per Bessemer's tracking.”

    — Korix, DEV Community (May 2026)

Why incumbents
can't fix

Salesforce ($125-$550/user), CrewAI ($99→$1,000/mo ladder), Lindy ($50/5k credits), Sintra ($47-$97/mo) all anchor pricing to seats or credits. Each is locked in by existing ARR.

Coverage

✅ Shipped Solo $19/mo, Pro $49/mo, both flat regardless of slot count. Free tier (3 slots, quote-tweet) is genuinely usable.

Pain C

Closed roster — can't add the agent you actually need

  • “Sintra ships 12 fixed Brain personas with locked roles. There is no custom agent builder.”

    — Arahi AI alternatives (2026)

  • “What if I don't want a Soshie? What if I want a bot that reads Reddit for complaints about my competitor, makes them into a haiku... Sintra can't do that.”

    — Noca, Emergent vs Sintra (Jan 2026)

  • “Sintra Brains don't carry context between chats — every conversation resets.”

    — Arahi AI alternatives (2026)

Why incumbents
can't fix

Sintra's brand IS the cast (Cassie/Penn/Soshie/Emmie). Opening the roster dilutes the marketing wedge and forces a builder UI that contradicts "hire, don't build."

Coverage

✅ Shipped foundr.team roster is open: any MCP-conformant agent slots in. Pro tier removes the slot cap entirely.

Pain D

Credit anxiety blocks experimentation

  • “I burned through those credits incredibly fast... I was so worried about running out that I couldn't properly test everything I wanted to build.”

    — Annika Helendi, Lindy review (Jul 2025)

  • “I found myself avoiding experimenting or having casual conversations with my AI agents because each interaction costs credits.”

    — Annika Helendi, Lindy review (Jul 2025)

  • “Lindy Builder got stuck and consumed over 1,000 credits. I had to delete the task, not sure if it even stopped consuming credits.”

    — Nitai B., Lindy Community (Aug 2025)

Why incumbents
can't fix

Lindy, Sintra, CrewAI all use consumption-based credits because LLM costs scale with usage. Removing credits requires absorbing token cost into a flat fee — which their unit economics make unprofitable.

Coverage

✅ Shipped Flat $/mo + hard per-agent budget cap means experimentation is bounded by a known weekly number, not eroded by silent credit drain.

Pain E

No weekly retro / per-agent ROI surface

  • “Production agents fail silently. The user says "it didn't work" and you have nothing to look at. No tool call log. No decision trace. No cost record. Just a black box.”

    — Mukunda Katta, dev.to (May 2026)

  • “Once I started spending fifteen minutes every Friday reviewing my Antigravity sessions and adding a line or two to AGENTS.md, the agent's behaviour visibly settled down.”

    — Masaki Hirokawa, Antigravity Lab (Apr 2026)

  • “Six weeks of production traffic had gone through the system with no traces, no per-step latency tracking, no cost attribution.”

    — Sapota, dev.to (May 2026)

Why incumbents
can't fix

Helicone, Langfuse, LangSmith all stop at traces — they're observability primitives, not retro rituals. Productising the retro requires opinions about cadence and format that observability vendors refuse to take.

Coverage

✅ Shipped foundr.team Sunday retro is a built-in product surface: per-agent cost, loop detections, top tool calls, kill recommendations, editable skill diff.

Synthesis

Where SAM × incumbent vulnerability × unaddressed pain converges.

A wedge counts only when all three columns align. Status = what we've actually shipped against it.

WedgeSAM segmentIncumbent vulnPain solvedStatus
MCP-first roster (no UI rewire)Solo founders already using Claude Desktop / Cursor / Claude CodeCrewAI, Lindy, Sintra all force you into their UI; Anthropic ships no roster primitiveI run 6-8 agents in different terminal panes; I want one roster, not 6 windows✅ Shipped
Hard pre-flight $/wk cap per agentDevs burned by overnight loops ($47, $80, $313, $400, $1,800, $6,000 documented stories)Anthropic dashboard updates "with a delay of several days"; no `--max-cost` flag in `claude -p`I want my AI bill to be a known number on Sunday✅ Shipped
Flat $/mo, agents as infra not seatsSolo founders running 4-23 agents across 3-15 projectsPer-seat fell 21%→15% in 12mo; CrewAI tier ladder $99→$12kAdding an agent shouldn't be a budget conversation✅ Shipped
Weekly retro (per-agent perf, ROI, kills)Multi-agent operators ("14 agents daily", "23-agent system")Langfuse/Helicone/LangSmith give traces, not retros; the "iterate on AGENTS.md weekly" pattern is hand-rolledWhich of my agents earned their keep this week, which to delete✅ Shipped
Free quote-tweet onboarding (3 slots)First-time agent-curious solo foundersLindy free plan "basically useless" (premium actions gated); Sintra paywalls everythingI want to try a roster without giving you a card✅ Shipped
Open agent registry (BYO personas)Builders who reject Sintra's closed castSintra: 12 fixed Brains, no custom builder; Marblism: 6 fixedLet me bring my own agent and let others install mine⚠️ Partial
Outcome-based billing tier (per shipped PR / retro / ticket)Operators who can attribute outcomesIntercom Fin proved the model ($1M→$100M ARR / 24mo); CrewAI/Lindy/Sintra all consumption-basedPay vendor when work shipped, not when tokens burned⚠️ Partial
Cross-IDE handoff (vibe-switch style)6-8 parallel-pane orchestratorsNone of CrewAI/Lindy/Sintra integrate with terminal-native CLIsHand off context from Claude Code agent to Codex agent without copy-paste❌ Gap

Capture strategy

Where foundr.team actually wins.

Each angle ties SOM capture to a specific incumbent vulnerability above.

See how we sell into that gap.

The market thesis lives here. The pricing, MCP surface, and feature list live on the features page.

Full features & pricing← Back to foundr.team