dreamteam / competitive intelligence

OverviewDossiersAttio

Threat level 5 · most dangerous AI readiness 4.0/5 Updated Jul 5, 2026

Attio

The flexible, beautiful data-model CRM that added a genuinely real agent. Same ICP as us, fastest ship cadence in the cohort — the default modern choice for our exact buyer.

The bet

Attio is betting that a beautiful, flexible data model is the durable foundation, and AI is a layer you keep adding on top. Founded 2019 by Nicolas Sharp (ex-VC); $116M raised including a $52M Series B led by GV in Aug 2025 — the "$141M" some aggregators cite is unverified Tracxn data, not the company's own figure. The strategy: own the "modern Salesforce replacement" slot for startups, then move upmarket with Call Intelligence, sequences, permissions, and now a real agent (Ask Attio, Feb 2026). They want to be the system teams configure and control — the opposite pole from "the CRM disappears behind agents."

Why they're #1 threat

Same buyer, a shipped agent with a credible technical architecture, and a ~biweekly release cadence aimed squarely at the gaps we'd attack. This is not vaporware — treat the window as quarters, not years.

Snapshot

$116M
total raised · $52M Series B (GV)
4.4/5
G2 · 284 reviews
$690/mo
10-seat Pro (annual)
~biweekly
release cadence
115–177
headcount (conflicting)

Verification gaps (flagged, not asserted): exact customer count & segment mix are undisclosed by Attio and unfound in secondary sources; headcount rests on aggregators (PitchBook 115 vs Tracxn 177); no neutral third-party benchmark of Ask Attio's task accuracy exists.

What's genuinely good

  • Data-model flexibility with consumer-grade UX — custom objects/relationships that feel like Notion, not Salesforce setup. Best-in-class, and the #1 praise theme on G2 ("doesn't feel like a traditional rigid CRM").
  • A real, documented agent architecture — Ask Attio runs across Anthropic, Vertex (Google) and OpenAI with automatic failover, a dynamic per-turn tool registry, and a streaming structured-UI parser. The engineering write-up is credible primary evidence, not a chatbot wrapper.
  • Genuine action-taking — as of Apr 29 2026, Ask Attio executes CRM operations from plain language, not just Q&A.
  • Call Intelligence — joins Zoom/Meet/Teams, transcribes, and extracts MEDDPICC/BANT/CHAMP fields into records (shipped ~Apr 2025, not a 2026 feature despite bundled marketing).
  • Cold-start magic — auto-enrichment from email/calendar sync at onboarding makes the empty-CRM problem feel solved.

What's weak

  • Reporting — the single most-repeated complaint, direct from G2: "The reports are limited. There should be built-in sales metrics for regular tracking." Weak below Enterprise; limited drill-down, no scheduled delivery.
  • Manual upkeep remains — Attio is a system you manage: views, workflows, hygiene. It doesn't run the pipeline for you.
  • Specific gaps users list — no unified inbox, no forms, no automated duplicate detection; one migration account reports deal-stage history lost and activity logs imported inconsistently.
  • Credit-metering opacity — two credit pools (per-seat + workspace), and the cost per AI action is unpublished, so teams can't predict what a workflow will cost before running it; users report burning workspace credits "very quickly" on complex automations.
  • AI depth vs. positioning — multiple reviews note capabilities "don't always match the depth suggested by the platform's positioning".

AI capability lens

The dimension the whole market is fighting on — graded from changelog + engineering docs, not the landing page.

Attio · AI readiness4.0

Shipped & verifiable

  • Ask Attio (Feb 4 2026) — conversational agent: search/update/create records, draft follow-ups, summarize calls/emails via "Universal Context"
  • Ask Attio in Slack (Mar 10) + action-taking (Apr 29) + custom per-workspace context (May 28)
  • Multi-model routing across Anthropic / Vertex / OpenAI with failover; Claude Opus 4.7 added Apr 16, Opus 4.8 + Gemini 3.5 Flash late June
  • Autonomous Workflows (Jun 9) — describe a goal in natural language, AI builds the automation
  • MCP server (Feb 19) — OAuth-scoped read/write to Claude/ChatGPT

AI strengths

  • Multi-model resilience → no single-vendor lock-in
  • Covers the full loop: search → summarize → act
  • Fast cadence tightly coupled to a public changelog
  • Deep call-to-field automation via qualification templates

AI weaknesses

  • Unpredictable/unpublished AI action costs → budget risk
  • Zero neutral accuracy benchmarking exists (all evidence vendor/investor)
  • Thin reporting substrate caps how far "AI insights" can go
  • Agent assists inside the workspace; it doesn't run the pipeline
So what

Attio has the strongest architecture in the cohort but the same structural ceiling as everyone else: an assist layer on a CRM you still operate, feeding a reporting layer users already call weak. Our opening is outcome ownership + predictable pricing — agents that finish jobs end-to-end, with costs a founder can predict before running them.

Customer voice

Praise

"This is the best solution I've seen in the market in quite a while… it can compete with HubSpot and Salesforce when needed, but it's also easy to use right out of the box."G2 review · platform: G2
"Attio has completely changed our business. I have recommended it to many people."G2 review · platform: G2

Complaints

"While it's extremely flexible and visually polished, it can take a lot of time to navigate through all the layers of information."G2 review · platform: G2
"Activity logs, call notes and email threads migrated inconsistently… deal stage-progression history was lost during migration."Independent HubSpot→Attio migration account · platform: Medium

Reddit quotes ("no unified inbox, no forms, weak reporting… returned to old CRM") circulate via review blogs but could not be independently verified this session — direct Reddit access was blocked. Flagged as a gap, not asserted as fact.

Pricing teardown

Per seat + two credit pools. Annual billing is 20% cheaper:

TierAnnual /user/moWhat unlocksSeat credits
Free$0 (≤3 seats)Core CRM100
Plus$29No seat limit500
Pro ★$69Call Intelligence, sequences, advanced permissions1,000
EnterprisecustomUnlimited objects/teams, advanced security2,500

10-seat Pro = ~$690/mo ($8,280/yr) — 3–5× the cheap AI-native challengers (Day, Clarify) but roughly half of HubSpot Pro. Gotcha: Call Intelligence is Pro-only (Plus→Pro is a $29→$69 jump the moment you need it), and workspace credits (shared pool for workflows/AI) are buyable at +5,000 for $85/mo up to +50,000 for $595/mo — with per-action cost unpublished.

Head-to-head (10-seat deal)

✓ We win when

The buyer wants outcomes, not a workspace to configure. Attio still asks the team to build and maintain the system; our agents own jobs (inbound triage, research, deal health) end-to-end with draft-approval. We also win on predictable pricing (vs. their credit opacity) and on phone/social-heavy motions (their "last interaction updates only via email" blind spot).

✕ We lose when

The buyer is a product-minded founder who enjoys configuring a data model, needs custom-object depth today, or wants a proven vendor (thousands of teams, real Series B, big-name investors) over a pre-launch product. Ask Attio + Call Intelligence lets their champion say "we get AI too, with less risk."

Watch

Their 2026 trajectory is to close exactly the gaps we attack. No new changelog entry shipped in the 48h before this snapshot (latest: Jun 29) — a pause, not a slowdown; re-check in 1–2 weeks. The strategic tell to monitor: if Attio ships real reporting depth or published, predictable AI pricing, two of our three wedges narrow at once.

← Full matrix · Next: HubSpot →