folk

The bet

folk is betting that relationships, not pipelines, are the primitive a small, network-driven team actually operates on — so the product optimizes for effortless capture (LinkedIn/Gmail extension), ambient enrichment, and a beautiful, zero-tutorial setup rather than sales-process rigor. Founded April 2020 in Paris, privately held, shipping weekly. Funding is a genuine open question: folk's own blog cites a $4.5M seed led by Accel, Tracxn corroborates $3.3M from the same 2021 round, but PitchBook lists $12.3M total raised — no Series A was found in any source checked, so the higher figure is unconfirmed. The strategy: own the fast-onboarding, LinkedIn-capture wedge for 1–25-seat relationship-driven teams, and ship AI assistants at every price tier rather than gating them as an Enterprise upsell.

Snapshot

Review-site scorecard click a card → source

Consistent 4.5–4.8 rating across every review site checked — the disagreement is on volume and funding, not on product satisfaction. Strong, repeated praise for onboarding speed and LinkedIn capture; strong, repeated complaints about mobile, automation depth, and dedupe.

Verification gaps (flagged, not asserted): funding conflict — folk ($4.5M) and Tracxn ($3.3M) agree with each other but not with PitchBook's $12.3M; no Series A found anywhere. G2 review count itself is inconsistent across snapshots — seen as 280, 304, 320 and 330, likely a mix of stale caches and genuine growth, not a data-quality problem with the rating.

What's genuinely good

• Fastest onboarding in the cohort — reviewers report diving in and using folk productively with zero tutorials; "I was able to dive right in and start using folk without watching any tutorials".
• folkX capture extension — one-click contact add from LinkedIn and Gmail, repeatedly named the standout feature across Capterra reviews.
• AI shipped at every tier, not gated — all four assistants (Follow-up, Recap, Research, Workflow) are included on Standard, Premium and Enterprise alike, differentiated only by monthly credit allocation.
• Broad ambient data ingestion — LinkedIn, WhatsApp, email, calendar and Fireflies meeting data all feed the assistants, genuinely differentiated vs. legacy CRMs.
• Responsive support — "they actually did the work for me instead of giving me a to-do list".

What's weak

• Scaling ceiling around ~10 people — reviewers and secondary sources converge that folk works beautifully solo or in a small team, then strains once headcount grows past roughly ten.
• No native mobile app — the single most-repeated complaint across G2 and Capterra: "Lacking an app. Not a fan of using a web browser all the time," corroborated across 26+ G2 mentions per aggregator analysis.
• Thin automation and pipeline logic — "there are still a lot of features missing," with reviewers specifically flagging campaign management and deals-pipeline gaps at lower tiers; no trigger-based sequences beyond email, no stage-fired actions.
• Manual dedupe — "we have to check for every lead if it's a duplicate or not manually," notably on Sales Navigator imports; no automated duplicate detection.
• Apparent changelog stall — no dated entries found between end-April and the July 2026 research date, plausibly a consolidation pause (the magicField→textField migration) rather than a genuine slowdown — worth a spot-check before any competitive claim ships.

AI capability lens

Four named, shipped assistants with monthly-cadence iteration — real, not vaporware — but every one is explicitly framed as suggest/summarize/draft. No autonomous multi-step execution or agentic layer disclosed.

Shipped & verifiable

• Follow-up Assistant — scans email/WhatsApp to detect inactive threads, suggests a personalized follow-up (human sends)
• Recap Assistant — AI summary of a relationship across email, meetings, notes, WhatsApp and LinkedIn activity
• Research Assistant — enriches company profiles and generates research notes
• Workflow Assistant — trigger-based automated email outreach on record create / field update
• AI Fields / Auto-fill AI — autonomous field population, cleaning and categorization from ambient data, shipped Mar 17 and expanded Apr 29 2026

Customer voice

Praise

"Folk is such a breath of fresh air – it's like if Notion and a CRM had a beautiful, productive baby."Faisal Karkoh, Product Hunt · platform: Product Hunt

"The company associated to contacts is the ultimate game changer here. It's something that HubSpot struggles with."Matteo Luigiodaro, Product Hunt · platform: Product Hunt

Complaints

"Lacking an app. Not a fan of using a web browser all the time."Capterra review · platform: Capterra

"We have to check for every lead if it's a duplicate or not manually."Capterra review · platform: Capterra

Secondary aggregators quote Reddit sentiment — folk's Chrome extension as "noticeably faster" than competitors, and that folk "seems to get expensive quite quickly" past Premium — plus a paraphrase that agencies churn "once they grew past ten people." None of this could be independently verified this session: reddit.com was unreachable via direct fetch and site-restricted searches returned no indexed results. Flagged as a gap, not asserted as fact.

Pricing teardown

Per seat, three tiers, all include the full AI-assistant suite (differentiated by monthly credit allocation, not feature gating). No free tier; 14-day trial defaults to Premium:

10-seat Premium = $5,760/yr (10 × $48 × 12) — the tier most Dreamteam-comparable buyers would actually need for deals/pipeline/API. Standard-tier equivalent is $2,880/yr but lacks a real pipeline. Gotcha: several third-party aggregators still quote $20/$40/$80 instead of $30/$60/$100 monthly list — those figures are stale; folk raised list prices at some point in 2025/2026, and the direct fetch of folk.app/pricing is the trustworthy figure here.

Head-to-head (10-seat deal)

Watch

The changelog gap (no dated entries found May–July 2026) is the live tell: it could mean a genuine slowdown, or that folk is consolidating before a bigger release — the magicField→textField migration in March suggests folk is generalizing its AI-field mechanism rather than standing still. Re-check folk.app/changelog directly before any competitive claim goes into a deck; a stale re-check here would be worse than not checking at all.