State of vertical AI in France 2026.

The vertical vs generic divide.

The French B2B conversational AI market polarized in 2025-2026 around two opposing philosophies. On one side, generic tools (Tidio + GPT, Intercom Fin, Crisp + AI, custom ChatGPT) that bet on the flexibility of a well-prompted LLM. On the other, verticalized AI agents (Orkelia and a few rare competitors) that bet on domain depth.

2026 marks a turning point: the first 12-month retention studies show that generic tools lose 60% of their B2B customers in the first year — wear caused by contextual slip-ups — whereas verticalized solutions maintain retention above 85% across the same cohorts.

Size and growth France 2026.

The French B2B conversational AI market represents ~€280M in 2026, growing ~45% YoY. Of that total, verticalized AI still weighs a minority share (~€60-80M) but with ~85% YoY growth — the inflection is underway.

For comparison: the B2C conversational AI market in France weighs ~€520M in 2026 (dominated by self-serve and plug-and-play tools). The B2B segment remains 2× smaller in value but 3× more profitable per customer in LTV.

Clinics · the underexploited captive market.

France has ~58,000 self-employed doctors, ~24,000 dentists, ~52,000 physiotherapists, and ~6,000 private clinics. The healthcare market is regulated (healthcare GDPR, HDS), which de facto excludes generic tools that don't cover compliance.

Across 30 aesthetic practices observed in 2025-2026, the finding is recurring: 47% of patient requests outside opening hours (Instagram DMs, WhatsApp, website form) never receive a timely reply. The average net loss per practice: ~€5,400/month.

Main players: Doctolib (booking + calendar, €109/month/practitioner), Maiia (Cegedim, calendar), Orkelia Clinique (conversational AI agent, €497-697/month). See the detailed comparison.

Structural HR vs operational brain.

France has ~175,000 restaurants. The market is served by HR tools (Combo ~€89/month/venue, Skello, PayFit) that handle structural data entry but automate neither urgent shift replacements, nor Instagram reservations, nor the daily team brief.

Average measured gain: ~6h/week for the manager on scheduling + replacements + briefing. With a manager hourly value of €50/h, that represents €1,100-1,600/month in hidden cost.

Orkelia Restauration enters pilot in Q3 2026 with 5 spots open. Pricing from €297/month (1 restaurant) to €697/month (2-5 restaurants). See vs Combo.

The illusion of self-serve outbound.

The B2B outbound market polarizes between self-serve tools (Lemlist €59/user, Apollo $99/user, Smartlead $39/user) and human SDRs (~€52K/year all-in).

The trap: Lemlist costs €708/year on the surface but requires ~30-45h/month of operations (templates, warmup, monitoring, reply classification, suppression list). That's €36-54K/year of founder time captured. The real cost of Lemlist in self-serve mode exceeds that of a human SDR.

Emergence of the "managed service" model in 2026: €500-1,500/month for 2-5K personalized emails/month, all-in. Vs Lemlist + SDR at €35-55K/year. Savings of ×7 to ×9 for the same result. See the detailed comparison.

Quotes + tracking so archaic.

~38,000 road transport companies in France. The market suffers from extreme fragmentation (most <10 vehicles), manual processes (Excel quotes, phone tracking), and a digital lag of about 5-7 years vs healthcare.

Clear opportunity: automate the quote cycle (request intake → rate grid → send in 30 sec) + real-time delivery tracking + WhatsApp driver coordination. Orkelia Transport in pilot in Q4 2026 (3 spots).

Bloomberg vs the informed investor.

Bloomberg Terminal costs ~25,000 USD/year for institutions managing >€100M. 90% of the market — informed retail investors, mid-size family offices of €30M-500M, private wealth managers — has no access to this data and makes do with Yahoo Finance + Trading View + Twitter Finance.

The Predator app (Orkelia Trading) addresses this gap: cockpit of 98 assets, adaptive scoring across 5 regimes, smart money SEC EDGAR Form 4, Polymarket signals, daily AI brief. B2C pricing €14.99/month or €120/year. B2B/Family Office pricing €1,497/month (vs Bloomberg ×16). See the comparison.

Shopify vs brands with identity.

Shopify dominates the global mass market. But for brands with specific business needs (technical catalogs, complex logistics, niche markets) that want more than a standard catalog, the generic tool isn't enough.

Emergence in 2026 of the agent-driven e-commerce model: an AI agent that watches over the catalog, orders/stock and customer support, instead of a simple back-office. Live case: Padel Congo, a new padel gear store delivering to Congo, the DRC and France, payment in euros via Stripe and Madrid→Congo logistics in grouped shipments. Orkelia Shop builds these storefronts from €1,800 excl. VAT setup. See vs Shopify.

What will happen by the end of 2027.

Based on 2024-2026 observations, my predictions for the vertical AI market in France:

1. Rapid consolidation in healthcare: 2-3 verticalized players will emerge as leaders (probably Orkelia + 1-2 others), capturing 60-70% of the private medical market by the end of 2027.
2. Collapse of self-serve B2B outbound: Lemlist, Apollo, Smartlead will lose market share to managed services (Orkelia Commercial and emulators).
3. Bloomberg challenged on mid-size family offices: Predator + 1-2 competitors will erode this segment from below (€30-300M in assets).
4. Vertical market in restaurants: emergence of a dominant player on chains (5-50 restaurants) operating in synergy with Combo/Skello. Orkelia Restauration pilot positioned.
5. Takeoff of the Tailor model: lawyers, real estate agencies, premium plumbers, coaches will start equipping themselves with dedicated AI agents (€15-50K Build + monthly Run).
6. EU AI Act regulation enters into force progressively over 2026-2027. Marked competitive advantage for players hosted in Europe with native compliance.