# Converteo vs DEPT

> Comparing AI Development options: Paris vs Amsterdam. Side by side: published rates, team size, service and industry coverage, and the curator's assessment of both — pairings are never sold.

Canonical page: https://www.aiagencies.eu/compare/converteo-vs-dept
Generated: 2026-09-20 · Curated by Gabor Kiss · No paid placements — inclusion, score, and rank are never sold.

## Facts
| | Converteo | DEPT |
|---|---|---|
| Editorial score | 70.8/100 | 64/100 |
| Hourly rate | not published | €130-170/hr (directory-listed) |
| Team size | 250+ | 1000-9999 |
| Location | Paris, France | Amsterdam, Netherlands |
| Founded | — | 2016 |
| Last reviewed | 2026-08-05 | 2026-08-06 |
| Services | AI Consulting, AI Agents, AI Development, AI Marketing | AI Development, AI Marketing, Generative AI |
| Industries | E-commerce, Manufacturing | SaaS & B2B |

## Our take on Converteo
Reviewed 2026-08-05 by Gabor Kiss against published evidence.

Sharlie, the voice agent behind Orange's Sosh brand, is the most completely documented build in this register, and the detail is the point.

Converteo committed to a voice-to-voice architecture in April 2024, before the model reached general availability, specifically to avoid a later migration off a speech-to-text pipeline—a decision that bought roughly one second of latency against three. Around thirty specialized agents hand off between themselves, escalate to a human, and act on Orange's catalog, billing and account systems rather than only answering. Validation was modeled on how call-center advisors are trained. Virtual clients replay conversations at volume against an LLM-as-a-judge scoring price accuracy and brand posture, and an AI-specific bug bounty ran with around forty white hackers before launch. It went to production on 16 March 2026 with a mechanism to stop it.

The rest of the library is a decade of measurement and pricing work with named clients: Carrefour, Leroy Merlin, L'Oréal, ENGIE, BUT, Picard, Arkema.

The caveat is that this depth is exceptional rather than typical—most of the other cases describe data and marketing engineering, not AI.

### Key strengths
- One engagement documented end to end, including the architecture decision, its latency cost, the evaluation rig and the kill switch—no other agency here publishes at that level
- Evaluation as a discipline rather than an afterthought: virtual clients, an LLM-as-a-judge scoring price accuracy and brand posture, and an AI bug bounty before exposure to customers
- Ten-plus named clients across retail, energy and industry, with the work described case by case rather than listed as logos

### Good to know
- Outside the voice-agent work, most published cases are data and marketing engineering—confirm which practice your engagement lands in
- No headcount, founding year or rates published; scale and price both have to be established in conversation

## Our take on DEPT
Reviewed 2026-08-06 by Gabor Kiss against published evidence.

Signs is the most visible AI work here: a sign-language learning platform using computer vision to check the learner's signing, which reached more than 20 million people in its first week and recorded 20,000 signs learned in ten days. It is a collaboration with NVIDIA and the American Society for Deaf Children rather than a commissioned client project, which is worth knowing when you read it as evidence.

The commissioned AI work is more ordinary and more useful as a guide: a return-value prediction model for the fashion retailer Omoda that fed its marketing hub, Clair for CaryHealth, a reference app its own case study credits with 90% faster clinical search, and nine CGI avatars generated for Inter's INTERISTA fan platform, published as 13.5 times faster and 92% cheaper than the comparable manual workflow.

Size is the real variable: DEPT operates across many markets, so who staffs your project matters more here than at a twenty-person firm.

The weaker fit is an enterprise wanting a production AI system governed and operated long-term; this bench is strongest where AI meets the audience.

### Key strengths
- Named clients with figures attached across three sectors: retail returns, clinical search, sports fan engagement
- Computer vision applied to genuinely hard interaction problems, not generative demos
- International delivery capacity, so multi-market rollouts do not need a second agency

### Good to know
- Signs is collaborative work with NVIDIA, not a client engagement—judge the commissioned cases separately
- Published AI work is weighted toward audience-facing products—ask for an operated system with an SLA if that is what you need
- At this scale, insist on knowing the named team, not the agency's aggregate credentials

## Which to choose
Choose Converteo for:
- closer collaboration at a smaller team scale
- disclosed specialization in E-commerce / Manufacturing
- broader service offering — also covers AI Consulting and AI Agents
- a higher overall editorial score

Choose DEPT for:
- larger team capacity for multi-stream or enterprise-scale programs
- disclosed specialization in SaaS & B2B
- broader service offering — also covers Generative AI

## Questions buyers ask

### Which scores higher overall, Converteo or DEPT?
Converteo scores 70.8/100, 6.8 points higher than DEPT at 64/100. Both portfolio quality and business credibility feed into the score — see our methodology for the full breakdown.

### When should I consider both Converteo and DEPT?
Consider running parallel discovery briefs with Converteo and DEPT if your project spans multiple workstreams, you want competitive proposals to compare scope and approach, or you're undecided between the specialization angles each brings (see the Best for cards above). Most engagements ultimately go with one — but the parallel-brief phase is a low-cost way to validate fit.

---
Source: AIAgencies.eu — the curated register of European AI agencies. Rate provenance: "agency-confirmed" = disclosed via submission or claimed profile; "directory-listed" = Clutch or two agreeing B2B directories (lone unverified sources are never written). Full method: https://www.aiagencies.eu/methodology · Rate index: https://www.aiagencies.eu/rates