The Context Translator — pitch deck
A concept for inclusive EU democratic participation

The Context
Translator.

Democracy that meets every EU citizen where they already are — on TikTok, in their inbox, on LinkedIn, in their letterbox.

Open concept · 2026
Problem · Diversity

Diversity isn't a footnote. It's the point.

Europe is larger, more multilingual and more institutionally varied than any other polity attempting digital citizen participation.

Problem · Diversity
449M
EU citizens

One consultation is addressed to all of them, in one register, at one moment.

24
official languages
27
institutional cultures, in one Union
~240
NUTS-2 regions, each with its own context
~40%
have low or no digital access
16→90+
age span participating
3 in 10
live in rural or remote areas
Problem · Tech interface

Every tech tool comes with an interface.

Whichever one you pick, you have already picked who you can reach.

Problem · Tech interface

Pick an interface — and you have picked your audience with it.

Interface
Web app
Who it reaches
45+ urban desktop fluent reader
Interface
Mobile app
Who it reaches
25–45 urban smartphone-fluent data plan
Interface
Desktop / SaaS
Who it reaches
35+ professional work email policy-curious
Problem · Tech interface

The same engaged demographic survives every interface choice.

For shoes, that's fine. For an EU democracy, it isn't.

Problem · Consequence

How a digital interface narrows the EU audience

Each digital filter strips a layer of the audience the same way.

EU citizens
449M · 100%
Daily internet users
~380M · 85% −15%
Above-basic digital skills (Eurostat)
~260M · 58% −27%
Use eGov services regularly
~160M · 36% −22%
Follow EU policy content
~50M · 11% −25%
Participate in an EU consultation
~5M · 1% −10%
Problem · Consequence
1%
the same people, every time

The same 1% at the bottom of every digital consultation.

Different platform, same demographic — urban, digitally fluent, already engaged. That isn't a bug. It's the design philosophy of the tools.

Solution · one of three
Beat 1

One topic in.

A stakeholder describes what they want to ask EU citizens. Nothing else is required of them.

Stakeholder brief

“We want public input on the EPBD revision — cost, comfort, climate. EU-wide, 6 weeks.”

Elena · policy officer, DG ENER
Solution · two of three

Many registers out.

AI intake and audience analysis — scope, depth, geography, partners — then one adapter per register.

Channel 01
Academic expert email
Dr. Anna Kowalski
TU Delft
Channel 02
TikTok / Reels via youth NGO
Léa, 17
Lyon
Channel 03
LinkedIn long-form
Markus, 41
Sustainability lead, Berlin
Channel 04 · non-digital
Plain-language paper letter
Mária, 72
village in Hungary
Channel 05 · non-digital
European Citizens' Panel
Randomly selected citizens
Solution · three of three · the return path

One synthesis back.

Academic email TikTok / Reels LinkedIn Paper letter European Citizens' Panel SYNTHESIS DASHBOARD Weighted · comparable · traceable

A paper tick-box from Mária lands in the same weighted synthesis as a LinkedIn comment from Markus. The loop closing is the differentiator.

Validation · worked example

What this looks like: the EPBD revision.

DG ENER wants citizen input on the next Energy Performance of Buildings Directive. One brief, nine interfaces — all live in this prototype.

Validation · nine clickable mockups
01 · Stakeholder side
Intake + dashboard
Elena · DG ENER policy officer
Open mockup →
02 · Citizen side
Expert email
Dr. Anna Kowalski · TU Delft
Open mockup →
03 · Citizen side
LinkedIn long-form
Markus, 41 · Sustainability lead, Berlin
Open mockup →
04 · Citizen side
TikTok short
Léa, 17 · Lyon
Open mockup →
05 · Citizen side · low digital access
Paper letter
Mária, 72 · village in Hungary
Open mockup →
06 · Citizen side · deliberative
Panel invitation
Mária again · drawn by lottery
Open mockup →
07 · Partner side
Panel organiser dashboard
Ágnes · Sortition Foundation, lottery coordinator
Open mockup →
08 · Contribution loop
Civic-tech console
Adapter gaps become open-source issues
Open mockup →
09 · Stakeholder side · return path
Synthesis dashboard
Elena, 6 weeks later
Open mockup →
Click the examples to open them in a new tab — continue the slideshow when you're ready.
Validation · what came back
428 314
responses
21%
low-digital-access · vs 4% in 2022

#1 theme across every channel: heating cost is the lived experience of energy policy.

Alternatives
Existing landscape
Decidim · CitizenLab · Consul

Browser-first participation platforms. The citizen has to come to them. Strong tools for those already engaged — invisible to everyone else.

EU Citizens' Panels · ECI

High-quality but single-format, slow, expensive. One channel, one register — by design.

Paid social campaigns

Multi-channel, but outbound only. Optimises for clicks, not for legible democratic input. No return path.

The Context Translator

One topic, every register, both directions.

Multi-channel by default

Five+ channels from one stakeholder brief — orchestrated, not duplicated by hand.

Non-digital is first-class

Paper letters and assemblies are channel adapters too, not a fallback.

Closes the loop

Inbound aggregation across channels, weighted so the quiet voices aren't drowned by the loud ones.

Viability

This is buildable — and it stays alive.

No new technology is invented. The novelty is the orchestration and the inclusion model.

Feasibility
Orchestration of existing parts
  • AI for segmentation — off-the-shelf LLM + EU demographic open data
  • Channels already exist — TikTok, LinkedIn, postal services, NGOs, academic networks
  • NGOs already do this work per channel — we coordinate, we don't replace
  • Synthesis — LLM-assisted summarisation + human rapporteur loop
Long-term sustainability
Open-source consortium per adapter
  • Each channel adapter = an open-source component maintained by an NGO + academic group
  • Funded through EU programmes: CERV, Erasmus+, Digital Europe
  • Orchestration core also open-source — runs on EU infrastructure
Model proven by Decidim · OS Politics · Consul
Implementation

One small core. Many small adapters.

Every channel is a separately-maintained adapter with the same contract:

“Given a brief and an audience segment, generate a context-adapted message and collect a structured response.”

Adding a new register — say, a community radio adapter for a particular region — is a contribution, not a platform overhaul.

Stack: Node/TS orchestration · LLM via API · adapters in any language · all open-source.

CORE Orchestration + Synthesis Academic email LinkedIn TikTok / Reels Paper letter EuropeanCitizens' Panel Community radiofuture
01.5
Section 1.5 · Open network

Not a platform. An open network of channels.

Civic-tech contributors plug new connectors in — and over time, EU-rooted channels can replace the platforms we depend on.

§1.5 · how the network grows

One connector at a time — registry today, federation tomorrow.

Today · registry
Central registry, distributed delivery
REGISTRY radio paper mastodon

Local radio · regional weekly · municipal Mastodon · any civic-tech custom adapter — all plug in via the same contract.

Soon · protocol
ActivityPub-flavoured federation
radio paper civic mstdn

Same content model, no central relay. Channels syndicate to each other; the registry becomes optional discovery.

Long term · sovereign
An EU-rooted social fabric
radio paper mstdn civic EUstack

A federation of EU-rooted channels carrying participation, education and commerce — gradually substituting US-platform dependency.

Engineering rule
The adapter contract is the ActivityPub contract from day one — Actors with inbox/outbox, JSON-LD payloads, HTTP-signed activities, channel-owned canonical URIs. If we ship a bespoke REST API today, the transition becomes re-platforming, not federation. (see research/03-federation-transition.md)
§1.5 · why it compounds

Each new connector shifts another conversation back inside the EU.

A Galician community radio
A Greek-island weekly
A German civic Mastodon
02
Section 2 · Same channels, certified content

From participation to qualification.

The same infrastructure carries partner-university content — issuing micro-credentials per EU Council Recommendation 2022/C 243/02, recognised across Member States.

§2 · a micro-credential, end to end

A credential, end to end — every standard already shipping.

A citizen earns an “Energy Citizenship” micro-credential issued by TU Delft, and it lands on a Hungarian employer's HR system six months later.

1
University
Issues a Verifiable Credential
ELM · Open Badges 3.0 · W3C VC
2
Trust anchor
Signing key attested in the Trusted Issuers Registry
did:ebsi · EBSI TIR · DEQAR-aligned
3
Context Translator
Triggers issuance through the citizen's chosen channel
OID4VCI
4
Citizen
Receives and holds it in the EUDI Wallet
EUDI Wallet · mandate end-2026
5
HR / Employer
Verifies the credential cross-border
OID4VP · Europass DCI
§2 · the architectural claim
Architectural claim

The Context Translator never custodies credentials.

We are the issuance trigger and the discovery surface. The trust chain is the Commission's — we are a conformant issuer client and we plug in. (see research/02-micro-credentials-wireframe.md)

Works in production today

ELM / W3C VC / OB 3.0 issuance · EBSI pre-production VC chain · OID4VCI/VP libraries · DEQAR-listed QA bodies.

Pending EUDI Wallet scale-out

Uniform Member State coverage of Trusted Accreditation Organisations (TAOs) in the EBSI TIR — forced by the end-2026 EUDI Wallet mandate.

03
Section 3 · Self-financing, sovereign

The economic self-sufficiency layer.

Once levels 1 and 2 reach critical mass, EU-SME advertising on the same channels can sustain the network — and begin to repatriate marketing spend that today leaves the Union.

§3 · where the ad spend goes

EU SMEs spend on digital ads — most of it leaves the EU.

European digital advertising is a ~€100 billion per year market. The lion's share is captured by US-headquartered platforms — the working capital of European small-business marketing is, in net, a transfer to a foreign tech economy.

We aren't trying to compete with Google. We're trying to keep €1 of marketing spend inside EU democratic infrastructure for every €99 that already leaves — and let that €1 sustain the channels citizens already trust.
Digital advertising market · EU advertisers · illustrative
Total spent by EU advertisers~€100B
Captured by US-HQ platforms~€85B
Stays inside EU media~€15B
~80–90% lands with non-EU platforms. Order-of-magnitude figures · Bruegel / IAB Europe / industry trade press. Verify before citing.
§3 · the model

The channels that reach EU citizens are the channels that reach EU customers.

Channel · dual use
Same infrastructure, two purposes

A municipal Mastodon that carries an EPBD consultation can carry a regional cooperative's invitation to a solar group buy. A community radio that broadcasts a citizens' assembly call can carry a Galician winemaker's seasonal ad.

  • Contextual, not behavioural — no cross-site tracking required
  • EU-rooted demand-side · EU-rooted supply-side
  • Audience-of-trust priced as such, not as data
Revenue · who gets paid
Revenue back to the channel maintainers

Each adapter is maintained by an NGO, public broadcaster, civic-tech consortium or local outlet. The advertising revenue that the channel attracts is what funds the maintainer that keeps the channel running.

  • Federation core takes a transparent small fee
  • Bulk of revenue stays at the channel — the local maintainer keeps the lights on
  • Crowds out the need for permanent EU grant subsidy
§3 · honest framing
Honest framing

Level 3 is the weakest evidence base in the proposal today.

The economic-sovereignty argument is politically resonant but the academic-policy case is thin. Our current commitment: we will commission research with Bruegel / ECIPE / European DIGITAL SME Alliance before going public on revenue numbers.

We treat this section as a direction of travel, not a decided model. (see research/landscape.md §7)

Democracy at the EU scale only works if it is polyphonic.

An open infrastructure for context-adapted citizen participation.
Open concept · 2026
01 / 28
← → arrows · space · F fullscreen · A sync: off