ChatGPT remains the most widely recognised AI assistant among Dutch consumers, but Google’s Gemini is gaining traction fastest — a shift Telecompaper’s Dutch Consumer AI Monitor (Q4 2025) frames as the latest sign that the generative‑AI battleground is moving from a single breakout product into a multi‑assistant ecosystem where distribution and product integration increasingly matter.
AI assistants went from curiosity to routine within a few years, and the Netherlands — a highly connected market with strong smartphone and cloud adoption — is showing the same pattern: high awareness of a dominant name (ChatGPT) alongside rapid growth in recognition and trial for second‑tier assistants, notably Google’s Gemini and Microsoft’s Copilot. Telecompaper’s summary notes headline awareness figures (ChatGPT ~65%, Gemini ~32%, Copilot ~28%) and calls out that awareness is strongly correlated with age, education and digital engagement; older or lower‑educated respondents report markedly lower recognition. That local snapshot mirrors broader telemetry and market‑tracking seen across 2025: public web traffic and app telemetry still show ChatGPT as the largest single destination for public generative‑AI use, while Gemini’s share and momentum have accelerated rapidly in mid‑to‑late 2025 as Google surfaces Gemini across search, Workspace, Android and mobile app surfaces. Independent trackers and industry reporting document this directional shift even as absolute numbers differ between measurement methods.
Source: Telecompaper Telecompaper
Background
AI assistants went from curiosity to routine within a few years, and the Netherlands — a highly connected market with strong smartphone and cloud adoption — is showing the same pattern: high awareness of a dominant name (ChatGPT) alongside rapid growth in recognition and trial for second‑tier assistants, notably Google’s Gemini and Microsoft’s Copilot. Telecompaper’s summary notes headline awareness figures (ChatGPT ~65%, Gemini ~32%, Copilot ~28%) and calls out that awareness is strongly correlated with age, education and digital engagement; older or lower‑educated respondents report markedly lower recognition. That local snapshot mirrors broader telemetry and market‑tracking seen across 2025: public web traffic and app telemetry still show ChatGPT as the largest single destination for public generative‑AI use, while Gemini’s share and momentum have accelerated rapidly in mid‑to‑late 2025 as Google surfaces Gemini across search, Workspace, Android and mobile app surfaces. Independent trackers and industry reporting document this directional shift even as absolute numbers differ between measurement methods. What Telecompaper’s findings say — and what they don’t
The headline: awareness versus active use
Telecompaper’s monitor reports brand awareness figures that matter because awareness is the upstream metric for trial, retention and eventual habit formation. Awareness data are useful for understanding which assistant names reach the mainstream psyche, but they are not equivalent to active usage, market share, or time‑on‑task.- Awareness (brand recognition) is an early indicator: if 65% of respondents know ChatGPT, that’s a strong awareness lead.
- Active usage (daily/weekly users, sessions, prompt volumes) and embedded use (in‑app assistants inside Gmail, Docs, Office) can tell a very different story about where attention and productivity are actually happening. Web‑traffic trackers and app telemetry often capture different slices of that picture.
Demographic skews matter
Telecompaper flags an important nuance: familiarity is unevenly distributed. Younger, higher‑educated and digitally engaged respondents show higher familiarity, while older or lower‑educated groups lag behind. This matters because it predicts how quickly different user cohorts will trial, pay for premium tiers, and accept in‑app copilots as part of daily workflows.- The practical effect: orgs and consumer marketers will see faster adoption among knowledge workers, students and early adopters; mass‑market penetration will require frictionless integration into familiar surfaces (mail, search, phone assistants) rather than expecting broad voluntary migration to stand‑alone apps.
Methodology caveat (why you should be cautious)
Surveys and trackers differ in how they define and measure awareness. Telecompaper’s monitor provides a useful snapshot but does not, in the published summary, show full sampling details or the exact questionnaire text. That makes it prudent to treat the reported percentages as directional and valuable — but not definitive for cross‑market comparisons without seeing full methodology appendices.Why Gemini is the fastest‑growing name (analysis)
Gemini’s growth is not an accident; it is the product of three structural advantages that translate to rapid visibility and trial:- Distribution and product integration: Google can surface Gemini in Search, Chrome side panels, Gmail, Docs, Drive and on Android phones — reaching users where they already work and search. Those in‑product hooks reduce friction and create frequent short interactions that compound into habit.
- Viral multimodal features: Breakout features — notably image editing and quick multimodal creative tools — created social sharing loops (memes, short videos, image edits) that drove downloads and attention spikes. Those viral moments accelerate awareness faster than slow, organic discovery.
- Enterprise packaging and connectors: Gemini’s enterprise positioning (Workspace integration, governance controls) makes it straightforward for organisations already in Google’s ecosystem to pilot and then scale the assistant for workplace tasks — a different path to growth than pure consumer install numbers.
How to read awareness vs. usage vs. market share
Three common measurement traps
- Confusing “recognition” with “use”: A name can be widely known without being actively used for work. Telecompaper measures the former; app telemetry measures the latter.
- Over‑reliance on web‑visit trackers: Tools that track visits to public web chat pages undercount embedded usage inside productivity suites, mobile apps, and APIs. Google and Microsoft can capture enormous usage inside Workspace and Microsoft 365 that never hits a public web page.
- Mistaking spikes for durable adoption: Viral features create attention peaks; durable adoption requires product fit in repeated tasks and sound governance for enterprise use.
Implications for consumers and Windows users
Short‑term (consumer) impacts
- Discovery and experimentation will remain the default mode: users will try multiple assistants until one fits their habitual workflows.
- Consumers who live in a single ecosystem (Google or Microsoft) will be nudged toward that company’s assistant because it reduces friction when composing emails, summarising documents or searching the web.
- For privacy‑sensitive consumers, the differences in data‑use defaults and training policies across vendors matter. Consumer tiers can and do differ from enterprise contractual guarantees; avoid pasting highly sensitive data into public tiers.
Short‑term (Windows power users and admins)
- Expect more, not fewer, assistant touchpoints inside Windows and Office: Copilot features will appear in Office flows and in OS side panels; admins need to treat assistant governance like any other app policy.
- DLP, conditional access and SSO logs become the primary telemetry for real usage: public web traffic is an imperfect proxy for in‑org adoption.
Enterprise and Windows admin playbook (practical steps)
- Inventory assistant touchpoints: catalogue where assistants appear (Search, Gmail, Docs, Outlook, Office add‑ins, browser sidebars, API integrations).
- Define data classification rules: treat PHI/PII/IP as prohibited from public assistants unless covered by an enterprise contract with explicit non‑training and retention guarantees.
- Implement DLP and Purview controls on endpoints and cloud connectors.
- Use tenant‑level enterprise plans for high‑risk workflows; require contractual assurances around training data and data residency.
- Pilot with measurable success criteria: time saved, reduction in repetitive tasks, error rates from hallucinations.
- Require output verification workflows for regulated work (legal, finance, healthcare).
- Maintain a multi‑assistant fallback and an incident response playbook for outages or data incidents.
Safety, privacy and legal risks — the fast checklist
- Hallucination risk: generative models can assert plausible but incorrect facts; always require human review for external‑facing or regulated outputs.
- IP and copyright exposure: image‑generation and text‑training disputes continue to generate litigation risk; check vendor indemnities and licensing terms.
- Data training defaults: vendors’ consumer defaults on whether prompts are used to train models vary; enterprise contracts can change that — insist on written non‑training clauses where necessary.
- Plugin and integration attack surface: third‑party plugins or browser extensions can leak credentials and amplify data exfiltration. Treat plugins as third‑party code and vet them.
- Concentration and regulatory scrutiny: as a few platforms accumulate distribution power, antitrust and competition policy questions sharpen — especially where assistants become default search or mail assistants.
Strengths and limits of the Telecompaper snapshot — critical analysis
Strengths
- Timely local insight: Telecompaper’s Dutch Consumer AI Monitor gives a focused view of awareness in an important European market, which is valuable for local policy makers and businesses.
- Demographic detail: the report’s attention to age, education and digital engagement helps target adoption and training efforts.
Limitations and risks
- Methodology transparency: the public summary lacks full disclosure of survey methodology (sample size, weighting, margin of error, exact question wording) in the accessible summary. That omission makes it difficult to compare percentages directly with other surveys. Users should request the full methodology appendix before making high‑stakes comparisons.
- Awareness vs. active use: as noted, recognition numbers do not equal usage metrics or enterprise penetration; complementary telemetry is needed to make procurement or governance decisions.
- Rapidly moving landscape: 2025 has been a year of rapid product releases and feature rollouts (e.g., new Gemini models, ChatGPT upgrades). Survey snapshots can be outdated within months; continuous measurement matters.
Cross‑checks and independent corroboration
Telecompaper’s direction — ChatGPT widely known, Gemini gaining fastest — aligns with multiple independent indicators:- Web and app telemetry firms show ChatGPT as the largest single public destination while Gemini’s web traffic and app installs grew rapidly through 2025.
- Major outlets report Google’s focused distribution strategy and viral multimodal features as the primary driver of Gemini’s acceleration.
- Industry analyses for enterprise usage emphasise that Microsoft Copilot’s strength is its deep embedding in Microsoft 365 and Windows — a different path to durable adoption, particularly in corporate environments.
Tactical recommendations for readers
- Consumers: try multiple assistants, but avoid sharing sensitive personal or business data on consumer tiers; test premium enterprise tiers or vendor addenda before using them for regulated work.
- Power users: evaluate assistants by task class — creative long form, short micro‑tasking, code generation, or multimodal editing — and keep a small toolkit rather than betting on a single assistant.
- IT leaders: prioritise inventory and governance now. Add assistant usage to standard IT procurement checklists: DLP, indemnities, non‑training clauses, and provenance requirements for generated media.
- Policy makers & regulators: focus inquiries on how distribution advantages influence competition, how training datasets are built and contracted, and what consumer protections are in place for derivative content and image use.
Looking ahead: what to watch next
- Retention vs. trial: can Gemini convert rapid trials into habitual, daily usage at scale?
- Enterprise contract terms: will vendor non‑training guarantees and data‑residency options become standard?
- Benchmark and governance convergence: will independent benchmark suites, provenance metadata and content credentials become mandatory for commercial publishing of generated media?
- Regulatory moves: increased scrutiny of distribution tie‑ins and potential consumer harms may shape product design and rollout strategies.
Conclusion
Telecompaper’s Dutch Consumer AI Monitor offers a crisp, local snapshot: ChatGPT is still the most recognised assistant among consumers in the Netherlands, but Gemini is growing fastest in awareness — a pattern that fits into a larger 2025 trend where distribution, in‑product integration and viral multimodal features change the competitive map. The headline numbers are useful for marketers and product strategists, but they should be combined with active‑use telemetry, enterprise contract reviews and careful governance planning before organisations or policymakers draw definitive conclusions. The practical takeaway for Windows users and IT professionals is straightforward: prepare for a multi‑assistant reality by inventorying touchpoints, enforcing DLP and identity controls, piloting with measurable goals, and insisting on clear vendor commitments for data use and provenance.Source: Telecompaper Telecompaper
