Key takeaways
- The 5th edition of Genesys' "State of Customer Experience" surveys 5,811 consumers and 1,560 CX leaders across more than 20 countries. It contains a figure few commentators will pick up: 43 % of French CX leaders make loyalty and empathy their primary optimisation focus, against 29 % in Europe and 18 % worldwide. No country scores higher.
- The same report shows that France is also the European country where previously collected information is least often passed automatically to the agent: 34 %, against 44 % in Europe and 52 % worldwide. On reducing customer effort, France reports 6 %, against 24 % worldwide.
- The strongest relational ambition on the continent, served by the weakest execution. This is not a cultural paradox, it is an instrumentation gap: you cannot pass on context you never captured, and you cannot steer empathy you never measure.
- The report proves the point itself: 90 % of companies consider themselves effective at limiting wait times, while 44 % of consumers waited more than fifteen minutes at least once during the year and 83 % consider that threshold unacceptable. A perception gap of that size is always a measurement gap.
- Genesys concludes that you should modernise your platform. The diagnosis is sound and the conclusion legitimate for a platform vendor. We draw a different conclusion about sequencing: conversation measurement comes before the platform overhaul, because it is what tells you where the platform falls short.
A vendor's report, but survey data
Genesys has published the fifth edition of its "State of Customer Experience" study. The survey was carried out by an independent research firm between March and April 2026, among 5,811 consumers and 1,560 CX leaders and executives, across more than twenty countries and all sectors: banking, insurance, healthcare, retail, telecoms, energy, transport, public sector.
Let us say it upfront, because it shapes the reading: Genesys is neither a research firm nor a professional federation. It is a customer experience platform vendor, serving more than 8,000 organisations. A study published by a vendor rarely ends on a recommendation that contradicts its own catalogue, and this one is no exception: it calls for modernising the CX stack around a unified cloud platform.
With that caveat in place, the data is worth reading. The protocol is serious, the consumer sample is large, and above all the report publishes country-level results that nobody else publishes. What interests us here is what that data says about France, not the commercial conclusion the vendor draws from it. It is also, let us be clear, the report of a player whose platform is among the environments on which a conversational analytics layer naturally sits.
This reading extends those we devoted to the EY / SP2C Barometer 2026 and the KPMG CEE Barometer 2025. With one difference: both of those described the market as seen by companies. This one finally provides the consumer counterpart.
The French paradox
Let us start with the figure the sector in France should be proud of. Asked which focus they prioritise in their optimisation efforts, 43 % of French CX leaders cite loyalty and empathy. That is the highest rate of any country studied, well above the European average of 29 %, itself well above the global average of 18 %.
In other words: where the rest of the world optimises first for operational efficiency and cost to serve, France has chosen the relationship. It is a deliberate singularity, and it is consistent with what consumers ask for, since 94 % of them place as much importance on service efficiency as on the empathy they receive.
The problem appears when you look at execution.
| CX leaders' self-reported data | France | Germany | Ireland | United Kingdom | Other EU | Europe | World |
|---|---|---|---|---|---|---|---|
| Loyalty and empathy as primary optimisation focus | 43 % | 31 % | 37 % | n/a | n/a | 29 % | 18 % |
| Previously collected information automatically provided to the agent | 34 % | 46 % | 43 % | 48 % | 47 % | 44 % | 52 % |
| Manage to significantly reduce customer effort | 6 % | 5 % | 11 % | 21 % | 29 % | 14 % | 24 % |
| Offer "extremely personalised" service | 21 % | 18 % | 27 % | 28 % | 31 % | 25 % | 35 % |
On automatic context handover to the agent, France ranks last in Europe, ten points below the European average and eighteen below the global one. On customer effort reduction and personalisation, it ranks second to last, behind Germany. The United Kingdom reports significantly reducing customer effort three and a half times more often than France.
France therefore displays the strongest relational intent on the continent alongside one of the weakest delivery capabilities. This is not a moral contradiction, and it is certainly not a lack of seriousness on the part of the teams. It is a tooling problem, and it can be read in the first figure of the table.
A note on method, because it matters. Europe accounts for 23 % of the 1,560 leaders surveyed, roughly 360 people spread across the United Kingdom, France, Germany, Ireland and other EU countries. The French base therefore runs to dozens, not hundreds. These results are also self-reported: they measure what CX leaders think of their own organisation, not what their customers experience. Read them as orders of magnitude and gaps between countries, not as precision measurements.
34 %, or the most expensive figure in the report
Only one third of French organisations automatically make available to the agent the information the customer has already given. In the other two thirds, the agent searches manually, or asks again.
The report quantifies precisely what that costs on the customer side:
- 95 % of consumers consider it important that their information follows them from one channel to the next.
- 52 % of companies still do not pass on to human agents what was collected earlier in the interaction.
- 26 % of consumers had to repeat to a human agent what they had already explained to a chatbot, and 33 % had to repeat it to several human agents in succession.
- 20 % state that this repetition irritates them enough to want to end their relationship with the brand for good.
One consumer in five considers leaving a brand over an internal plumbing issue. Set against the fact that 85 % of consumers say they have already cut spending or stopped buying after a poor service experience, and that 47 % would switch after only two or three bad interactions with a brand they actually like, the bill is no longer a matter of comfort.
The most useful passage in the report sits in its European section: Europe is the region where customer information is least often made automatically available to agents, and the one where agents most frequently have to look it up themselves. The time a French advisor spends reconstructing context is time not spent doing what their organisation says it prioritises: the relationship.
A perception gap is always a measurement gap
The report contains a contradiction its author does not comment on, and which is nonetheless the heart of the matter.
| What companies report | What consumers report |
|---|---|
| 90 % consider themselves effective at limiting wait times | 44 % waited more than fifteen minutes at least once during the year, 83 % consider that threshold unacceptable, 29 % gave up seeking help |
| 93 % believe virtual agents are a good way to start a self-service interaction | 84 % grant a virtual agent no more than three attempts, fewer than 20 % grant more |
| 35 % worldwide, 21 % in France, say they offer extremely personalised service | 92 % expect every organisation to deliver an experience matching the best they have ever had |
Nine companies in ten believe they perform well on an indicator where nearly one customer in two experienced the opposite. No organisation deliberately deceives itself to that extent. The gap has a simple, technical cause: the company measures what its tools report, namely average wait times and a sample of monitored calls, while the customer remembers the call where they waited twenty minutes.
An average wait time says nothing about the tail of the distribution. A sample of 2 to 5 % of conversations, still the norm in manual quality control, says nothing about the remaining 95 %. As long as steering relies on those two instruments, the gap between internal perception and lived reality does not narrow, it persists structurally.
It is the same reasoning we applied to the EY / SP2C Barometer: you do not steer what you do not observe. The Genesys report supplies the numerical demonstration, across twenty countries.
Empathy has become a measurable criterion, and that is new
For a long time, empathy was treated as the unmeasurable part of the job, assessed by feel during a quarterly side-by-side listening session. The report's data moves it firmly into the category of indicators to be steered:
- 94 % of consumers place as much importance on empathy as on efficiency.
- 57 % would certainly or probably turn to a competitor after perceiving a lack of empathy during an interaction.
- 91 % believe a company is only as good as its customer service, up nine points in a year.
- Being unable to reach a human agent tops the list of situations that make a consumer no longer want to deal with a company.
When more than one customer in two leaves over perceived lack of empathy, empathy is no longer a diffuse human quality, it is a churn driver. And a churn driver can be measured.
That is precisely what conversational analytics makes possible: identifying across all exchanges the reformulations, active listening, talk-time ratio, ownership markers, and the moments where a customer expressed an emotion that went unacknowledged. Not to grade advisors, but to equip their skills development with examples drawn from real interactions, and to detect departure signals before they become cancellations.
For a French sector that has put loyalty and empathy at the top of its priorities at 43 %, this is the missing link between intent and proof.
The virtual agent's three attempts
The report provides one immediately actionable operational data point: 84 % of consumers grant a virtual agent a maximum of three attempts to resolve their problem, and fewer than 20 % are willing to grant more. Beyond that threshold, the study notes, tolerance does not decline gradually, it collapses.
The consequence is clear. The quality of a virtual agent is not judged by its share of handled conversations, but by its ability to resolve within three exchanges or hand over cleanly, context included. The report adds that 48 % of consumers with a worsened perception of AI attribute it directly to a bad experience in the past twelve months. A poorly calibrated bot does not just degrade one interaction, it damages trust in every AI initiative the brand runs, including the ones that come next.
That means treating the virtual agent like any other production channel: with a quality control setup that evaluates it against the same criteria as human teams. This is the subject we developed in our article on benchmarking internal teams, outsourcers and AI tools: as long as the bot sits outside the quality monitoring perimeter, nobody in the organisation really knows what it produces.
What we do not take on board
The report closes with five recommendations, the first being to modernise CX around a unified platform, treating cloud migration as a structural investment. The underlying finding is real: only 31 % of CX infrastructure is fully cloud-hosted, a figure unchanged since 2025, and for the first time since 2023, maintaining service quality on ageing systems outranks rising customer expectations among operational challenges.
We dispute neither the finding nor the value of a modern platform. We question the order of operations, for a reason the report supplies itself: 42 % of CX leaders cite demonstrating AI value and return on investment among their main challenges, while planning to devote 30 % of their customer service budget to it over the next twelve months. In Europe that pressure is identical, and 35 % of leaders make reducing cost to serve a priority, against 29 % worldwide.
A modernisation programme that starts without prior conversation measurement starts without a baseline: it does not know which contact reasons concentrate repetition, at which precise point in the journey context is lost, or which irritants actually weigh in cancellations. Eighteen months later it will be asked to demonstrate a ROI it has no means of establishing. That is exactly the challenge cited by 42 % of respondents.
Conversation measurement is inexpensive, deploys in weeks rather than quarters, does not require replacing what exists, and plugs into the infrastructure in place, including via SIPREC, a standard supported by the main contact centre platforms, or via a native connector. It produces the map that makes the modernisation programme defensible, then measurable once launched. It does not replace the platform: it indicates where to start and then allows its effect to be proven.
What the report implies for a French contact centre
| Study finding | What it implies | What conversation measurement brings |
|---|---|---|
| France last in Europe on context handover (34 %) | The advisor manually reconstructs what the customer already said | Identify the reasons and channels where context is lost, quantify the time and repeat contacts it costs |
| Only 6 % report significantly reducing customer effort | Customer effort is not steered because it is not observed | Measure real effort across 100 % of exchanges: repetitions, transfers, repeat contacts, callbacks within 48 h |
| 43 % make empathy their primary optimisation focus | A priority without an indicator remains an intention | Make empathy observable: reformulation, active listening, acknowledgement of emotion |
| 90 % of companies think they handle wait times well, 44 % of customers waited over 15 min | Steering relies on averages and a sample | Move from 2-5 % of calls monitored to 100 % analysed, and look at the tail of the distribution, not the average |
| 84 % grant a virtual agent only three attempts | The bot must be assessed as a production channel | Apply the same quality criteria to the bot as to human teams, and monitor escalation quality |
| 31 % fully cloud infrastructure, AI ROI hard to prove (42 %) | Modernisation must be prioritised and defensible | Provide the baseline before the project, and the proof of effect afterwards |
| 94 % of consumers want to know they are talking to an AI | Transparency becomes an execution requirement | Verify across all exchanges that the disclosure is made, like any other mandatory statement |
Where to start. Take the three contact reasons that generate the most volume. Across all conversations in a single month, measure just three things on those reasons: the rate of repetition of already-provided information, the repeat-contact rate within 48 hours, and the presence of listening markers. Within weeks you will have the map the Genesys report describes in global averages, but on your own customers. That is the prerequisite for any investment decision, including a platform overhaul.
Closing thought
France emerges from this study with a distinctive profile: the most assertive relational conviction in Europe, and among the weakest means of living up to it. That is uncomfortable reading, and it is rather good news. A conviction gap takes years to fix. An instrumentation gap takes weeks.
The report says 91 % of CX leaders believe their organisation will be able to orchestrate customer experience end to end within three years, and that 88 % anticipate a deep transformation of contact centre roles over the same period. Between those ambitions and today's 34 %, there is work that begins with neither a platform nor an autonomous agent. It begins by listening to what customers already say, in full rather than by sampling.
Make execution measurable
- Try it free: app.raisetalk.com/try
- Contact us: www.raisetalk.com/contact
French CX leaders rank loyalty and empathy higher than anywhere else in the world. They also pass customer context to their advisors less often than anywhere else in Europe. Both figures come from the same report, and they do not contradict each other: they describe a sector that knows what it wants to do and has not yet given itself the instruments to do it. The gap between 43 % of intent and 34 % of execution will not be closed by more conviction. It will be closed by making observable what actually happens in conversations, across all of them, continuously, in service of the advisor and never of their surveillance.
Source: Genesys, "State of Customer Experience 2026", 5th edition, French edition. Survey conducted by an independent research firm between March and April 2026 among 5,811 consumers and 1,560 CX leaders and executives across more than 20 countries. Professional results are self-reported; European country-level data rests on sub-samples of a few dozen respondents and should be read as orders of magnitude. See also our analyses of the EY / SP2C Barometer 2026 and the KPMG CEE Barometer 2025.
Download the full study: "State of Customer Experience" on genesys.com