ENFR
Book a demo
Product · WizAI

Wizville's AI agents

Nearly twenty specialised agents work on your customer feedback continuously, analysing all your data and supporting your strategic decisions. With the agents, focus on what really matters.

Your Wizville agents
They run inside the platform, on your data
WizAI
Your agents, by family
5 families
Nearly twenty agents, from collection to action
Collection 3 getting customers talking
Analysis 7
Review management 3 replies and priorities
Summaries & actions 6 summaries and action plans
Dashboard Reviews Competition Reports
See the agents in a demo
Why agents

One agent per task, not another chatbot

The voice of the customer arrives faster than anyone can read it. Wizville's agents take on the repetitive part of the work: following up, reading, comparing, drafting a first version. They prepare, your teams decide.

  • Specialised
    Each agent has a single job and works on your data, with your rules and your brand's tone of voice.
  • Where you already work
    No separate tool, no second login. The agents run inside the platform, alongside your dashboards and your reviews.
  • People stay in control
    What an agent produces is a proposal, visible and editable. Nothing goes out to a customer without approval.
Discover "Retain your customers"
Reviews handled in one hour Example
With suggested reply ≈ 40
Replies written from scratch ≈ 10
The agent writes the first draft, your teams review and publish.
Illustrative example, to be calibrated on your own figures
The catalogue

Five families of agents, from collection to action

Each agent does one specific thing, on your data. One illustrated example per family, and all the others in plain view.

01
Collection

Get useful feedback, not just scores.

Open question · initial verbatim didn't like it at all
AI follow-upWhat exactly did you not like?
MD
Marie D.“Twenty minutes waiting at the checkout, nobody on the shop floor.”
Enriched verbatimCheckout waitShop floor service
In pictures
AI verbatim follow-up

A score with no explanation? The agent asks the follow-up question that gets the customer to be specific, and turns a “didn’t like it” into an actionable pain point.

Product reviews

Picks the products to be rated and tailors the request to the channel.

AI voice collection

The customer answers by voice; the agent transcribes, then asks its follow-up questions in the same box.

Voir en image Back
02
Analysis

Read your results and the market's, without building a single pivot table.

Lyon Part-Dieu area · topic “Wait”
Your store3,9
Competitor A4,3
Competitor B3,5
Gap −0,4 vs A
In pictures
Local competition

Around each store, how you are perceived compared with competitors in the area, based on public reviews.

Autonomous semantic taxonomy

Builds and maintains the taxonomy of your topics, tailored to your sector.

Semantic classification of verbatims

Links each comment to its topics, with the associated sentiment.

NPS and satisfaction analysis

Isolates what is really moving in your scores and names the topics behind it.

Network performance

Who is slipping, who is improving, and what sets them apart, site by site.

National competition

Your brand positioned against the other players in the sector, perception gaps included.

Sales and satisfaction correlation

Observed links between sales and satisfaction, topic by topic.

03
Review management

Handle the right review, at the right time, with the right reply.

★★☆☆☆Marie D. · “Wait too long on Saturday”
Hello Marie, thank you for your feedback. Twenty minutes' wait is too long: we are opening an extra checkout on Saturdays from next week…
Review and publishRewrite
In pictures
Suggested reply

A reply ready to proofread, in your brand’s tone, based on what the customer wrote.

Customer satisfaction summary

A customer's full history in a few lines: scores, visits, comments.

Handling prioritisation

The queue, sorted: what is urgent, what is sensitive, what can wait.

04
Summaries

Go from a wall of data to three useful lines.

Week 36 · South-East network
Checkout wait12 magasins
Product availability4 magasins
Fitting room cleanliness2 magasins
In pictures
Priority actions summary

Findings translated into actions: what to handle first, at which level of the network, and why.

Multi-source summaries

Surveys, online reviews and verbatims from a given period brought together in a single summary.

Clustering by priority

The sites and regions that share the same problem, grouped together.

Chart and table summary

A plain-language reading of a chart, a table or a dashboard.

05
Actions

Go from findings to what needs doing, at every level of the network.

Topic “Checkout wait”
Head office
Review the Saturday staffing schedule across 12 stores
Region
Prioritise hiring extra checkout staff in Lyon
Store
Open checkout 4 from 2 p.m. on Saturdays
In pictures
Action recommendations by level

At head office, in the regions or in store, everyone receives the actions within their reach, phrased for their level of decision.

Sector business policy

Sets out the rules of your sector and your brand: what matters, what needs escalating.

The agents work on your Wizville data and on the public reviews of your market. The agents propose, your teams decide.
In detail

Four agents you will see at work in the demo

ChatGPT Image 2 sept. 2026, 15_53_41 (1)
ChatGPT Image 2 sept. 2026, 15_52_11 (2)
ChatGPT Image 2 sept. 2026, 15_52_12 (3)
ChatGPT Image 2 sept. 2026, 15_53_42 (4)
01

AI verbatim follow-up

The customer left a score without writing anything, or just wrote "didn't like it"? The agent asks the follow-up question that draws out the detail, and turns a silent score into a specific reason, ready to be analysed.

02

Local and national competition

Around each store and at chain level, the agent compares how you are perceived with the market, based on public reviews, then points out the gaps.

03

Suggested reply

The agent drafts the reply to a review in your brand's tone, drawing on what the customer wrote and on the history of the relationship.

04

Priority actions summary

Across all sources, the agent writes down what to handle first, at which level of the network, and groups the sites that share the same problem.

What an agent never does on its own

Agents speed up the work, they do not replace it. Three principles that apply across the platform.

Nothing goes out without you
A suggested reply remains a draft until the person replying approves it.
Your data stays yours
The agents work on your account's data and on the public reviews of your market, within each user's scope.
Enabled one by one
Each agent can be enabled, disabled and configured according to your needs and the organisation of your network.
The agents propose, your teams approve: that is the rule across the whole platform.
FAQ

Frequently asked questions

An assistant specialised in a single task: asking the right follow-up question, analysing a score, preparing a reply, writing a summary, recommending an action. It works inside the platform, on your data, and delivers a result that your teams approve.

A chatbot waits to be spoken to. An agent has a precise mission and runs within your workflow: when a review comes in, when a period closes, when a committee meeting needs preparing.

On the data in your Wizville account (surveys, reviews, verbatims, scopes) and, for competitor analyses, on the public reviews of your market. Each user stays within the scope assigned to them.

No. The agent prepares a proposed reply; the person in charge reviews it, adjusts it if needed, then publishes.

Yes. Agents are enabled family by family and agent by agent, according to your needs and your configuration. Your Wizville contact helps you choose the ones that matter for your teams.

Yes, that is their home ground: analyses by site and by region, review prioritisation store by store, grouping of sites that share the same problem.

Want to see the agents at work on your reviews?

30 minutes with a Wizville expert to see the agents running on your data, your reviews and your network.

Request a demo