WhatsApp & Messaging AI: Who answers the same customer messages, again and again, across WhatsApp, email and chat?
We build messaging agents that read each message, look up the answer in your own systems and draft the reply in the channel you already use, including WhatsApp Business, email and your helpdesk. Agents draft and prepare; a person on your team approves every send.
Order status, bookings, opening hours and resets: the same questions arrive in every channel, and someone types the same answers.
02
Answers spread across tools
To reply, your team checks the order system, the courier and the inbox, one tab at a time.
03
Messages that arrive after hours
Questions come in when nobody is at the desk, and they wait until someone is.
What we build
What our messaging agents do
Understand each message
Name the intent of every message, with a confidence and a reason, and send the unusual ones to a person.
Answer from your own data
Look up the order, booking or account before drafting, so the reply is about this customer, not a template.
Draft in your channel
Replies are drafted where your team already works: WhatsApp Business, Gmail, Microsoft 365 and Teams, Slack or Zendesk.
One-click approval
A person reads the draft and sends it with one click, or edits it first.
Playbooks your team can change
Each kind of question has its own playbook, so your team can change how a category is handled without waiting for a developer.
Voice, and messaging without internet
Speech recognition and a synthetic voice that run on a small device, for teams that work where the network fails.
You stay in control
Agents draft and prepare. A person on your team approves.
Nothing goes out unapproved
Agents create drafts. A person on your team approves every email, message, filing and payment record. By default, nothing is sent, filed or paid by an agent alone.
Block, don't guess
When records disagree, the agent stops and asks a person. A silent guess is treated as a defect.
A record of every run
What was done, why, and what is still open. Each decision shows its confidence and reasoning.
Inside the tools you already use
No rip-and-replace. The agents work inside your existing inbox and systems, and your team keeps the tools it knows.
Code does the arithmetic
Fetching, totals and filing are done by ordinary code. The model reads, matches and drafts.
Rules we live with ourselves
We run our own finance and admin on agents, under these same rules.
Proof
Proof
Our customer-support agents
Support agents that draft the answer before a person opens the ticket
Fleek runs a B2B marketplace for second-hand wholesale clothing. Its support team was answering the same "where is my order?" questions by hand, across five tools.
We built agents that read each ticket, name its intent, look up the order and leave a ready-to-send reply inside Zendesk, the helpdesk Fleek already used. The agents never reply to a customer directly: every answer is a draft that a person approves.
Each kind of ticket has its own playbook, so the support team can change how a category is handled without waiting for a developer. Our engineer ran a hands-on AI master class for the team, so they could shape the playbooks themselves.
“The responses have become a lot more reliable … one click send. That's amazing.”
Fleek customer-experience team, daily sync, October 2025
Source: Fleek–Wisflux working sessions, October–November 2025
~10 days
from the first ideation call to agents drafting replies in the live queue
84–85%
of "where is my order" tickets classified correctly, checked against 50–100 tickets hand-labeled by Fleek's team
Our offline voice AI work
Voice AI and messaging that keep working when the network is gone
When a disaster takes out mobile networks, first responders still need to call for help, share what they see and report it upward. For ATR, one of Japan's leading communications research institutes, we engineer X-FACE: a portable box that creates its own network and runs speech AI on the device itself. A local language model, which pulls the disaster facts out of each report offline, is being added in 2026.
Japanese and English speech recognition, a wake word and a synthetic voice run on the box, radio chat is transcribed live, and field reports sync to Japan's national disaster information systems once any link is available.
The fine-tuned language models are ATR's own, developed with its AI partner. Our part is making them run reliably offline on small hardware.
Disclosure: X-FACE is being commercialised as VICINO AI by NEI Globe Networks, Kyoto, where Wisflux's founder is Director and CTO.
Source: ATR–Wisflux contracts, activity reports and correspondence, 2020–2026
Since 2020
consecutive annual contracts with ATR
2 countries
of field drills and deployments with responders, in Japan and the Philippines
—ISO/IEC 27001:2022 and ISO 9001:2015, certified by TNV System Certification, valid to October 2028.
—GDPR-ready practices.
—Read-only access to your data first; write access only when it is earned.
—Background checks for engineers, secrets in a vault, anonymized data outside production, and on-premises deployment where required.
—You own everything: code in your repository from the first commit, running in your cloud or on your premises.
How we start
One workflow first. Real data, drafts first.
01
Workflow review call
Our founder, Chandraprakash Sharma, walks through one workflow with you and says plainly whether an agent can take it on, and if not, says that too. There is no charge for this call.
02
Agent pilot
One workflow, on your real data, drafts first, with success criteria agreed before we start. Your team compares the drafts with its own answers.
03
Roll out, one category at a time
We switch on the standard cases first and train the people who will use it. Hard cases stay with your team until it is confident, and every draft is still approved before it goes out.
We stay until it is used: a project ends when your team uses the system daily, not when it goes live. Every engagement is quoted to its scope.
You see the work as it happens: daily progress notes, a demo every sprint, and a work report behind every invoice showing where each hour went.
Other ways to work with us: a fixed-scope build with milestones, a dedicated team of named engineers, or run and support for systems we built or inherited.
Questions
WhatsApp & Messaging AI: common questions
Do you work with WhatsApp Business?+
Yes. WhatsApp Business is one of the messaging channels we have integrated, alongside Gmail, Microsoft 365 and Teams, Slack and Zendesk.
Will the agent reply to customers on its own?+
Not by default. Every reply is a draft that a person approves. Later, if you choose, a category that has proved itself can run without approval; that is your decision, never a default.
Who sees our data?+
We are ISO 27001 certified. We start with read-only access, use read replicas instead of production credentials, keep secrets in a vault, and can keep sensitive data on private, open-weight models.
What if the agent gets it wrong?+
Agents start by drafting only, and a person approves everything that goes out. We measure accuracy daily against your own examples, and every miss becomes a test case. Unclear cases stop and ask. Later, if you choose, a category that has proved itself can run without approval; that is your decision, never a default.
We've already tried ChatGPT or Claude for this. What do you add?+
Many teams now try. The demo is easy; production is not: connecting your systems, measuring accuracy, keeping a person in the loop, handling the unusual cases and keeping it running. We build with the same AI tools, and we can teach your team as we go.
Are we locked in if it doesn't work out?+
No. You own everything from the first commit: the code lives in your repository and runs in your cloud or on your premises. If we part ways, you keep everything.
Start with the questions customers ask most.
Tell us which messages your team answers most often today. Tell us about one workflow. We'll walk through it with you and say plainly whether an agent can take it on — and if it can't, we'll say that too.
