The Centigen AI Sales Suite: one governed pipeline from first message to booked meeting

Most businesses in the UK and the Gulf have already tried an AI sales tool. A chat widget on the website, perhaps a WhatsApp bot, perhaps a browser plug-in that writes emails. The pattern is familiar: a promising first week, a few odd answers to customers, a duplicate contact in the CRM, and then someone quietly switches it off.

The tool was not the problem. The absence of governance was. Nobody had decided what the agent was allowed to say, what it must never say, when it should hand a conversation to a person, or what happens to the data it collects. The Centigen AI Sales Suite exists to fix that, and this article explains how.

AI sales fails on governance, not technology

Every AI vendor can now show you an agent that answers questions. That is no longer the hard part. The hard part is running one in front of real customers, every day, under the same accountability that applies to a member of your sales team.

That means four things have to be true before an agent goes live:

1. It answers only from an approved knowledge base, so a customer never receives an invented price or a made-up policy.
2. It holds a never-answer list: the topics where the correct response is to defer to a person, not to improvise.
3. Every conversation, decision and CRM write is logged, so a manager can audit what happened and why.
4. Personal data is collected with consent, kept for a defined period, and handled in line with the law that applies to you: the UAE Personal Data Protection Law in the Gulf, UK GDPR in Britain, and GDPR and the EU AI Act for anyone selling into Europe.

The Centigen AI Sales Suite is built around those four conditions first. The agents come second, as proof that the method works.

What the suite actually is

The Centigen AI Sales Suite is not a bundle of separate tools. It is one governed sales pipeline, with three connected capabilities running on a shared layer.

A website sales agent. Orion sits on your site, answers visitor questions from your approved knowledge base, qualifies the visitor against your criteria, and books a meeting directly into your calendar. It creates the CRM record properly, once, with the right fields populated.

A WhatsApp sales agent. The same qualification logic and the same knowledge base, delivered on the channel Gulf buyers actually prefer and UK buyers increasingly expect. Opt-in is captured explicitly at the start of the conversation, and the consent state travels with the contact.

Automated pre-call research. When a meeting is booked or a lead crosses your qualification threshold, the suite researches the prospect's company from public and permitted sources and writes a graded pre-call brief onto the CRM record. Your salesperson opens the meeting already knowing who they are speaking to.

The moment that sells this is simple. A visitor chats on your website, leaves, messages you on WhatsApp two days later, and the suite recognises them. Their qualification score carries across, the conversation history is intact, and a pre-call brief is sitting on the CRM record before anyone picks up the phone.

The shared layer is where the value sits

Connecting two chat channels to a CRM is not difficult. What makes the Sales Suite defensible at management level is the layer underneath the channels.

One identity. A contact is recognised across web and WhatsApp as the same person, so your CRM does not fill with duplicates and your salesperson does not ask the same qualifying questions twice.

One qualification model. A single scoring rubric, agreed with you once, applied identically on every channel. Two people looking at the same lead would place it in the same stage.

One knowledge base. The web agent and the WhatsApp agent cannot drift apart, because they read from the same approved source. Update a price or a policy once and both channels reflect it.

One governance and consent layer. Consent is captured once and honoured everywhere. Escalation rules are channel agnostic. Every action the suite takes passes through the same approval gate and lands in the same audit trail.

That audit trail is not a technical nicety. For a regulated business, or for any business whose board is asking hard questions about AI, it is the artefact that turns "we use an AI agent" into "we operate a governed AI capability". It is also what the EU AI Act now expects: from August 2026, customers interacting with an AI system must be told so, and the operator must be able to show how the system behaves. A logged, disclosed, human-supervised agent meets that standard by design rather than by retrofit.

Human handover, by design

An agent that never hands over is a liability. An agent that hands over too often is a chatbot with extra steps. The Sales Suite draws the line deliberately.

Every deployment defines the conditions under which the agent stops and a person takes over: a request outside the approved knowledge base, a topic on the never-answer list, a sentiment signal, or a customer simply asking for a human. When handover happens, the person receives the full conversation and the qualification context, not a blank slate.

In the current release, handover runs on the web channel. WhatsApp handover is in development and follows the same rules.

Your data stays yours

Two design decisions matter here and both are worth asking any vendor about.

First, every third-party subscription the suite depends on is held in your name, not ours. If you ever choose to walk away, the infrastructure, the data and the accounts are already yours.

Second, the research capability draws only from sources you would be comfortable defending: the prospect's own website, official registries, regulator registers, published bios and cited public sources. We removed social profile scraping from the design because it sits in a grey area under UAE PDPL and UK and EU GDPR alike, and a governed deployment should not depend on grey areas.

How an engagement starts

We do not start by installing agents. We start with an Agent Readiness Assessment: one business process, a specification of what the suite will do for you, the governance controls it will run under, and a clear estimate for the build. Whether you are a Dubai trading group, a Riyadh distributor or a Manchester services firm, the assessment is the same: it establishes what the agent may do, what it must never do, and which regulation it answers to before anything goes live.

That order is deliberate. The specification and the controls are written before a single line of the agent is configured, because that is the only way to put an AI agent in front of your customers and keep it accountable.

If your sales team is fielding the same questions every day, losing leads between channels, or walking into meetings without a brief, the Sales Suite was built for that problem.

Book your Agent Readiness Assessment and we will map your first governed sales process with you.