Skip to content

We launched our new Conversational AI platform for enterprises

We rewrote the stack from scratch for a new AI-first generation: unified omnichannel, conversational AI CRM, enterprise Reasoning and measurable operations, built to scale without stacking tools.

Marlos Carmo

Founder of Tolky

Updated on November 15, 2026

11 min read
We launched our new Conversational AI platform for enterprises

Share

Update, November 15, 2026. This text records the May 2026 launch. The new platform is now available to all customers: see the release notes in the public changelog (latest published version: v2.3.0, June 22, 2026) and plans at /precos. Active channels: WhatsApp, webchat, Microsoft Teams and voice; Instagram, SMS and email are listed as "coming soon" at tolky.to. Passages below that mention other channels describe the launch vision, not what is live today.

We're launching our new B2B Conversational AI platform: a structural leap that bundles years of lessons from high-volume customers into one environment. This is not a feature pack on top of the previous generation: it is a full rebuild of our stack, designed for speed, scale, mobile, internationalization and the pace of change generative AI demands in production.

We start from a simple, demanding premise: conversation stops being one channel among many and becomes the primary interface between customers, teams and internal systems. Instead of phone queues, rigid forms or keypad-driven bots (which break when intent drifts from the script). The operation works in natural language, with history, CRM context and brand-aligned business rules.

Why we rewrote the platform, not just updated it

In October 2025, we began rewriting the product. After two years in conversational AI, it became clear that the platform that brought us this far would not reach the next level: volume, deep integrations, enterprise governance and iteration speed the market now expects.

The rise of vibe coding and new development paradigms reinforced our decision: we needed a new foundation in architecture, performance and experience, not patches on top of 2023 choices. In May 2026, selected customers were already running the new platform and the rollout to the installed base was starting, alongside the new site at tolky.to.

If you're evaluating automation platforms, the message is direct: we're entering a new growth phase focused on organizations where interaction volume, legacy integration and governance cannot live in fragmented tools.

The problem we're solving with this generation

We see most enterprises with huge message volumes operating on a mosaic of tools: a CRM, a helpdesk, a WhatsApp sender, BI spreadsheets, generic automations and, on top, a chatbot that does not truly talk to any of it.

The symptoms are predictable:

  • Customers repeat context at every handoff or channel switch
  • Marketing launches campaigns, but service does not see lead origin
  • Finance resolves invoices in one system; support does not know the bill was already sent
  • Management exports CSVs for reports; conversations do not become data in real time
  • AI looks great in demos but does not write to the ERP or honor exception policy

We built this generation to attack that pattern with an integrated conversational ecosystem: service, management and intelligence on the same omnichannel layer, without hiding complexity behind outdated menus or promising a "magic bot" in the corner of the site.

AI-first architecture: AI at the center, not on the surface

For us, AI First means artificial intelligence is not limited to a superficial text layer. It runs through:

  • Dynamic knowledge base: approved, versioned, curated content, not unconstrained hallucination
  • Conversational CRM: every message updates pipeline, tags and next steps
  • Human console: when someone takes over, context is already assembled
  • Active campaigns: outbound and follow-up share the same history as inbound
  • Semantic voice: web, WhatsApp and phone (smart IVR) in one operation
  • Integrations: ERPs, finance and corporate data when resolution requires reading or writing real records in the customer's infrastructure

The outcome we're pursuing is not only "automation to shrink headcount." We want higher volume with quality, lower cost per contact, faster first response and a guarantee that when a human joins after the AI, they already know why the customer wrote in, what was tried and what sensible next step follows.

Context-aware handoff (history, intent, CRM data and explicit escalation reason) is central to our value proposition, not a UX detail.

Our operations and BI dashboard

The conversational ecosystem: five fronts, one operation

We organized the product into five connected fronts, from first contact to operational intelligence. Everything shares the same thread of context; not modules you "integrate later."

FrontWhat it coversKey capabilities
Conversations & OperationsHybrid AI + human serviceUnified inbox (WhatsApp, web chat, Microsoft Teams and voice live; Instagram and SMS coming soon), smart queues, real-time SLA, departments with permissions, AI Voice
Relationship & AI CRMSales and lifecycleNative pipeline, live profile fed by conversation, segmentation, opportunity alerts, WhatsApp campaigns (other channels coming soon)
Automation & OrchestrationFlows and agentsBranching builder, APIs for legacy systems, SDR/Sales/Closer/Support agents, multi-agent Reasoning, real-time triggers
Knowledge & IntelligenceBrand and accuracyCentral knowledge base, canned replies, continuous refinement, sentiment and intent, audio/image/video on WhatsApp
Governance & BIEnterprise controlAutomatic improvement alerts, analytics by channel, CSAT/NPS, audit trails, SSO (Enterprise plan), encryption and LGPD

Instead of stacking Service + CRM + Contacts + Content + Departments + Flows + Campaigns + Automations + Alerts + Follow Up + AI Forms + BI across vendors, we concentrate that ecosystem in one operation connected to the customer's data.

The five pillars of our launch

1. AI-first architecture

Conversational resolution at the center, with generative models wired to approved knowledge, business rules and CRM context, not a standalone LLM answering without memory or policy.

2. True omnichannel

WhatsApp, web chat, Microsoft Teams and voice live, with other channels (Instagram, SMS and email) coming soon, in one operation with unified history. Customers do not "start over" when switching channels; teams do not lose the thread on escalation.

3. Measurable operations

Native BI, Follow Up, AI Forms and conversational CRM turn customer service into actionable data for management, without parallel spreadsheet workflows or constant manual exports.

4. Context-aware handoff

When AI escalates to a human, the agent sees full history, intent and CRM in plain sight. Customers do not repeat name, order and ticket number for the third time.

5. Enterprise-ready

Documented APIs and a Reasoning orchestrator for critical flows: integrate ERPs, finance and corporate data with governance, curation and clear boundaries where the company always prefers a human in control.

Our conversation workspace

AI CRM for the conversational era

Our positioning is explicit: we're not "another bot." We're relationship infrastructure for an era when businesses and customers already talk via message, audio and voice, not only forms and email.

We're the Conversational AI platform that turns customer service, sales and relationships into an intelligent, omnichannel and humanized operation.

AI CRM for the conversational era. Relationships became conversational; we bring service, management and intelligence into one living ecosystem.

Three axes sustain that vision at once:

AxisWhat changes in practice
CommercialFast first reply still warms leads; qualification happens in conversation and the lead is routed to the sales team
SupportEffective automation frees humans for negotiation, exceptions and empathy
DataEach interaction becomes input for product, campaigns and process; it does not evaporate in chat

The market we see: beyond "magic bot" hype

Brazil's generative AI market has matured. Operators want SLAs, security, traceability and reports. Marketing and sales need conversion and pace in first replies after campaigns. Finance needs negotiation and payment copies on the right channel, without friction.

The shift underway is from form-, queue- and menu-driven operations to intelligent conversation-driven operations. Enterprises with huge volumes in care, sales, collections, support and lifecycle must serve more people with higher quality in less time, without losing personalization or control.

We believe mature Conversational AI is how you scale without multiplying operating cost or degrading CX. That's why we position this platform as infrastructure for those outcomes, not a static widget.

The next generation of enterprises won't only be digital. They'll be conversational, intelligent and integrated, able to serve at scale without treating people like numbers.

Our manifesto

Enterprise trajectory and what we learned in the field

Our trajectory includes institutional and industrial clients with major public impact, including a public case with Brazil's National Council of Justice (CNJ) (see our cases page) and large brands that depend on true omnichannel reach.

The governance, traceability and legacy integration bar those profiles require shaped this product generation: documented APIs, Reasoning for critical flows, curated knowledge and controls where regulated sectors or internal policy require a human on the final decision.

With this new structure, we serve organizations where volume, legacy and compliance do not fit in shallow automation layers or disconnected SaaS stacks.

Our automation and orchestration

Applications by area

The cases below are intended use cases, not measured results. Our platform serves enterprises with high volumes in customer service, sales, marketing, collections, technical support, Customer Success and HR. Our modular design lets you connect what hurts most today first and expand without swapping stacks.

AreaUse casesGoal
Care / SupportFAQ, order status, ticket creation, proactive NPSQuality deflection; humans only on complex cases
Sales / MarketingLead qualification, routing to sales, post-campaign follow-upConversion in the first response window
Finance / CollectionsInvoice copies, negotiation, payment confirmationResolution on the preferred channel (e.g. WhatsApp)
CS / RetentionConversational health score, alerts (as configured)Action before cancellation
HR / InternalBenefits, time-off, policy questionsScale without an infinite HR queue
OperationsInbound triage, CRM ↔ ERP reconciliationLess manual work between systems

Deep integration: what separates curiosity from product

Patrick Bonnereau, our co-founder, puts it well:

When AI talks to ERP, finance, CRM and operations in real time, the company stops patching tools together and starts operating with end-to-end context.

Our bet is AI First with governance: scale with maturity, not empty promises. That includes architecture for continuous evolution, response quality metrics, human handoff with context and a technical base that sustains growth without sacrificing experience.

Operational intelligence and reporting

What our founders say

Marlos Carmo, co-founder:

That is why we rebuilt everything. We are launching the new site at tolky.to: in May, some customers were already testing the new platform and it was starting to reach everyone. A new phase for Tolky.

FAQ about the launch

Is the new platform available to everyone?

Yes. It has been live since launch and keeps evolving (see the public changelog). If you still run any flow on the previous version, our team supports the transition according to your account plan.

Do I need to migrate all channels at once?

No. The modular design lets you start with the highest-pain channel or process (e.g. WhatsApp + lead qualification) and expand later to the full CRM or API/MCP integrations. Other channels are coming soon. See plans at /precos.

How are we different from a chatbot + Zapier?

An isolated chatbot answers text; it does not operate as a relationship layer with live CRM, SLA, native BI, governance and multi-agent orchestration. Zapier connects apps but does not unify conversational operations or native context-aware handoff; the company remains the integrator. We deliver the full ecosystem.

Does the platform meet enterprise requirements (LGPD, audit)?

Our governance front includes immutable history, audit logs, end-to-end encryption and LGPD; SSO is available on the Enterprise plan. For security questions, read our guide on security and privacy in AI platforms and talk to our team.

Where is the official press release?

Journalists and editorial teams can access the release, media kit and fact sheet on our press page, with materials in Portuguese, English and Spanish.

Next steps

The new platform is available. Compare plans (from free Individual to Enterprise) at /precos and follow its evolution in the public changelog.

For editorial coverage: negocios@tolky.to · materials at /press

Share

Tags

About the author

Marlos Carmo

Founder of Tolky

Marlos Carmo is an AI entrepreneur and founder of Tolky, the conversational-era infrastructure and AI CRM that unifies intelligent service, multi-channel support (such as WhatsApp and voice), live CRM, and operational intelligence in a single ecosystem. He is a finalist for the SXSW Innovation Awards and a member of Francesco's Economy, a global network of young entrepreneurs focused on innovation and social impact. He works connecting Artificial Intelligence and digital transformation in projects for large organizations.

Next step

Want to bring this to your operation?

Tati, Tolky's AI assistant, understands your scenario in a few questions, suggests the best path and connects you with someone from the team, with full context.

Read also