Pilot program
Prove it in production, in 60–90 days.
A production pilot with success metrics agreed up front. Your customers, your systems, one channel. Then you decide with real data.
How it works
Four phases, one decision
Discovery
We map your top use cases, the systems to connect and the channel to start with, and agree on the success metrics.
Configuration and knowledge base
We configure the agent's personality, knowledge base, tools and guardrails, and connect your systems through MCP.
Go-live on one channel
The agent starts serving real customers on web chat, WhatsApp, voice or a branch avatar.
Measurement and decision
We measure results against the agreed metrics, and you decide whether to scale, adjust or stop.
What you get
What's included
An agent in production
Configured with your knowledge base, tools and guardrails, serving real customers.
Connected to your systems
Integrations through MCP with the systems defined during discovery.
Success metrics agreed up front
Clear targets defined before go-live, for example resolution rate, handling time, CSAT or escalation rate.
Continuous improvement
LLM-as-judge evaluation of every interaction and A/B testing to improve the agent during the pilot.
Full visibility
CSAT, compliance and escalation metrics, with auditable conversations.
A data-driven decision
A results review at the end of the pilot to decide the next step.
Commercial terms
Pricing and terms
There is no price list. Pricing depends on the scope and the difficulty of the project — above all on the integrations our developers have to build — so every pilot is scoped and quoted on its own, after we look at what you already have running.
What drives the price
The channels you launch, the volume of conversations, whether a branch avatar is included, how much of the agent is custom and, most of all, the integration work. A custom setup is analyzed on its own and quoted in its own proposal.
Why integrations change the number
If you run a market-standard ERP, CRM or e-commerce platform, our MCP connectors plug in relatively fast. If your operation runs on systems with no APIs to work with, the integration takes more analysis and more development, and it costs more.
Why we start with discovery
Quoting without looking at what you use would be guessing. In discovery we map your systems, channels and processes, and only then write the proposal: use cases, integrations, success metrics and timeline. MindLab does not replace what you already have — it connects to it and adapts to improve the experience your customers get.
After the pilot
You review the measured results against the metrics you agreed and decide whether to scale to more use cases and channels. The terms for that next stage are agreed in the proposal.
FAQ
Frequently asked questions
How long does the pilot last?
Between 60 and 90 days in production, depending on the scope agreed during discovery.
Which channel should we start with?
Usually the one with the most volume or the clearest use case: web chat, WhatsApp, voice or a branch avatar. Other channels can be added later with the same agent.
What do we need from our team?
A project owner, access to the knowledge sources and systems to connect, and agreement on the success metrics.
What happens after the pilot?
You review the results against the agreed metrics and decide whether to scale to more use cases and channels.
Next step
Start your pilot
Book a demo to scope your pilot, or ask our AI agent how it works right now.
ROI calculator