Answers your customers. Knows when to hand over.
An AI agent configured around your business — your services, policies, and the questions you actually get. It handles the repeatable conversations on website chat and WhatsApp, qualifies the enquiry, and routes anything sensitive to your team.
Configured, tested, and handed over — not a self-serve subscription. Starting at $5,000.
Illustrative interface. The agent answers from approved sources and stops when a conversation needs a person.
One agent, configured around your business.
This is not a general-purpose chatbot dropped onto your site. The agent works from approved sources, follows rules you sign off, and stops when the answer is out of scope. The result is a support layer that is useful on day one and honest about its limits.
Answer common questions
Service details, availability, hours, delivery, returns, and policies — answered from the sources you approve rather than guessed.
Guide visitors to the right next step
Point people to the relevant service, page, or process instead of leaving them to browse.
Qualify enquiries
Ask the few questions that separate a real project from a browse, and capture contact details with your consent language.
Escalate and route
Hand over when asked, when confidence is low, or when the topic is outside the agreed scope.
Connect to agreed tools
CRM, help desk, booking, or messaging tools where the integration is agreed and technically supported.
Stay inside its knowledge
Answers come from configured sources. When it does not know, it says so and routes the conversation rather than inventing an answer.
Where it earns its keep.
The highest-value conversations are usually repetitive, time-sensitive, or happening outside office hours. These are the patterns we configure most often.
Professional services intake
Answers questions about services, process, and fit, then collects the details needed for a useful first call.
E-commerce support
Product, delivery, returns, and order questions — with an order lookup only where that integration is agreed and available.
SaaS and onboarding
Common setup questions and plan clarifications, without pulling the team into every thread.
Local service businesses
Hours, coverage, quotes, and availability — then a clean handover for the visit or estimate.
Every use case is tested against questions you supply. We do not claim a use case is live until it has been configured, reviewed, and accepted in your environment.
From your knowledge to a working conversation.
Four stages, in this order. Skipping the first two is how AI support projects end up sounding confident and being wrong.
Knowledge map
We inventory the questions you actually get and the sources that answer them: services, pricing rules, policies, FAQs, and internal notes.
Conversation design
We define tone, what the agent may answer, when it must stop, and how it hands over. Your team approves the rules before build.
Channel setup
Website chat and WhatsApp are configured and tested on the devices your customers use, including the handover path.
Test, launch, refine
We run the agreed use cases before launch, then refine for 30 days after handover within the agreed scope.
Two channels. Done properly.
Website chat and WhatsApp are the approved channels for this product. Other channels are separately scoped — we do not include them by default or imply they are already connected.
Website chat
An agent embedded on your site that answers common questions, guides visitors to the right service or page, and captures the details your team needs. It can work alongside your existing contact form rather than replacing it.
Typical scope: conversation design, knowledge sources, escalation rules, lead capture fields, and a handover path to email or your agreed inbox.
Conversations that continue on the channel many customers already use. WhatsApp has its own business account, template, and messaging-window rules, so this is configured within the platform's policy rather than around it.
Typical scope: business account setup or connection, approved message templates, routing, and escalation to a human number.
Instagram DMs, Facebook Messenger, email, and SMS can be discussed as separate work. Each has its own access, platform policy, and support requirements — none of them are part of the standard scope.
What it knows, and what it must never guess.
What we configure
- Approved knowledge sources: services, pricing rules, policies, FAQs, and internal notes you want used.
- Tone and boundaries, including phrases and claims the agent should avoid.
- Escalation rules: which topics, keywords, or confidence levels must reach a person.
- Opening hours, response expectations, and a fallback when nobody is available.
- Lead capture fields, consent language, and where the details are sent.
- What the agent must never answer — legal, medical, account-specific, or anything you classify as sensitive.
What we need from you
- Source material: current service descriptions, policies, FAQs, and pricing guidance.
- A named owner who can approve the knowledge and the escalation rules.
- Access to the agreed tools, or a contact who can grant it through official permissions.
- Decisions on sensitive topics and the wording you want used.
- A set of real customer questions, including the awkward ones, for testing.
Integrations depend on the agreed technical scope. We confirm what is supported before committing to it — we do not assume every CRM, help desk, or messaging provider can be connected, and we do not describe a connection as live until it has been tested.
From first call to handover.
A clear sequence with agreed checkpoints. You always know what is being configured, what is waiting on you, and what has been tested.
Understand the business goal
What should the agent take off your team’s plate, and what must stay with a person?
Select the workflow and tools
We agree the primary workflow and confirm which tools can realistically be connected.
Configure the agent and knowledge
Brand context, approved sources, tone, boundaries, and escalation rules.
Integrate the agreed systems
Only the integrations confirmed in scope, connected through supported methods and tested.
Test and refine
Real questions, edge cases, fallbacks, and handover — reviewed with you before launch.
Hand over with guidance
Documentation, basic operating guidance, and 30 days of refinement within the agreed scope.
Starting at $5,000.
The baseline package covers one agent and one primary workflow. Final implementation pricing depends on the agreed scope — a second workflow, a complex integration, or a large knowledge base is quoted separately.
Starting price. One agent, one primary workflow, configured and tested before handover.
- One AI agent and one primary workflow
- Brand and business-context configuration
- Configuration of the agreed knowledge sources
- Testing against agreed use cases
- Handover and basic operating guidance
- 30 days of post-handover refinement within the agreed scope
What sits outside the starting price
The following are separate costs or separately scoped work. They are not included in the $5,000 baseline, and we will not imply that they are.
Separate costs
The 30-day refinement period is bounded: it covers adjustments within the agreed scope, not unlimited training, unlimited revisions, or permanent support.
Broader agent strategy, model selection, and deployment work is handled through Softlogs.io.
Answered honestly.
Can the agent be customized to our business?
Which systems can it connect to?
Who oversees the agent once it is running?
What does usage cost?
How long does implementation take?
Can it answer in more than one language?
What happens when it does not know the answer?
Is this a subscription?
Tell us what your customers keep asking.
Share the questions, the channels, and the systems you have. We will tell you what can be configured, what needs an integration, and what the agent should hand to a person.
Not sure which product fits? Choose “Another enquiry” in the form and describe the problem.