B2B SaaS companies
How-to questions, access issues and billing — answered from your docs.
Customer support, grounded in your knowledge
SupportPilot is a proposed AI support platform that answers routine questions from your approved help content — always with the source — sorts and routes every ticket, and hands anything uncertain to a person with a summary.
Built forB2B SaaS companies
Hi! I answer from Nimbus Notes’ approved help articles, with sources. If I’m not sure, I’ll hand you to a person.
Figures from the prototype’s sample data.
Core features
Can it answer our customers?
Answers come only from articles your team has approved, quoting the relevant passage and citing it.
Who handles this ticket?
Billing, access, data, bugs, cancellations — each ticket gets a category and a priority the moment it arrives.
What happened in this thread?
A short summary of the request and the key details, so agents never start from scratch.
Where should it go?
Billing questions to billing, bugs to engineering, cancellations to retention — rules you can read and change.
And when it isn’t sure?
Below your confidence threshold, or for anything sensitive, a person takes over with the context attached.
Are we seeing the same problem?
Spots clusters — four tickets about invoices this morning — before they become a backlog.
Is it working?
Drafted, sent, edited and escalated, by category — measured honestly so you can trust the numbers.
The differentiator
Knowledge-grounded support automation with source references, and human escalation when answers are uncertain. No guessing, no invented policies.
Annual plans can be refunded in full within 14 days of purchase.
Source: KB-03 · Refund policy
KB-03 · approved · updated 2026-04-30
Annual plans can be refunded in full within 14 days of purchase. Monthly plans are not refunded, but you can cancel at any time to stop future charges. If you were charged twice for the same invoice, we refund the duplicate charge within 5 business days.
Human escalation
Set the confidence threshold. Questions below it go straight to your team with a summary. Move the slider and see which of these would be answered automatically.
Recurring issues
From twelve sample tickets this morning, SupportPilot groups the ones that share a topic, so a manager can fix the cause — or update the help article.
Who it’s for
How-to questions, access issues and billing — answered from your docs.
Orders, refunds and returns policies, routed to the right people fast.
Plan changes, cancellations and invoices, handled consistently.
Decision-makers: customer experience directors · support managers · operations executives
Business model
Monthly or annual subscriptions based on support volume and agent seats. Illustrative pricing for a proposed product.
Planned AWS infrastructure
SupportPilot is a proposed customer support platform for B2B SaaS companies, ecommerce businesses and subscription services. It must answer from each customer’s own approved knowledge, cite its sources, and keep ticket records reliable for agents and analytics. Designing on AWS lets it use managed retrieval over approved documents, generate grounded drafts and process tickets as they arrive.
Amazon S3
Support teams upload approved help articles, policies and macros as knowledge documents.
Amazon Bedrock Knowledge Bases
Questions are matched to passages from approved content only, with their source references.
Amazon Bedrock
Grounded response drafts and conversation summaries are generated from the retrieved passages.
AWS LambdaAmazon API Gateway
Tickets arrive from helpdesk and chat integrations, are categorised, routed and escalated by rule.
Amazon DynamoDB
Tickets, drafts, citations, agent actions and outcomes are stored for queues and analytics.
Across every stage
AWS Activate · company description
SupportPilot is a proposed AI customer support automation platform for SaaS companies, ecommerce businesses and subscription services. The planned application will retrieve answers from approved knowledge sources, categorise customer inquiries, summarise conversations and route complex cases to human support agents. We plan to use Amazon Bedrock and Amazon Bedrock Knowledge Bases for grounded response generation, Amazon S3 for knowledge documents and Amazon DynamoDB for ticket records. AWS Lambda and Amazon API Gateway will support processing and integrations. Support teams will approve knowledge sources and configure escalation policies.
Planned, not yet deployed. This website and the support console run in your browser on fictional data.
Founding team
Founder
Victor founded SupportPilot after watching support teams answer the same password and billing questions hundreds of times a week while complex cases waited. Victor leads product, customers and every pilot, with a focus on grounded answers and clear hand-offs to people.
Co-founder
Yetunde leads customer success at SupportPilot: onboarding support teams, shaping escalation rules with support managers and turning agent feedback into product changes.
CTO
Kelechi leads engineering: retrieval from approved knowledge, citations and confidence, ticket categorisation and routing, summaries and the AWS platform it runs on.
Progress & demos
SupportPilot is a proposed product in its building stage. A demo walks through knowledge upload, test conversations, ticket classification, drafts and the escalation queue. The MVP covers knowledge-base uploads, test conversations, ticket classification, response drafts and human escalation queues.
By Victor Adebayo, to answer routine questions well and give agents their time back.
Answers only from approved sources, citations on every reply, confidence thresholds and human escalation.
Knowledge base, test conversations, ticket classification and routing, drafts, escalation queue and analytics on fictional data.
Retrieval from approved knowledge, response drafting, ticket records, processing and APIs designed.
Knowledge-base upload and test conversations for one support team, with agents approving every draft at first.
FAQ
It is designed not to. Answers come only from approved sources and cite them; when no approved source covers a question, or confidence is low, it escalates to a person.
No. Teams can require approval for every draft, edit before sending, and set the escalation threshold and routing rules.
The planned product connects through APIs to common helpdesk and chat tools. The prototype runs on its own with fictional tickets.
Not yet. It is a proposed product; the support console is a prototype running in your browser on a fictional knowledge base.
Contact
Questions about the product, helpdesk integrations, data handling, or partnerships. Messages go directly to the founder.