An AI insurance assistant that handles service, not advice
We build an AI assistant for your website, text line or phone that answers routine questions (office hours, how to pay, how to get ID cards or a certificate, what to do after an accident) from your agency's approved answers, takes policy change and certificate requests into the right workflow and hands any question about coverage, claims decisions or advice to a licensed person. It discloses that it is automated. Projects typically cost $2,500 to $7,500 and take five to eight weeks.
The problem
What an insurance chatbot should and should not do
CSRs spend much of the day on the same requests: send my ID cards, add a vehicle, how do I pay, I need a certificate for a job. Each is simple, but together they push renewal and claims work into the evening.
An insurance AI chatbot can take that load only if it is narrow. It cannot tell someone whether they are covered, recommend a policy or quote a price it did not get from a carrier. It must answer from approved wording, collect clean requests and pass anything else to a licensed agent.
What we build
What our AI insurance assistant includes
Approved answer library
Office hours, payment options, carrier claims lines, ID card and certificate steps written and approved by the agency.
Website chat and text line
A chat widget on your site and, optionally, the same assistant on SMS or WhatsApp.
Clean service requests
Policy changes, certificate requests and document requests collected in a structured format for CSRs.
Carrier claims lines
After an accident or loss, the right carrier claims number and the agency's FNOL form.
Licensed hand-off
Questions about coverage, pricing, recommendations or complaints go straight to a person.
Disclosure and transcripts
It says it is automated, offers a person at any time and logs every conversation.
Who does what
Automation does the chasing, people make the calls.
-
01
Automated
Question asked
A policyholder asks on the website, by text or by phone.
-
02
Automated
Answer from your library
The assistant answers only from approved wording.
-
03
Automated
Request captured
Change, certificate or document requests are structured and sent to the right workflow.
-
04
CSR or account manager review
CSR completes
CSRs process requests and reply to anything the assistant could not answer.
-
05
Licensed agent or principal decides
Coverage advice
Anything about coverage, pricing or recommendations goes to a licensed agent.
How it behaves
How the AI insurance assistant behaves
What the policyholder sees
- A clear note that they are talking to an automated assistant
- Quick answers from the agency's own wording
- A licensed person whenever they ask about coverage or price
What CSRs see
- Structured change and certificate requests
- Hand-offs with the full conversation
- A weekly list of unanswered questions
What the automation does
- Answers only from approved wording
- Refuses coverage, pricing and advice questions and hands over
- Logs every conversation
Edge cases and fallbacks
- Question not in the library
- It says it does not know, offers a person and logs the question.
- Client asks if a loss is covered
- No answer; hand-off to a licensed agent and the carrier claims line.
- Complaint or distress
- Immediate hand-off to the named manager.
- AI provider outage
- The widget falls back to a contact form and the text line to a holding reply.
Architecture
An AI service assistant for an agency at a glance.
- Input
- Logic and services
- AI step
- Data
- Output
- People
How it flows
- 1 A policyholder asks a question and is told they are talking to an automated assistant.
- 2 The assistant service retrieves approved answers and passes them, with fixed rules, to the model.
- 3 Routine questions are answered and service requests are structured for CSRs.
- 4 Coverage, pricing and complaints go to licensed staff, and every conversation is logged.
Project planner
How the project runs, phase by phase.
AI service assistant for policyholders project plan
Typical time to launch: about 6 weeks, then support
-
01
Discovery
About 1 week
Collect real questions from calls, emails and texts; agree what the assistant must never answer.
-
02
Answer library
About 1 week
Write and approve answers, hand-off rules and disclosure wording.
-
03
Build
About 2 weeks
Widget or text line, retrieval, request capture, hand-offs and logging.
-
04
Test
1 to 2 weeks
Real, tricky and coverage-bait prompts; licensed staff sign off.
-
05
Launch
2 to 3 days
Soft launch after hours, then all day.
-
06
Support
Ongoing
Weekly transcript review at first, then monthly library updates.
What it runs on
Real products only, connected through the access each vendor provides. See all integrations
Cost estimate
What it costs, line by line.
-
Discovery and answer library
Real questions collected, answers written and approved.
$500 – $1,200
-
Build
Assistant service, widget or text line, retrieval and hand-offs.
$1,200 – $4,000
-
Integrations
CRM or AMS request routes, alerts and messaging.
$300 – $1,000
-
Testing and evaluation set
Real, tricky and coverage-bait prompts before launch.
$300 – $900
-
Launch and training
Soft launch, transcript review and team walkthrough.
$200 – $400
-
Total project estimate
$2,500 – $7,500
Optional monthly care after launch is quoted with your proposal. Fixed quote after a free call. A voice line costs more to build and run than web chat. Model usage and messaging are billed to your accounts directly.
Monthly running costs you pay directly
Estimates from the vendors' published pricing, checked November 2026. Billed to you by each vendor, never marked up.
- AI model usage: Billed by the provider per use; short service answers are usually a small monthly amount.
- Messaging or voice: Twilio at its published per-message rates if you add a text line; a voice line costs more to run than web chat.
- Hosting: n8n Cloud from about €20 a month, or your own server.
What pushes cost up
- Systems without an API, or with integration access that must be requested from the vendor
- Writing data back into your core system rather than only reading from it
- Several locations, brands or very different processes to support
- Custom apps, dashboards or a voice line on top of the workflows
What keeps cost down
- One location and one clear process to start with
- Cloud software with a documented API
- Using tools you already pay for instead of adding new ones
- Approved wording and sample data ready at the start
Privacy and consent
Rules we design for, with the sources.
EU: tell people they are talking to AI
Article 50 of the EU AI Act requires AI systems that interact with people to inform them they are interacting with an AI system, from 2 August 2026.
Source: artificialintelligenceact.euUS: insurer AI governance
The NAIC model bulletin on insurers' use of AI, adopted December 2023, expects governance of AI systems. Agencies working with carriers should expect questions about AI tools, so ours stays out of decisions.
Source: content.naic.orgUK: Consumer Duty
The FCA Consumer Duty requires firms to put retail customers' interests first, so the assistant offers a person at any time and avoids steering.
Source: www.fca.org.ukFree tools
Try these first, free.
How we deliver
Agile delivery, in 1 to 2 week sprints.
Every sprint
- 1 Sprint planning. At the start of each sprint we agree which backlog items we will finish, each with written acceptance criteria you can check.
- 2 Daily stand-up. A short internal check-in on progress and blockers. You get a written update in your Slack or Teams channel, not another meeting.
- 3 Sprint review and demo. At the end of each sprint we demo what was built on test data or in a shadow run, and you accept it or ask for changes.
- 4 Retrospective. The team looks at what went well and what to change in the next sprint, and we share anything that affects you.
How scope stays under control
- Product backlog
- Every feature, fix and question in one prioritised list you can see at any time.
- Acceptance criteria
- Plain-language conditions for "done" on each item, agreed before work starts.
- Change requests
- New ideas go into the backlog with an estimate. Nothing changes scope or price without your written approval.
- User acceptance testing (UAT)
- Before launch your team tests the real workflow against the acceptance criteria and signs off.
Tools we use
Project manager / Scrum master
Runs the sprints, keeps the backlog and timeline honest, and is your single point of contact.
Business analyst
Maps how the work runs today, writes the workflow specification and acceptance criteria with your team.
Automation engineer
Builds the n8n workflows, integrations, AI steps and error handling.
Backend developer
Writes custom code where a no-code step is not enough: APIs, data sync, secure storage.
Frontend / mobile developer
Builds forms, booking pages, dashboards, web apps or Android apps when the project needs them.
QA engineer
Tests every workflow against the acceptance criteria, including edge cases and failure paths.
DevOps engineer
Sets up hosting, access, backups, monitoring and alerts, and the deployment pipeline.
Who does what at each phase
- Free call
- Project manager, business analyst
- Understand the problem, check it is worth automating and outline a rough range.
- Discovery
- Business analyst, automation engineer
- Map the current process, systems and data, and agree success measures.
- Design
- Business analyst, automation engineer, backend developer
- Write the workflow specification, data flows, compliance notes and backlog with acceptance criteria.
- Sprints
- Automation engineer, developers, QA
- Build, test and demo in 1 to 2 week sprints; you accept or adjust after each demo.
- UAT and launch
- QA, DevOps, project manager
- Your team tests and signs off, we switch on, monitor closely and train the team.
- Support
- Automation engineer, DevOps
- Stabilisation window, then optional monthly care with monitoring and changes.
A point of contact
One person who can answer questions and approve decisions, with about an hour a week during the project.
Access
Admin or API access to the systems involved, created as separate accounts with the least access needed.
Sample data
Real but non-sensitive records (or de-identified ones) so we can design and test against reality.
Approved wording
Message templates, policies and consent wording (for SMS, your opt-in and opt-out text) reviewed by you.
Vendor agreements
Signed agreements with the software vendors that will handle personal data, such as a BAA or processor contract.
Time for testing
A few hours from the people who will use the system, for the demos and user acceptance testing.
Working system
Workflows, integrations and any apps running in your accounts, under your keys.
Documentation
Workflow specification, data-flow map, run book and exported workflow files.
Training
A walkthrough for the team and a short recorded video for new starters.
Handover
All access, credentials and code handed over; you can keep us or take it in-house.
Stabilisation window
Fixes for anything that does not meet the agreed acceptance criteria after launch.
Support options after launch
No plan
You run it yourselves after the stabilisation window and contact us when you need a change.
Monthly care
Monitoring, error alerts, small changes and updates each month, with an agreed response time.
Priority care
Faster response times for business-critical workflows and a monthly review call.
How a project moves through our team
- Input
- Logic and services
- People
- Output
- Data
How it flows
- 1 A free call with the project manager and business analyst to understand the problem and whether it is worth building.
- 2 Discovery and design produce a written specification, data-flow map and a backlog with acceptance criteria.
- 3 The team builds in 1 to 2 week sprints; you review a demo at the end of each and accept or adjust.
- 4 Your team runs user acceptance testing, we launch, then monitor and improve under a support plan.
FAQ
AI service assistant for policyholders questions, answered.
Answer routine service questions, collect policy change and certificate requests, point clients to the right claims line and hand everything else to a person. It never advises on coverage.
Ours does not. Selling, quoting and recommending cover require a licensed person, and prices must come from carriers. The assistant collects details for a producer instead.
Yes. It says so at the start and offers a person at any time. The EU AI Act requires this from 2 August 2026, and we do it everywhere.
It collects a complete change request, such as a new vehicle with VIN and driver, and sends it to a CSR to submit to the carrier.
It gives the carrier's claims line, the agency's claim form and emergency steps, and flags the conversation for staff in the morning.
Transcript logs, a weekly list of unanswered questions, the share of requests resolved without a hand-off and complaints, if any, about the assistant.
Free 20-minute call
Want this for your insurance broker?
Book a free 20-minute call. We will look at how the work runs today, tell you honestly whether it is worth building and give you a fixed quote if it is.
More for insurance brokers
All servicesInsurance agency CRM & sales pipeline automation
Every quote request, referral and cross-sell opportunity in one insurance agency CRM pipeline linked to your AMS, with fast first replies, follow-up tasks and a weekly report by source and line.
From $1,000 · About 4 weeks Details
Renewal pipeline & client information requests
Build the renewal pipeline from AMS expiry dates by line, send information requests months ahead, chase replies, highlight changes since last year and give account managers time to market the risk.
From $1,000 · About 4 weeks Details
First notice of loss intake & claim updates
Capture the facts of a loss in a structured form, pass the notice to the carrier the same day, chase photos and documents and keep clients updated, with coverage questions left to people.
From $1,000 · About 4 weeks Details