What problem does it solve?
When not every branch offers every service, the question "where do I go" is hard to answer. Opening hours vary, and specialists such as mortgage or business bankers may work by appointment in only a few locations. Customers who turn up at the wrong place, at the wrong time or without the right documents waste a trip, and staff meet them unprepared.
Booking by phone ties up contact centre time for a simple task, and web booking forms often do not know which specialist handles which need. The same pattern appears wherever physical visits need to be planned: a retailer's service desk, a clinic, a property viewing.
How does it work?
- Understand the need. The agent asks what the visit is for (a mortgage, a business account, a cash deposit, help with the app) because that decides where and with whom.
- Find the right location. It searches location data for the nearest branch or ATM that offers that service, with opening hours, accessibility details and any temporary closures.
- Offer alternatives. Where a video call, the app or a phone call would serve the customer better, it says so, and books that instead if the customer prefers.
- Book the slot. It checks the specialist calendars, offers real times and books the appointment through the scheduling system, with a confirmation and a reminder.
- Prepare the visit. It records the reason for the visit and tells the customer which documents to bring, so the specialist is ready.
- Pass special needs to a person. Accessibility requests and anything unusual go to staff rather than being handled by a rule.
- Audience
- Customer facing
- Autonomy
- Autonomous
- Adoption
- Early adopters
- Channels
- Web chat, Mobile app, WhatsApp, Phone and voice
What is it worth?
Benchmarks are computed from the public deployments below: one data point per organization per KPI, with who made each claim.
| KPI | Median | Reported range | Data points | Claimed by |
|---|---|---|---|---|
| Interactions handled | Not pooled | at least 3 billion | 1 | 1 organization |
Value drivers: Customer experience, Lower cost to serve, Revenue growth, Inclusion and access.
Indicative value
A bank that books 100,000 branch and specialist appointments a year
USD 72,000 to USD 480,000
Booking handling cost avoided per year
How this is calculated
Formula: appointments * agentShare * minutesPerBooking * costPerMinute. The low scenario uses every low input, the high scenario every high input.
| Input | Low | High | Basis |
|---|---|---|---|
| Appointments booked per year appointments, appointments per year | 100,000 | 100,000 | The reference bank. |
| Share of bookings moved from phone and branch staff to the agent agentShare, fraction of appointments | 0.3 | 0.6 | Editorial assumption, replace with your own channel mix. |
| Staff minutes per booking handled by a person minutesPerBooking, minutes per booking | 4 | 8 | Editorial assumption covering the call, calendar lookup and confirmation. |
| Fully loaded cost of a staff minute costPerMinute, USD per minute | 0.6 | 1 | Editorial assumption, replace with your own fully loaded cost. |
What it leaves out: Counts booking time only. It leaves out fewer wasted visits and no shows, better prepared appointments that convert more often, the cost of the AI and the calendar integration.
Who already uses it?
4 public deployments, strongest evidence first. Grades: A regulator or audit, B the organization itself, C vendor case study, D anonymous or estimate.
Bank of America
United States · Banking · 2025
Erica, launched in 2018, is Bank of America's virtual financial assistant in its Mobile Banking app. Beyond answering questions it delivers proactive, personalized insights: BankAmeriDeals cash back deals based on the client's spending, where balances are trending over the next seven days and eligibility for the Preferred Rewards program. It also gives guidance on investment topics for Merrill clients and hands off to people by scheduling appointments. The bank reports that clients have received and interacted with more than 1.7 billion of these insights, and that most users find the information they need, which it links to lower call centre volume. Bank of America says Erica selects answers from a predefined set and does not use generative AI or large language models.
- Users served: about 50 million, since launch in 2018, as of August 2025
"assisting nearly 50 million users since launch, surpassing 3 billion client interactions, and now averaging more than 58 million interactions per month"
Claimed by: organization - Interactions handled: at least 3 billion, client interactions since launch in 2018, as of August 2025
"surpassing 3 billion client interactions"
Claimed by: organization
Best Buy
United States · Retail and ecommerce · 2024
In April 2024 Best Buy announced, with Google Cloud and Accenture, a generative AI virtual assistant for BestBuy.com, its app and its customer support line, expected to launch in late summer 2024, to help customers troubleshoot product issues, change order delivery and scheduling, and manage subscriptions and memberships. Google Cloud reported in its April 2026 list that Best Buy now guides shoppers through technical specifications, issue resolution and appointment scheduling autonomously, using Agent Assist powered by Gemini Enterprise for Customer Experience. It is a retail example of the same job a bank has when it books a branch or specialist appointment. No outcome figures for the assistant were published.
No outcome disclosed.
MOGUL.sg
Singapore · Real estate · 2025
MOGUL.sg, a Singapore property platform, launched MAIA in February 2025: an AI agent on WhatsApp that searches listings and books viewing appointments, built with Vertex AI, Gemini and the Google Maps API. Location lookup plus booking in one conversation is the same pattern a bank needs for "find my nearest branch and book me in".
No outcome disclosed.
Hemominas
Brazil · Healthcare · 2024
Hemominas, Brazil's largest blood bank, partnered with Xertica to develop an omnichannel chatbot for donor search and scheduling of donations. Google Cloud describes the expected impact in terms of potential lives saved, which is a projection, not a measured result, so the record is kept at the announced stage.
No outcome disclosed.
How do you implement it?
A model agnostic playbook: what to prepare, the order to build in, and what goes wrong.
Data you need
- Branch and ATM locations with opening hours, services, accessibility details and closures
- A mapping from visit reasons to services and specialist roles
- Specialist calendars and booking rules
- Document checklists per visit reason
Systems to integrate
- Scheduling or appointment system
- Location data or maps service
- CRM for visit reasons and follow up
- Messaging for confirmations and reminders
- Contact centre for handover
Complexity: Low
Mostly reads location data and writes to one scheduling system. The effort goes into clean, current branch and service data, and a calendar integration that reflects real specialist availability.
- 1
Clean the location data
Make one owned source for every location's hours, services, accessibility and closures. A branch assistant can only be as current as the branch data behind it, so fix this first.
- 2
Map reasons to specialists
Write down which visit reasons need which service and role, and which can be handled by video or in the app instead.
- 3
Integrate the real calendar
Offer only slots the scheduling system confirms, and write the booking back with the reason for the visit.
- 4
Add reminders and rescheduling
Send a confirmation and a reminder with the document checklist, and let the customer move or cancel the appointment in the same conversation.
- 5
Measure kept appointments
Track bookings, no shows and outcomes by channel so you can see whether guided booking actually improves visits.
Guardrails
- Services and hours only from the owned location data, never inferred by the model
- Bookings only into slots the scheduling system confirms
- Accessibility needs and special requests passed to staff, not handled by rule
- Personal data in bookings disclosed, minimised and stored in region
KPIs to instrument
- Bookings completed by the agent and share of all bookings
- No show rate by booking channel
- Wrong location or service reports from customers and staff
- Handover rate and reasons
- Customer satisfaction after the visit
Human in the loop
Branch staff own the appointment once booked and handle accessibility and special requests. Location data owners keep hours and services current, and a sample of conversations is reviewed monthly for wrong locations or services.
Common failure modes
- Promising a service the branch does not offer
- The customer arrives and cannot be helped. Keep service lists owned and current, and answer only from them.
- Phantom slots
- The agent offers times that are no longer free. Read and write the live calendar.
- Accessibility handled by rule
- A wheelchair user is sent to a branch with steps. Pass accessibility needs to a person and keep accessibility data current.
What are the risks and rules?
EU AI Act
Limited risk (transparency)
Article 50(1): people must be told they are interacting with an AI system unless that is obvious. Finding locations and booking appointments does not fall under any Annex III category. If a healthcare version starts to triage patients by urgency, or a public body uses it to decide eligibility for a public service, reassess it against Annex III point 5.
Rules that apply
Guidance
- Directive (EU) 2019/882 on the accessibility requirements for products and services (European Union, Europe). The European Accessibility Act sets accessibility requirements for consumer banking services and ecommerce services, so digital booking channels in those sectors need to meet them.
Controls to put in place
- AI disclosure and a clear route to a person
- Owned, dated location and service data with change control
- Consent and retention rules for booking data
- Monitoring of wrong location reports and no shows
Frequently asked questions
- Is a branch booking agent worth it when most banking is digital?
- It depends on your visit mix. Where branch visits are mostly for specialist needs such as mortgages or business banking, sending the customer to the right person, prepared, saves a wasted trip on both sides. Bank of America, whose Erica assistant has served nearly 50 million users since 2018, uses it to schedule appointments as a handoff to human service.
- Who already books appointments with AI agents?
- Bank of America uses Erica to schedule appointments as a handoff to high touch service channels. Outside banking, Google Cloud reports that Best Buy uses AI to guide shoppers through appointment scheduling, and MOGUL.sg books property viewings through a WhatsApp agent. In healthcare, Hemominas in Brazil worked with Xertica on a chatbot for donor search and scheduling. Public outcome data specific to booking is scarce.
- What should you get right first?
- Stale location data is the risk to plan for first. Wrong opening hours or a service a branch no longer offers send the customer on a wasted trip however good the conversation is, so fix data ownership before tuning the agent.
How to cite this page
Blits.ai AI Use Case Library, "AI agent for branch finding and appointment booking", last verified 27 September 2026, https://www.blits.ai/ai-use-cases/branch-and-appointment-booking-agent. Licensed under CC BY 4.0. Method: how we verify use cases.
Changelog
- 27 September 2026: First published