A Radley host presenting a herringbone tote on the shop floor during a live show.
All customer stories
Handbags

Radley London

The London-store answer, on every product page.

Nineteen live shows, hosted from the shop floor between March and December 2024. The recordings stayed shoppable afterwards.

Where it runs
On the product page, on every visit
Industry
Handbags and accessories

When a shopper gets an answer, they buy.

The problem

A handbag is decided on the details a product page lists but does not answer. Will it hold what I carry. How long is the strap on me. Does that colour age well. In a London store somebody settles all three in a sentence, and the sale happens in the same minute.

Online the question usually goes nowhere. It is not big enough to open a support ticket over and not small enough to guess at, so the shopper closes the tab — and the store records a bounce rather than a question it could have answered.

What changed

The AI Sales Agent answers on the product page itself, from Radley's own catalogue: what is in stock, what the material does, what goes with what, and where to go next. It is built to sell rather than to deflect — the point of the answer is the purchase, not the closing of a ticket — and it does not clock off at six.

Behind it Radley also runs live events and keeps the recordings, so a shopper who wants more than an answer has somewhere to go. A browsing session can become a conversation, and a conversation can become an order.

The result

Across more than 8,000 AI-assisted sessions, and thousands of live chats beside them, Radley London shoppers who engage buy at roughly 10% — about one in ten.

A typical online store converts 1.7–2% of its visitors overall. That makes an assisted session worth close to five ordinary ones, at a volume no shop floor could be staffed to match.

-
-
.
,
1
2
3
4
5
6
7
8
9
0
#
£
%
+
K
B
M
X
-
$
-
-
.
,
1
2
3
4
5
6
7
8
9
0
#
£
%
+
K
B
M
X
-
$
-
-
.
,
1
2
3
4
5
6
7
8
9
0
#
£
%
+
K
B
M
X
-
$
-
-
.
,
1
2
3
4
5
6
7
8
9
0
#
£
%
+
K
B
M
X
-
$
-
-
.
,
1
2
3
4
5
6
7
8
9
0
#
£
%
+
K
B
M
X
-
$
+
AI-assisted shopping sessions
~
-
-
.
,
1
2
3
4
5
6
7
8
9
0
#
£
%
+
K
B
M
X
-
$
-
-
.
,
1
2
3
4
5
6
7
8
9
0
#
£
%
+
K
B
M
X
-
$
%
Of assisted sessions end in a purchase
~
-
-
.
,
1
2
3
4
5
6
7
8
9
0
#
£
%
+
K
B
M
X
-
$
×
The rate of a typical storefront

Baseline: typical online stores convert around 1.7–2%. Sources: IRP Commerce and Smart Insights, 2025. Immerss figures are conversation-level and attributed to Radley London.