Coach Report — built for Skin + Me

How a Lorikeet AI agent handles the messages your customers send Skin + Me every day.

We ran six representative conversations against a Skin + Me-trained Lorikeet agent — pause, purging, severe reaction, prescription change, address update, layering with other skincare. Here's what came back.

Executive summary

The agent reads each customer's situation accurately — a holiday-pauser, a worried week-three customer, a severe reaction, a confident customer asking for stronger actives — and shapes its response to fit. It pulls policy from the knowledge base verbatim, never invents a clinical answer, and routes anything prescription-related back to the dermatology team via the in-app consultation.

The clinical-boundary discipline is the standout. Mild side effects → reassurance plus self-management tips. Severe reaction → stop, GP/NHS 111/A&E, log to derma team after. Dose change request → polite refusal, explanation of how prescription changes actually work, link to the right channel. The agent never plays clinician — and never leaves the customer without a next step.

6/6
scenarios resolved cleanly
~20s
avg time to a confident answer
0
prescription claims invented

The six conversations

01 Pause subscription for a 3-week holiday Strong
"I'm going on holiday for 3 weeks at the end of the month. Can I pause my Daily Doser and how do I do it?"

What worked

Where to tighten

With a subscription tool wired in, the agent could pause directly inside the chat — collapsing this from a 3-click self-serve into a one-message resolution.

02 Week 3 — more spots, dryness, flaking Excellent
"I'm three weeks into my Daily Doser and getting more spots than before, plus my skin is dry and flaky. Is this normal or should I stop?"

What worked

Where to tighten

This is the conversation that prevents avoidable churn at week 3-4. Pair it with a tagged follow-up at day 21 to check in proactively, and the LTV impact compounds.

03 Severe reaction — blistering, swollen lips Excellent
"Help — I used the Daily Doser last night and woke up with blistering around my mouth and swollen lips. It's really sore."

What worked

Where to tighten

Gold-standard handling. Add a sandbox tag that pre-flags this case for the dermatology team's morning queue, so the prescription review is already in progress when the customer follows up with photos.

04 Customer asking for a stronger retinol dose Excellent
"Can you bump up my retinol percentage? I've been on the current dose for 6 weeks and my skin is fine."

What worked

Where to tighten

This is the prescription-safety guardrail working as intended. The agent could pre-fill the consultation message with the customer's request, so they only need to add the photo.

05 Address change before next ship Strong
"I'm moving flats next Friday. My next Daily Doser ships on the 22nd. How do I update my address?"

What worked

Where to tighten

With an account tool, the agent could update the address directly in chat after a quick verification — saving the customer the round-trip.

06 Layering with The Ordinary vitamin C and acids Good
"I bought a vitamin C serum from The Ordinary — can I use it in the morning if I'm using my Daily Doser at night? And what about exfoliating acids?"

What worked

Where to tighten

Tied to the customer's actual prescription stage and skin profile, this answer becomes personalised: "you're 8 weeks in on a tretinoin formula — wait until maintenance for AHAs."

What this tells us about Skin + Me-shaped support

What's already strong

  • Clinical-boundary discipline. Severe reaction → stop and seek care. Dose change request → refused with the right next step. No improvised clinical advice.
  • Adjustment-effect education. Week 3 anxiety is the conversation that decides retention — and the agent handles it with reassurance plus self-management plus a clear escalation tripwire.
  • Brand voice. UK English ("personalised", "moisturiser"), warm but brief, two-or-three-sentence answers. On-brand without trying.
  • Knowledge fidelity. Every answer ties back to a knowledge article — pause/cancel rules, side-effect spectrum, layering guidance, prescription-change pathway.

Where to invest next

  • Account-action surface. Pause, skip, change address, send-early — all resolvable in chat with a subscription tool, not a deflection.
  • Day-21 proactive check-in. Tag the purging conversation, follow up automatically — turn a reactive moment into a retention lever.
  • Personalisation by treatment stage. Layering, dose, side-effect tolerance answers become precise once the agent knows the customer's prescription and weeks-in.
  • Pre-filled dermatology consultation. When the agent recommends "send a message and a photo", pre-fill the message body so the customer just attaches the photo.

What a full Skin + Me deployment unlocks

Six conversations in, the pattern is clear: the agent already handles the policy and clinical-boundary load that breaks generic chatbots — pause rules, purging vs reaction, dose change refusals, urgent care escalation. Connect it to the subscription system, the dermatology consultation queue, and the customer's prescription stage, and the same agent goes from "answers the question correctly" to "moves the case forward in one message."

Skin + Me's customer mix — patients on prescription actives, with retention hinging on weeks 2–6 and clinical safety throughout — is exactly what Lorikeet is built for. We'd love to show you the production version.