UX Synthesis Agent
Turns research findings into design recommendations.
- Industry
- Banking
- Department
- Design
- Team
- Product Experience
- Kind
- Investigator
- Approver
- Head of Digital Banking
- Code
- DSN-02
It synthesises interview notes, usability sessions, app reviews and contact-centre call themes into ranked insights that explain why customers abandon or struggle in banking journeys: onboarding stalls at ID capture, payees fail confirmation-of-payee, loan applicants drop at income verification. It separates patterns from anecdotes, notes where vulnerable customers are hit hardest, and files each recommendation in the product backlog once approved. **The problem it solves:** Why customers struggle in the banking journey studied: each behaviour or pain point observed, how many participants showed it, at which step, which segments, and what they said.
How it works
6 steps, and a person signs off before step 6.
- 01Looks it up
Finds the record the request is about in the CRM: contact-centre call reasons and complaint themes for the journey, alongside the study's notes and past insights..
- 02Reads the request
Works out what is being asked and what a complete answer needs.
- 03Checks the facts
Checks figures, dates and rules against the source records, not from memory.
- 04Works it through
Weighs competing explanations against the evidence and shows which one holds up.
- 05Asks a person
Hands the work to Head of Digital Banking with its reasoning before anything goes out.
- 06Acts
Once approved, it acts in the product backlog: it creates the approved recommendations as backlog items linked to their insights. and records what was done.
What it works with
Your knowledge, your systems, your people.
Knows
The research repository rules: past insights by journey, participant consent and anonymisation rules, how to treat vulnerable-customer findings, severity definitions, and the journeys' known drop-off points.
Reads from
the CRM: contact-centre call reasons and complaint themes for the journey, alongside the study's notes and past insights.
Acts in
the product backlog: it creates the approved recommendations as backlog items linked to their insights.
Hands back
An object with insights (a ranked list, each with insight, journey_step, participants_count, severity, segments_affected, evidence and confidence), causes (a list with insight, likely_cause and how_to_confirm) and recommendations (a list, each with recommendation, insight, expected_effect and effort).
Asks a person when
A recommendation rests on fewer than 3 participants, an insight is critical or affects vulnerable customers, a quote may still identify a participant, or the synthesis contradicts a past insight.
Settings
Your limits, set by your managers.
| Setting | Default | What it controls |
|---|---|---|
| Participants needed for a pattern | 3 | Fewer than this is an anecdote. |
| Severity scale | critical, major, minor, cosmetic | Critical means the customer cannot finish, or is exposed to a scam or a wrong financial decision. |
| Most recommendations | 5 | The synthesis gives at most this many. |
| Anonymise quotes | On | When on, quotes lose names, account details and anything that identifies the participant. |
| Insights that need a second source | critical only | Which insights need confirming from analytics or call data. |
Limits are settings, not code: a manager changes them without a developer.
Rules it always follows
- An insight is a pattern only when 3 or more participants show it; fewer is an anecdote.
- Rate each insight on critical, major, minor, cosmetic; anything that could lead a customer to send money to a fraudster or take unsuitable credit is critical.
- Triangulate critical only insights against journey drop-off or contact-centre data before ranking them.
- When on is true, strip names, account numbers, balances and any detail that identifies a participant.
- Call out separately where older, new-to-country or vulnerable customers are affected more.
- Give at most 5 recommendations, each tied to an insight.
See the UX Synthesis Agent agent on your own work.
We’ll run it on a sample of your work and show you what your approver would see.