It started as client work — and a wish we couldn’t grant.
A global strategy consultancy wanted personas they could act on — not documents. We couldn’t build that for a single client at a competitive price. Too expensive, back then.
Why we built Ask Sona
Better businesses are built by listening better.
Ask Sona brings customer perspective into everyday decisions — without turning every question into a long research project. This page explains how it works, what it’s built from, and why it isn’t just another AI tool.
Product lead“I’m pretty sure users want more automation here.”
Marketing“Our customers never read long emails. Trust me.”
Sales“The last client I talked to said the opposite…”
Founder“Let’s park it and discuss next sprint.”
Your customer’s perspectivePersona Response · decision support
“Automation isn’t my problem — I don’t trust the results enough to hand them off yet. Show me why I can.”
Grounded in interviewsTraceable
✓Grounded in research, not vibes✓Every answer traceable to its sources✓Your data is never used to train models✓Built by a research-led product agency
The shift
Look at the decision through your customer’s eyes — while it’s still open.
Customer-centricity is easy to say and hard to practice: when a real decision is on the table, the customer is rarely in the room. Drag the divider — same team, with and without a shared reference.
How it’s built
Three layers make a persona you can trust.
Understanding people needs a carefully designed lens. Ours has three layers — scroll, and each one explains itself.
Sona Typology Standard
Curated AI Models
Your Knowledge
Layer 1
The Sona Typology Standard
A curated research knowledge base: 80+ taxonomy layers, statistical guardrails, traceable reasoning.
Layer 2
Curated AI models
A hand-picked set of frontier models reasons through it all — a persona in under 60 seconds. Never trained on your inputs.
Layer 3
Your knowledge
A sentence, notes, or your own research — this is where the persona becomes truly yours. Templates cover the start.
Together
Personas grounded in reality that hold their ground.
Instead of a persona that just agrees with you.
Try it live
Ask a persona yourself.
Thomas Grabowski is a real Persona Template from the catalog. This is the same chat you get on the template page — no signup needed.
AI-generatedThomas GrabowskiGewinnt B2B-Neukunden und entwickelt Schlüsselkunden in Westdeutschland.
TG
Start a conversation with Thomas Grabowski to understand their perspective, motivations, and challenges.
Try asking about their goals, daily routine, or what motivates them.
No signup · meet all personas in the catalog →
What we stand for
Built on three commitments.
Our value
Research-Led
Grounded in years of UX, social science, and research practice.
Our promise
More Clarity
Helping teams move from guesswork to clearer customer understanding.
Our goal
Customer Priority
Making customer perspective part of everyday decisions.
Built by Dinghy
Ask Sona is made by Dinghy, a research-led product design agency — 10 years of UX design and research practice for product teams, agencies, and consultancies. More about Dinghy →
Fair questions
The things you’re probably wondering.
Is this just ChatGPT with a costume on?+–
No. Generic AI generates plausible text from the whole internet. Ask Sona responses are built on a defined, inspectable basis: your own inputs, the Sona Typology Standard, and your decision context — with visible Grounding under every answer, so you can check the basis instead of taking it on faith.
Is my data safe?+–
Your inputs are used to ground your personas — never to train AI models. Your research stays yours.
What do I need to get started?+–
Less than you think. A single sentence is enough to start; templates carry professional persona knowledge so there's no blank page. Notes, files, and research make everything richer — add them whenever you're ready.
Can I trust the answers for real decisions?+–
Trust them as decision support, not as oracle output. A Persona Response applies existing knowledge through a structured lens — it exposes assumptions and perspectives, and it tells you when its basis is thin. It doesn't replace research or predict every real person in your audience.
When teams understand people better, they build better things.
Customers feel seen. Businesses move with more confidence. Everyone wins. Our mission: bringing the voice of the customer into the room, so teams can build with more clarity and confidence.
Origin
We’re persona skeptics. That’s why we built this.
From a client wish we couldn’t grant, through a year of publishing about broken personas, to one PowerPoint slide that broke us — the whole story, with receipts.
Drag through the timeline →
Jun 2024 · Dinghy blog
Why Personas Suck
The post that started the public argument.
Read on dinghy.studio →So we went public with the problem.
Everyone wants personas; in the long run they never work out. We spent a year writing about why — a nine-piece series on what makes personas fail, and what would fix them.
Sep 2024 · Dinghy blog
What is a Persona?
Back to first principles.
Read on dinghy.studio →
Sep 2024 · Dinghy blog
Unsucking Personas — a step-by-step tutorial
The fix, as far as it went without software.
Read on dinghy.studio →
Sep 2024 · Video
EP14: Let's talk about User Personas
The podcast, same obsession.
Watch on YouTube →
Sep 2024 · Video
EP15: How do we make personas better?
Working through the fix, out loud.
Watch on YouTube →
Meanwhile · the day job
EnBW: a 2-week UX audit
The research practice kept running.
dinghy.studio →
Meanwhile · the day job
SumUp: continuous testing on a field-sales app
Listening to users as an ongoing practice.
dinghy.studio →Another persona project.
Research, interviews, synthesis — all done. Then came the moment every persona project ends with: distilling everything into the classic bento grid on a single PowerPoint slide.
“How do I cook all this down so it fits on one slide — and still works for the person using it?”
We let a language model try.
We defined the bento boxes as an output format for an LLM and ran the distillation with AI for the first time — and were amazed by how helpful a language model really is here.
A vibe-coding project became a prototype.
We showed it around. Our agency and freelancer friends were ecstatic about the premise — and wanted access.
Video
What If You Could Talk to Your Personas?
The question, asked on camera.
Watch on YouTube →
Jan 2026 · Dinghy blog
From research artefact to interactive artefact
The thinking, written down as it became real.
Read on dinghy.studio →Let’s make it a product.
Ask Sona: personas that are alive, grounded, and actionable — the thing we couldn’t build for one client, built for everyone.