01 · Discover & Design

Discover & Design

Understand what customers need before you invest — with evidence, not instinct.

Stage: Understand

Your problem

You decide what to sell or improve next on scattered feedback and gut feel — then invest before verifying anyone will pay.

The feedback that could answer the question already exists in your reviews, emails and calls. Nobody has time to read it all.

Why it matters

Research with product teams shows AI can meaningfully compress discovery work on specific, well-defined tasks — grouping, summarising, drafting test material. The judgment calls — sample bias, pricing, whether to invest — stay human.

Where this does not apply: If you have fewer than a dozen pieces of customer feedback, there is no pattern to mine yet. Talk to five customers first; AI cannot replace that.

Before, then with AI

Customer feedback becomes an evidence list

A business deciding what to offer or improve next, sitting on reviews, emails and call notes.

Illustrative example — not a live customer result

Before

  1. Feedback arrives in reviews, emails and calls, then scatters

    Staff

  2. The owner decides what to sell next mostly on instinct

    Owner

  3. Money goes into a new offer before anyone has verified demand

    Owner

With AI

  1. AI groups permissioned feedback into named problems, each linked back to the original quote

    AI prepares

    Output: Problem list with source lines

  2. A person checks for sample bias and picks one problem worth testing

    Human reviews

  3. A draft offer is tested with a real quote to a real customer before any build

    Human only

    Output: One priced test, real response

Where this fails: AI can invent patterns from thin data. Never let it paraphrase customers into quotes they did not say; every line keeps its link to the original.

Flagship use case

Turn permissioned customer feedback into a tested problem list

For businesses deciding what to offer or improve next: AI groups your existing reviews, emails and call notes into named problems, each linked to the original quote — then you test one with a real priced offer.

Open the full use case →

Other parts of this area

  • Listen at scale
  • Watch the market
  • Test before you build
  • Package & price

How to start

  1. Gather 6 months of feedback you have permission to analyse

    Owner · Permissioned corpus

  2. Run a small grouping trial and spot-check the links to originals

    Owner · Verified problem list

  3. Pick one problem backed by distinct customers

    Owner · One chosen bet

  4. A person approves the test offer and its price

    Owner · One real priced test

  5. Record responses; every learning updates the problem list

    Owner · Decision note

Tools to look at first

Can your current tools do this? These are category candidates, not a recommendation to buy.

ChatGPT (business plans)

General assistant

Good for: Feedback grouping, drafting and analysis tasks with pasted context · Testing a use case before buying anything specialised

Not for: Anything with customer personal data unless your plan’s data terms cover it · Unattended or automated sending of any kind

Works with: Standalone; check the data-handling terms of the specific plan before business use.

Official pricing · Data handling

Alternatives: Claude, Gemini, Microsoft Copilot · Disclosure: No commercial relationship. Data-handling and training settings differ by plan and change over time — verify on the official page, not here. · Reviewed 2026-10-19

Google Workspace + Gemini

Documents & knowledge

Good for: Shared Drive as a lightweight Company Brain · Drafting in Gmail/Docs where your team already works

Not for: Deep Microsoft-ecosystem shops · Line-of-business accounting tasks

Works with: Best if your business already runs on Gmail, Docs and Drive.

Official pricing · Data handling

Alternatives: Microsoft 365, Notion · Disclosure: No commercial relationship. Category candidate, not a verified fit for every business. · Reviewed 2026-10-19

Notion + Notion AI

Knowledge base

Good for: A structured Company Brain with owners and last-updated dates · Role cards, rules pages and checklists in one place

Not for: Heavy document/email workflows · Financial records

Works with: Works alongside either Microsoft or Google stacks as the knowledge layer.

Official pricing · Data handling

Alternatives: SharePoint, Google Drive + Docs · Disclosure: No commercial relationship. Category candidate, not a verified fit for every business. · Reviewed 2026-10-19

Evidence

External research

One person plus AI matched some team outputs

In a field experiment with product development teams, individuals working with AI matched some outputs that previously needed a small team — on specific, well-defined tasks.

What this does not show: A multinational product organisation; task-specific findings.

Source: Harvard Business School — The cybernetic teammate (field experiment with P&G product teams) · Independent academic study at one multinational; not an SME setting. · Reviewed 2026-10-19

Boteam self-experiment

We run this site the way we describe

This website is built and maintained by AI agents drafting code and copy, with a person reviewing and approving every change. The commit history and the published guides are the record.

What this does not show: Our own operation — one data point about our workflow, not a customer result.

Source: Boteam — Boteam self-experiment: rebuilding boteam.ai with AI agents under human review · Our own experiment, reported as such. Not a customer result. · Reviewed 2026-10-19

Do one thing

Do one thing: mine the feedback you already have

0/5 done · saved on this device

The prompt includes placeholders — never paste customer personal data or company secrets into a tool that has not passed your data check.

Next

What would make this useful for your business?

Ask a question or report a factual issue. Mason reads every submission; it shapes what we research and publish next.

AI for Customer Research and Service Design · Boteam