Business and Sales · News · 🇺🇸 United States · via agentusecases.ai
Business lunch in Houston chosen by an agent
Nielsen Norman Group researchers watched the agent run 10 searches and propose three restaurants.
- Profession
- UX Researcher
- Tools
- ChatGPT Agent
- Source
- nngroup.com ↗
- Country
- 🇺🇸 United States
- Found via
- agentusecases.ai (Alex Finn) ↗
- Report date
- no date
- Verified on
- October 25, 2026
What they did
The tester gave a short request: a business lunch near an address in Houston. The agent ran 10 searches across 96 sources and suggested three places; once Brennan's was chosen, it asked for the number of people and went ahead with the reservation.
How it went
NN/g evaluates the flow and points out usability problems and slowness.
Good to know
Short requests work, but the agent is faster with clear criteria.
The prompt
Suggest 3 restaurants for a business lunch near [address] on [day], for [people]: quiet atmosphere, up to [amount] per person, reservation available at [time]. Explain your choice.
Replace what is in [brackets]. Prompt written by us from the account, not a literal copy of the source. Do not paste other people’s personal data (privacy laws such as LGPD and GDPR apply).
Summary in our own words. Results as stated in the original account; we have not tested or audited them. Read the source ↗ Case found via agentusecases.ai (Alex Finn); text and prompt rewritten by us. Found a mistake?
Similar cases
01Competitor analysis as a presentation and a spreadsheet
The author asked for a presentation on the UK's 5 biggest coffee subscriptions and a table of differentiators.
02Disputing a contractor's prices on a building project
A project owner gave the AI every item, the purchase order and the emails; it found charges he hadn't noticed.
03Paying an overseas supplier with an agent and human approval
An agent checks balances in several currencies and prepares the payment; a person reviews and approves it.
04Canceling insurance by phone and recovering money with an agent
Forbes reports on people who used an agent to call the insurer and cancel a policy with a refund.
05Selling a VR headset on Marketplace with an agent
With access to Facebook, the agent listed the Quest 3 and talked to buyers within a minimum price.
06University recovers US$100,000 with a treasury skill
Cornell's finance and AI teams built a skill to identify and chase unidentified payments.
07Pre-qualify leads before they reach a salesperson
Futurepedia built an agent that researches each company, writes a fit summary and forwards the best leads.
08Bookkeeping for a Swedish company with agents
A cofounder had agents read receipts, suggest accounts and VAT, and prepare entries for approval.