AI Offer Creation

UI Design | UX Research

Cate is heycater!’s AI-powered catering assistant. Instead of filling out forms or waiting for Sales, customers can simply chat with Cate to describe their event. Compared to other projects, the design part was relatively small, the real effort went into the tech team building the data foundations, AI logic and integrations that make the experience work.

View hi-fi prototype

Discover

What was the problem?

Creating offers is a slow and manual process, Sales often had to chase missing details and rework requests, which made it inefficient. Customers had little flexibility to edit offers themselves or explore options quickly, leading to frustration. At the same time, corporate clients required clear, reliable offers with PDFs, tax details, and comparison tables, but what they received was often inconsistent. The result was long turnaround times, wasted effort, and a poor overall experience for both Sales and customers.

Research

We tested Cate both internally with colleagues and externally with 30 participants on Lyssna. This mix allowed us to validate usability early on, gather unbiased feedback from real users and compare how well the conversational flow worked across different perspectives.

Research Questions

  • How easy was it to understand what Cate does?

    87% said easy or very easy

    After the first chat messages, what did you think Cate could help you with?

    73% described it as “a faster way to get a catering offer

  • How easy was it to answer Cate’s questions about your event?

    80% rated the flow easy or very easy

    Did you feel the conversation length was too short, just right, or too long?

    67% said “just right”, 23% said too long

    Were there any questions that felt unnecessary or repetitive?

    18% pointed to duplicate event detail questions

  • Did you trust the offer that Cate generated for you? Why or why not?

    72% said yes

    How clear was the price breakdown (food, drinks, equipment)?

    65% highlighted this as the most valuable part

    Would seeing caterer logos, reviews, or past customer logos increase your trust?

    90% said yes

  • How enjoyable was using Cate compared to filling out a traditional form?

    82% preferred Cate

    What part of the experience did you like most?

    Top answers: “felt personal”, “easier than forms”, “clear pricing.”

    What part would you improve or change?

    Most common: shorten conversation length

  • Would you prefer to use Cate or a traditional form for your next catering booking? Why?

    76% chose Cate

    Would you recommend this tool to a colleague or friend?

    71% said yes

Competitor Insights

As part of our research, we analyzed a Kitchen AI Planner, an AI-powered conversational tool that helps customers design customized kitchen layouts. Similar to Cate, it uses a chat interface to guide users step by step through their needs and preferences.

Insights

Smooth onboarding: the conversational flow felt friendly and easy to start, lowering the barrier for first-time users

Clear visual previews: after the chat, users received instant visual mockups of their kitchen, which made the AI’s suggestions easier to understand and increased engagement

Pain Points

Overloaded chat: too many questions in one go made the process feel long and tiring

Limited flexibility: users couldn’t easily go back and adjust earlier inputs without restarting the conversation

Generic outputs: the AI-generated kitchen offers often felt repetitive and not truly personalized

Missing trust elements: no clear indication of accuracy, expert validation, or reviews left users uncertain about the results

Define

Goal

The AI Cate project aims to automate catering offer creation, reducing manual work for Sales while ensuring trust and accuracy. At the same time, it gives customers a smooth self-service experience that feels modern and reliable.

KPIs

  • Average time-to-offer (request → PDF)

  • Sales hours saved per offer created

  • Customer satisfaction score (CSAT/NPS) on offer process

  • Customer satisfaction with Marketplace

  • Conversion rate: offers sent → deals closed

  • Consistency of tax info and pricing across all offers

Out of scope

  • Full AI-driven price negotiation with customers

  • Deep integration with every third-party catering ERP

  • Multi-language natural language generation for offers

  • Offline/print design customization beyond standardized PDF templates

  • AI chatbots for full customer support

Persona Development

From my research, I noticed patterns among potential users of an AI-powered catering assistant. They liked the idea of a chat-based flow that feels personal and guided but were cautious about long conversations, repetitive questions and whether the AI could generate reliable offers. I also found that corporate clients appreciated trust elements like logos, reviews and clear pricing breakdowns to feel confident.

So I created the persona, Sarah Becker.

User Persona: Sarah Becker

Age: 32

Occupation: HR Manager at a multinational company

Work Style: Hybrid (2–3 days in office)

Location: Frankfurt

Goals:

  • Easily describe event needs without filling out complex forms

  • Get accurate, professional offers instantly

  • Save time coordinating with Sales for standard requests

Frustrations:

  • Lengthy or repetitive chat interactions

  • Generic or unclear offer outputs

  • Lack of transparency on pricing and subsidy handling

Behaviors:

  • Willing to try AI tools if they save time and simplify work

  • Relies on trust elements (logos, USPs, clear PDFs) before committing

  • Often adjusts event details (budget, headcount, extras) last minute

Develop

User Flow

Concepts

Chat interface: core of the project → built a friendly step-by-step flow that collects event details and auto-structures them into clean data (date, city, courses, cuisine, allergies, etc.).

Caterer matching & offers: key innovation → built real-time matching with a ranked shortlist and added AI-generated offer previews, including upsell suggestions like drinks or desserts.

Trust elements: strengthened credibility → added customer logos, NPS score, key platform USPs and the footer to build confidence and reassure users throughout the flow.

Saved chats: improved continuity → added a side drawer with full chat history so customers can revisit past conversations, pick up where they left off and reuse event details without starting from scratch.

Lead capture form: final step of the flow → built to collect the last missing details needed to create a qualified lead and enable the final catering booking.

Beyond Design

Tech & Operations Work

Inventory & data foundation

  • Cleaned and standardized caterer inventory for Berlin & Frankfurt, focusing on Breakfast, Fingerfood, and Business Lunch categories

  • Fixed allergen and dietary tags and replaced poor images with professional photos

  • Created a centralized inventory schema so Cate could reliably generate offers without manual corrections

Tech integrations

  • Synced Cate’s outputs directly into Salesforce, ensuring all structured event data and offers are logged

  • Integrated Stripe checkout inside the chat for a seamless self-service booking flow

  • Built email triggers (confirmation, offer preview, booking summary) connected to the chat events

Automation & intelligence

  • Added matching algorithms to rank caterers not just by availability but also by historical success, customer preferences and dietary fit

  • Built an offer generator template for PDFs with consistent pricing, tax info and branding

  • Configured confidence scores so Sales can quickly see how reliable each AI-generated offer is

Testing & rollout

Piloted first with Berlin & Frankfurt caterers, gradually expanding inventory scope

Ran internal sessions with the Sales team to test accuracy and identify edge cases

Set up feedback loops: Sales and Ops could flag bad outputs, which trained the system further

Designed A/B tests comparing traditional manual offer creation vs. Cate flow to measure time saved and conversion impact

Deliver

What was the result and impact?

For customers: faster, simpler, and more flexible catering planning with less back-and-forth

For Sales: less manual qualification work, more focus on high-value deals

For the business: faster time-to-offer, improved offer quality, and higher closing rates

Cate turns a slow, manual process into an intelligent, automated experience that saves time and increases trust.

View hi-fi prototype