DE EN
Intro call · 30 min

What you know about your customers is already there. It's just in too many places.

I bring together what sales, service, your digital channels, and your studies know about customers, and make it usable for your decisions.

Intro call · 30 min Free, no obligation View example
Your data sources are brought together along your customers' journey in a curated data foundation. This produces AI-assisted interactions and five applications for your team. Your data sources Your data sources Your customers'journey Curated datafoundation AI-assistedinteraction What your teamdoes with it Need Selection Purchase Commissioning Service Your customer knowledge Query data innatural language Simulateconversations withtypical customers Test offers Package, price, promisebefore going to market Prepare conversations Know the objections beforethe meeting Test digital touchpoints Landing page,configurator, newsletter Understand leads which signals point topurchase intent Set priorities which break point to fixfirst Service tickets andcomplaints Ticketing system Surveys Voice of Customer Sales and dealers CRM Website and leads Web Analytics Email and webinars Marketing automation Studies Market research
  1. Your data sources
    • Service tickets and complaintsTicketing system
    • SurveysVoice of Customer
    • Sales and dealersCRM
    • Website and leadsWeb Analytics
    • Email and webinarsMarketing automation
    • StudiesMarket research
  2. Your customers' journey
    1. Need
    2. Selection
    3. Purchase
    4. Commissioning
    5. Service
  3. Curated data foundation
    Your customer knowledge
  4. AI-assisted interaction
    • Query data in natural language
    • Simulate conversations with typical customers
  5. What your team does with it
    • Test offersPackage, price, promise before going to market
    • Prepare conversationsKnow the objections before the meeting
    • Test digital touchpointsLanding page, configurator, newsletter
    • Understand leadswhich signals point to purchase intent
    • Set prioritieswhich break point to fix first

Where the ROI is

  • Time-to-Insight. Answer customer questions in minutes, without rounds of alignment between departments and without an ad hoc study.

  • Cost per Insight. Use existing data first, such as tickets, CRM notes, and dealer feedback. New research only for documented gaps.

  • Pre-testing. Test value proposition, price, offer, and campaign with synthetic customers before budget goes into launch and media.

  • Insight Adoption. Customer knowledge in the daily work of sales, service, marketing, and product management, tailored to each role, for example to prepare for objections in a meeting or for website and campaign content.

How I work

  1. 01

    Data screening

    Review existing sources, including digital signals such as Web Analytics, portal and campaign data. Check quality and coverage, and fill gaps where it matters.

    Result

    Source map with gaps for each phase of the customer journey

  2. 02

    Curated data foundation

    Reviewed, organized along the customer journey, every statement marked with its source or flagged as an assumption.

    Result

    A knowledge base that your team can query in their own words

  3. 03

    Use it day to day

    Ask the data questions or talk to customer types, in your own environment.

    Result

    Actionable results for sales, service, and marketing

This doesn't necessarily require an IT project. We'll work out the exact requirements and implementation together, based on your existing tech stack.

Two ways to use the data foundation

Most companies have plenty of customer knowledge: studies, tests, service tickets. I make it available in a deliberately curated data foundation that your team uses in two ways.

Ask your data

You ask a question in natural language, starting from your business problem. The answer shows what it is based on and where knowledge is missing.

Example

What content should visitors see on our website in each phase?

  • First-time visitors from search read application examples and rarely open datasheets.

    Web Analytics

  • People who use the configurator multiple times are significantly more likely to request a quote.

    Web Analytics and CRM

  • Existing customers with an open service ticket show little response to upgrade offers.

    Ticketing system and email campaigns
Result

From these signals, I derive rules for your Personalization Engine: which signal triggers which content.

Gap

Customers who buy through dealers leave few digital traces. Their visits cannot be attributed.

Conversations with customer types

Together, we design typical customers as conversation partners, known as synthetic personas. You decide which focus topics to include.

Example

The head of purchasing · built from purchasing interviews, visit reports, quote histories, and Web Analytics

  1. Team

    Where do you look for information before calling the dealer?

  2. Customer

    On your website and on comparison sites. If I find neither a price range nor a reference from my industry there, I don't even ask for a quote.

Personas based on two-hour interviews with 1,052 people answered General Social Survey questions 85 percent as accurately as the participants replicated their own answers two weeks later. Park et al., 2024, arxiv.org/abs/2411.10109

On the limits of chatbots and synthetic personas

You may know this from your own experience: upload a market research report to Copilot and have it answer like a specific customer from the report. Systems without conditioning respond more politely and agreeably than real customers¹, and they lose context in longer dialogues². That's why I developed a framework for targeted conditioning of the models: every answer is grounded in your data, and statements without a source are marked as assumptions.

  1. ¹ Sharma et al. (2024), Towards Understanding Sycophancy in Language Models, ICLR, arxiv.org/abs/2310.13548; Bisbee et al. (2024), Synthetic Replacements for Human Survey Data? The Perils of Large Language Models, Political Analysis.
  2. ² Laban et al. (2025), LLMs Get Lost In Multi-Turn Conversation, arxiv.org/abs/2505.06120

Let's talk about where you stand.

Intro call · 30 min Free, no obligation

Prefer email info@streetsmartcx.com