---
title: Where we use AI in customer analytics and where we don't
description: "A year of building customer analytics with AI agents: the 4 jobs where it reliably works and the 5 where we tell clients to use something else."
---

[![Data Agents](https://data-agents.co.uk/hubfs/raw_assets/public/data-agents-theme/images/data-agents-logo-white.png)](https://data-agents.co.uk/)

[Our approach](https://data-agents.co.uk/blog/where-we-use-ai-in-customer-analytics-and-where-we-dont#pillars) [Our work](https://data-agents.co.uk/blog/where-we-use-ai-in-customer-analytics-and-where-we-dont#cases) [Our thinking](https://data-agents.co.uk/thinking) [DiagnosticAI](https://data-agents.co.uk/diagnosticai) [Contact](https://data-agents.co.uk/blog/where-we-use-ai-in-customer-analytics-and-where-we-dont#contact) 

[Start with DiagnosticAI](https://meetings-eu1.hubspot.com/simon-spyer/discovery-call)

ai · Aug 19, 2026, 1:02:26 PM

# Where we use AI in customer analytics and where we don't

We've spent the past year building customer analytics with AI agents. The same few jobs keep coming up where AI reliably works well. There's also a clear set where we tell clients not to use it. This piece covers both.

A unified customer view and predictive analytics are the two data problems marketers name most often, both cited by 34% of teams in [Supermetrics' 2026 Marketing Data Report](https://supermetrics.com/marketing-data-report-2026). Those are the first two jobs below. Here are the applications that came up most, then the ones where AI doesn't belong.

## Building the single customer view

A lot of analytics problems come back to the same thing: you can't see the customer in one place, transaction history sits in one system, email engagement in another, loyalty and spend somewhere else again. The CRM often recognises someone only at their last transaction so nobody sees a whole lifetime of value.

A single customer view is one record per customer, pulled from every system they touch. Prediction, personalisation and measurement all need it before they can start. You can't work across a journey you can't see in full.

This is the job agents speed up most. Reconciling messy multi-source data used to take a hours of engineering and analyst time: matching identities, standardising fields, resolving the duplicates and the near-misses. An agent does the same reconciliation in hours, with every rule documented and auditable. Speed matters because everything else depends on this.

## Scoring who matters

Once you can see customers whole, the next question is which of them deserve the attention. Two techniques do most of the work.

Segmentation groups customers by how they behave. Clustering lets those groups form from the data itself rather than a rule written in advance so you find patterns you didn't know to look for. Value and lead scoring rank individuals by how likely they are to be worth something: to convert, to grow, to lapse.

An agent builds, tests and re-cuts a scoring model inside seamlessly. The harder work is defining what "valuable" actually means in your business, with your customer base, at your price point. That definition decides more than the choice of model. Get it wrong and you have an accurate model aimed at the wrong thing. Most teams decide the model before they've done that thinking.

## Measurement and effectiveness

The third job is telling what actually worked. Marketing teams have plenty of numbers and still can't say what worked or what it was worth. Different systems report different figures for the same thing. A campaign looks successful until you find the numbers behind it don't reconcile.

AI helps in two ways. It reconciles conflicting figures to one base everyone can trust. Then it sizes each point of loss in pounds so you can rank the fixes by what they're worth. The discipline that matters is separating lift from noise: the change that came from the activity, measured against what would have happened anyway.

## Making the data answerable

The fourth job is access. A model or a dashboard only helps if the people with the questions can reach it without waiting. Half of marketers say they wait 1-3 days just for data team support, which slows every decision that depends.

We've built natural-language query tools and live dashboards that sit in the client's own environment so a marketer can ask a question in plain English and get an answer there and then. The capability stays with the team. They read it, run it and extend it after we've gone.

This changes how marketing teams works day to day. A build that arrives in days means the first answer comes back while the decision is still open. The team has time to ask the next thing it raises so the model becomes something people use while they can still affect a decision and optimise performance.

## Where AI isn't the answer

Those are the four jobs where AI pays. There are also five where we tell clients not to use it.

- **The question isn't defined.** AI makes the answer faster. It can't choose which question is worth asking. That decision is yours, and it matters more than the speed.
- **The signal was never collected.** A model only works with what's in the data. If the behaviour you care about was never captured, no model will find it. That's a tracking job first.
- **You need proof of cause.** Reading history shows you what correlates. Proving what caused a result needs a designed test with a holdout, measured forward in time. A model run over the past can't do that.
- **The job is small or one-off.** An agent pays back when the work repeats or the data is large. For a one-time question over a small table, a person with a spreadsheet is quicker and cheaper.
- **The real work is judgment.** Positioning, creative and the trust of the room aren't analytics problems. AI can support those decisions. People make them.

## What a year of this taught us

Three things decided the outcome every time.

First, the question. AI answers it faster than ever. Choosing which question to ask is still your job.

Second, the data you own. Agents are the delivery mechanism. Data is the moat.

Third, speed. A fast build lets you keep asking while the decision is still live.

Only 41% of marketers can demonstrate a return on their AI investment in 2026, down from 49% a year earlier, according to Benchmarkit's State of AI in Marketing 2026. The teams who show a return use AI on the right jobs and their own data. They leave it alone for the rest.

Want to see how Data Agents can help you do more with less? [Book a quick strategy call](https://meetings-eu1.hubspot.com/simon-spyer/30-minutes).

![Simon Spyer, Founder of Data Agents](https://data-agents.co.uk/hubfs/raw_assets/public/data-agents-theme/images/simon-spyer.jpg)

Author

## Simon Spyer

Founder, Data Agents

Simon has 20 years of CRM, performance, and data-led marketing behind him — working with brands including Sainsbury's, IKEA, Barclays, Samsung, KFC, and News UK. He founded Data Agents to solve the problem he watched repeat itself across every organisation he worked with: great data, no time to use it.

[Start with DiagnosticAI](https://meetings-eu1.hubspot.com/simon-spyer/discovery-call) [Connect on LinkedIn →](https://www.linkedin.com/in/simonspyer/)

More from Thinking

ai

### [What 12 months of AI-assisted analytics taught us about customer data](https://data-agents.co.uk/blog/what-6-months-of-ai-assisted-analytics-taught-us-about-customer-data)

Oct 9, 2026, 10:31:00 AM

In the last year we've run AI-assisted analytics on customer data for businesses in leisure, restaurants, insurance, beauty, streaming and direct-to-consumer ecommerce. The ...

[Read →](https://data-agents.co.uk/blog/what-6-months-of-ai-assisted-analytics-taught-us-about-customer-data)

ai

### [AI agents in customer analytics: what a year of builds taught us](https://data-agents.co.uk/blog/ai-agents-in-customer-analytics-what-a-year-of-builds-taught-us)

Aug 13, 2026, 11:55:36 AM

We've spent the past year building customer analytics with AI agents. The work covered segmentation, single customer views, value prediction and funnel diagnostics across several ...

[Read →](https://data-agents.co.uk/blog/ai-agents-in-customer-analytics-what-a-year-of-builds-taught-us)

ai

### [Keyword research is dead. Prompt research is what matters now.](https://data-agents.co.uk/blog/prompt-research-beats-keyword-research)

Jun 15, 2026, 9:45:57 AM

Your keyword research is optimised for a behaviour that no longer exists. The tools are precise, the methodology is rigorous but the outputs are completely disconnected from how ...

[Read →](https://data-agents.co.uk/blog/prompt-research-beats-keyword-research)

## For marketers who are done waiting for answers.

[Let's talk](https://meetings-eu1.hubspot.com/simon-spyer/discovery-call)

![Data Agents](https://data-agents.co.uk/hubfs/raw_assets/public/data-agents-theme/images/data-agents-logo-white.png)

From Data to Decisions. In Weeks.

Contact

- [info@data-agents.co.uk](mailto:info@data-agents.co.uk)
- [Schedule a call](https://meetings-eu1.hubspot.com/simon-spyer/discovery-call)
- [LinkedIn](https://www.linkedin.com/company/data-agents)
- [YouTube](https://www.youtube.com/channel/UCptDGE-mPOVC5ZxdMzjzmfQ)

Sitemap

- [Home](https://data-agents.co.uk/)
- [DiagnosticAI](https://data-agents.co.uk/diagnosticai)
- [Thinking](https://data-agents.co.uk/thinking)
- [About](https://data-agents.co.uk/about)
- [Privacy Policy](https://data-agents.co.uk/privacy-policy-data-agents)

©2026 Foundry Data Ltd t/a Data Agents. Registered in England and Wales, company number 16257653. Registered office: 162-164 High Street, Rayleigh, Essex, SS6 7BS.

```json
{
  "@context" : "https://schema.org",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "Simon Spyer",
    "url" : "https://data-agents.co.uk/blog/author/simon-spyer"
  },
  "dateModified" : "2026-08-19T19:08:56.287Z",
  "datePublished" : "2026-08-19T17:02:26.000Z",
  "headline" : "Where we use AI in customer analytics and where we don't",
  "mainEntityOfPage" : {
    "@id" : "https://data-agents.co.uk/blog/where-we-use-ai-in-customer-analytics-and-where-we-dont",
    "@type" : "WebPage"
  },
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject"
    },
    "name" : "Data Agents"
  }
}
```

```json
{
  "@context" : "https://schema.org",
  "@id" : "https://data-agents.co.uk/#organization",
  "@type" : [ "Organization", "ProfessionalService" ],
  "address" : {
    "@type" : "PostalAddress",
    "addressCountry" : "GB",
    "addressLocality" : "Rayleigh",
    "addressRegion" : "Essex",
    "postalCode" : "SS6 7BS",
    "streetAddress" : "162-164 High Street"
  },
  "alternateName" : "Foundry Data Ltd t/a Data Agents",
  "areaServed" : "GB",
  "contactPoint" : {
    "@type" : "ContactPoint",
    "areaServed" : "GB",
    "availableLanguage" : "en-GB",
    "contactType" : "sales",
    "email" : "info@data-agents.co.uk",
    "url" : "https://meetings-eu1.hubspot.com/simon-spyer/discovery-call"
  },
  "currenciesAccepted" : "GBP",
  "description" : "Data Agents is a UK AI strategy and execution consultancy that builds AI agents for marketing teams. It turns marketing and customer data into decisions in weeks under fixed-time, fixed-price terms.",
  "email" : "info@data-agents.co.uk",
  "founder" : {
    "@id" : "https://data-agents.co.uk/about#simon-spyer"
  },
  "foundingDate" : "2025-02-17",
  "hasOfferCatalog" : {
    "@type" : "OfferCatalog",
    "itemListElement" : [ {
      "@type" : "Offer",
      "description" : "DiagnosticAI is a one-week, fixed-fee audit of your customer data, tech stack and channel performance. AI agents run it on your live data and it ends in a clear action roadmap.",
      "name" : "DiagnosticAI",
      "price" : "19495",
      "priceCurrency" : "GBP",
      "url" : "https://data-agents.co.uk/diagnosticai"
    }, {
      "@type" : "Offer",
      "description" : "A custom AI agent built around one data or marketing challenge. Deployed into the existing stack in six weeks.",
      "name" : "AgentLab",
      "priceSpecification" : {
        "@type" : "PriceSpecification",
        "minPrice" : "19495",
        "priceCurrency" : "GBP"
      }
    }, {
      "@type" : "Offer",
      "description" : "A full AI marketing programme with multiple agents, delivered in 12 weeks. Fixed fee, scoped before work starts.",
      "name" : "Accelerator"
    } ],
    "name" : "Data Agents services"
  },
  "identifier" : {
    "@type" : "PropertyValue",
    "propertyID" : "Companies House company number",
    "value" : "16257653"
  },
  "knowsAbout" : [ "AI marketing agents", "Single customer view", "Customer segmentation", "VIP and customer value scoring", "Loyalty and CRM analytics", "Generative engine optimisation (GEO)" ],
  "legalName" : "Foundry Data Ltd",
  "logo" : {
    "@type" : "ImageObject",
    "url" : "https://data-agents.co.uk/hubfs/brand/data-agents-logo-blue.png"
  },
  "name" : "Data Agents",
  "priceRange" : "££",
  "sameAs" : [ "https://www.linkedin.com/company/data-agents", "https://www.youtube.com/channel/UCptDGE-mPOVC5ZxdMzjzmfQ", "https://find-and-update.company-information.service.gov.uk/company/16257653" ],
  "url" : "https://data-agents.co.uk/"
}
```

```json
{
  "@context" : "https://schema.org",
  "@id" : "https://data-agents.co.uk/#website",
  "@type" : "WebSite",
  "inLanguage" : "en-GB",
  "name" : "Data Agents",
  "publisher" : {
    "@id" : "https://data-agents.co.uk/#organization"
  },
  "url" : "https://data-agents.co.uk/"
}
```

```json
{
  "@context" : "https://schema.org",
  "@id" : "https://data-agents.co.uk/about#simon-spyer",
  "@type" : "Person",
  "description" : "Simon Spyer founded Data Agents in 2025. He was previously CEO of Iris's global data-driven and performance marketing business and has 20 years of CRM, performance and data-led marketing experience on the brand and agency side.",
  "image" : "https://data-agents.co.uk/hubfs/raw_assets/public/data-agents-theme/images/simon-spyer.jpg",
  "jobTitle" : "Founder",
  "knowsAbout" : [ "CRM", "Performance marketing", "Customer data", "AI marketing agents" ],
  "name" : "Simon Spyer",
  "sameAs" : [ "https://www.linkedin.com/in/simonspyer/" ],
  "url" : "https://data-agents.co.uk/about",
  "worksFor" : {
    "@id" : "https://data-agents.co.uk/#organization"
  }
}
```