Back to insights

You’ve got the data, but you don’t have the answers

  • Artificial Intelligence
  • Strategy

Dan Llewellyn

 

So, you’ve just purchased a top-of-the-range data hoover.  

It promises to suck up almost everything from your technology estate, from supplier contracts to security documents, to wiki pages.

The problem is there’s a big gap between having a full bag of data and having the connections and context required to make the right decisions about modernising your tech estate.

“Enterprise context” and “context spine” have become industry buzzwords, and many platforms promise to get you 80% of the way there with a single click. However, the devil is in that missing 20% where the data is prioritised, connections are made and the answers become clear.

This is the problem we’re solving with Compass.

Real life example

Say you’ve been given next year’s budget to modernise part of your application architecture. You need to understand the costs and vulnerabilities in your existing system, where scaling is difficult and which dependencies could cause problems.

You also have supplier contracts, previous incidents and years of analytics to consider. Somewhere in all that information are the answers you need. But weighing up the trade-offs between cost, speed, simplicity and risk can take weeks or months.

The information exists. The challenge is bringing it together when a decision needs to be made.

Some organisations have a configuration management database (CMDB) to connect this information. Many don’t, or have one nobody trusts.

A better way to connect

Context matters. An invoice in a finance system needs a supplier’s contact details. An invoice in a technology estate graph needs to show the cost of each service.

Getting those choices right requires people who understand the business. AI can help connect the information, but people need to decide what matters.

Effective AI is very much a human issue and that’s why we’ve developed an accelerator called Compass that builds an enterprise knowledge graph. This is a connected view of how things relate across an organisation.

An application, for example, can link to its codebases, the infrastructure it runs on, the databases it depends on, the team that owns it and previous incidents it has caused.

With those relationships in place, AI can trace dependencies and answer questions that would otherwise take weeks of investigation.

Building and maintaining a graph like this by hand is slow, and it’s easy to miss things. Compass uses AI to propose how new information should connect to what’s already there.

People who understand the organisation review those proposals and shape the graph. Once a data source is understood, new information in the same format can flow in automatically.

How we built it

Compass began as a way to help customers make better decisions about complex modernisation challenges, from building AI-first platforms to scaling applications or consolidating systems after an acquisition.

We started with five to ten data sources. For each one, AI proposed what to add to the graph and how it should connect to existing information. Someone with knowledge of the organisation reviewed every proposal, often suggesting additional information worth capturing.

That early human input proved essential. It helped establish what the graph needed to contain and how the information should relate.

After working through the first few sources, we let AI handle most of the ingestion. Within a couple of days, the graph held the bulk of the data.

Our view is that a focused graph beats a bigger one. Start with a clear purpose, such as managing infrastructure risk or controlling software spend, and deliberately decide what information belongs in it. The result is a system that can give you answers you can stand behind.

When knowledge finds another job

Much of an organisation’s knowledge still lives in people’s heads. When someone moves off a system, that knowledge can go with them.

As AI takes on more coding work, this becomes a bigger risk. Organisations need a way to retain the knowledge that keeps their systems running.

Compass can uncover risks that are difficult to spot in individual systems, such as:

  • A critical platform with a sole code owner approaching retirement.
  • Infrastructure that relies on skills no longer readily available in the recruitment market.
  • Hidden dependencies between schema changes across repositories that appear unrelated.

These are the kinds of issues that can turn a manageable migration into a crisis.

Built to be used

Compass isn’t just a graph to look at. Teams can query it directly, connect their own AI coding tools or build agents that use its information.

It can also work within existing processes, flagging risks based on other teams’ backlogs or estimating the cost of a proposed change in the context of each organisation.

Most teams struggle to put a reliable business value on individual pieces of work. They often have to guess.

By connecting cost, risk, ownership and existing plans, Compass gives teams a stronger basis for deciding what to do next.

For organisations facing migration, modernisation or post-acquisition rationalisation, that means less rework, fewer blind spots and plans built on a clearer understanding of the estate.

Keep the hoover if you like. Just make sure something’s sorting through the bag.

Want to see what Compass could uncover in your own estate? We’d be happy to show you. Find out more about how we can connect up your decision-making here.

You might also be interested in Atlas, another accelerator that reads your estate to reveal risks and opportunities. Read more about that here.

 

FAQs

What is Compass?

Compass is CreateFuture’s enterprise knowledge graph. It connects information that usually sits in separate places, such as applications, infrastructure, bills, supplier contracts, migration plans and ownership, so you can see how it all relates and make better decisions.

AI proposes how new data should link to what’s already in the graph. Early on, people who understand the business review those proposals and shape what belongs in it. Once a data source is understood, new information in the same format flows in automatically.

Technology leaders facing decisions that depend on many sources at once, like where to spend a security budget, which systems to retire or how to approach a supplier renegotiation. It’s built for complex organisations planning a migration or modernisation, or running development teams too large for anyone to see the whole picture.

A platform that stays with you. Your teams can query it directly, point their own AI tools at it or build agents on top of it. It can also sit inside existing workflows, for example flagging which security tickets genuinely need fixing first.

Meet the author

Daniel Llewellyn is Director of Technology at CreateFuture. He specialises in cloud architecture, AI-led delivery and engineering enablement, with a background spanning solutions architecture, security and software engineering across financial services and technology.

 

Dan L author image

 

Industry Insights

Explore the latest thinking from our industry and tech experts.

What is AI native?
Digital Transformation

What is AI native?

by CreateFuture
Your discovery phase is already out of date
Artificial Intelligence

Your discovery phase is already out of date

by Chris Daly
CreateFuture expands its Anthropic AI capabilities as Version 1 Group achieves Preferred Services Partner status
Artificial Intelligence

CreateFuture expands its Anthropic AI capabilities as Version 1 Group achieves Preferred Services Partner status

by CreateFuture