The Data Advantage #10, August 2026

You are here: / / The Data Advantage #10, August 2026
Subscribe to the Data Advantage Book a Demo

Newsletter: Issue #10

FROM DATA TO CONTEXT: THE NEXT ENTERPRISE ADVANTAGE

Data tells you what happened. Context tells you what it means — and what to do next.

Welcome to the Data Advantage: Issue #10,

Enterprises have spent years building the infrastructure to collect more data.

Cloud platforms have made storage almost limitless. SaaS applications have multiplied the number of systems generating information. Analytics platforms have made data increasingly accessible. And AI has given organizations an entirely new way to interrogate it.

Yet access to more data has not necessarily made decision-making easier.

The problem is increasingly one of context.

A number on a dashboard can tell you that abandonment increased. It does not automatically tell you which queues were responsible, whether staffing changed, what happened to demand, whether the same pattern occurred last month, which customer segments were affected or what action should be taken next.

A customer record can tell you who purchased a product. It may not explain their service history, recent interactions, commercial importance or whether another business unit defines that same customer differently.

And an AI model can retrieve thousands of pieces of information without necessarily understanding which information matters.

That distinction is becoming increasingly important.

In April 2026, Gartner described context as emerging critical infrastructure for data and analytics, arguing that semantics and metadata are becoming mission-critical as organizations deploy AI agents. Gartner also reported that organizations with the highest maturity in AI-ready data and analytics capabilities are achieving up to 65% greater business outcomes, including revenue growth and cost optimization.

The next data advantage may therefore be less about how much information an organization possesses and more about whether it can establish the relationships, definitions, history, permissions and business meaning around that information.

Because data provides facts.

Context turns those facts into intelligence.

In this Issue:

  1. AUGUST — BY THE NUMBERS
  2. FROM DATA TO CONTEXT: THE NEXT ENTERPRISE ADVANTAGE
  3. THE AI-READY DATA GAP
  4. WHY AI NEEDS CONTEXT, NOT JUST DATA
  5. WHEN THE SAME METRIC MEANS DIFFERENT THINGS
  6. REAL-TIME DATA IS NOT ENOUGH
  7. GOVERNANCE IS PART OF THE CONTEXT
  8. THE CONTEXT LAYER: CONNECTING DATA, MEANING AND DECISIONS
  9. TECHNICAL DEEP DIVE: BUILDING AN ENTERPRISE CONTEXT ARCHITECTURE
  10. FROM DASHBOARD TO CONVERSATION
  11. THE KEY TAKEAWAY
  12. RESOURCES

TL:DR

1.AUGUST — BY THE NUMBERS

These are just some of the numbers you will find in this months Newsletter

4X

Organizations reporting successful AI initiatives invest up to four times more in data quality, governance, AI-ready people and other foundational areas than organizations experiencing poor AI outcomes

65%

Organizations with the most mature AI-ready data and analytics capabilities are achieving up to 65% greater business outcomes, according to Gartner.

17–23 pts

Accuracy improvement observed in a 2026 benchmark when frontier AI models querying structured enterprise-style data were provided explicit business semantics rather than database schema alone.

80% / 60%

By 2027, Gartner predicts organizations that prioritize semantics in AI-ready data will increase agentic AI accuracy by up to 80% and reduce costs by up to 60%.

75%

More than three-quarters of organizations are prioritizing investment in AI-ready data as data readiness becomes a major obstacle to scaling AI.

2 Aug 2026

The date Article 50 transparency obligations under the EU AI Act began applying.

2. FROM DATA TO CONTEXT: THE NEXT ENTERPRISE ADVANTAGE

Enterprises have spent years building the infrastructure to collect more data, but access to more data hasn’t made decision-making easier. As technology environments multiply, the same business concept ends up with several technically-correct-but-different definitions across systems — a “customer” in CRM isn’t the same as a “customer” in billing.

Experienced analysts used to quietly resolve these inconsistencies; AI removes that buffer and forces the gap into the open.

Gartner now frames context — semantics, metadata, relationships — as emerging critical infrastructure, and organizations with the highest AI-ready data maturity are achieving up to 65% greater business outcomes. The next advantage belongs to organizations that can explain what their data means, not just retrieve it.

Read the Full Article

3.THE AI-READY DATA GAP

AI has made sophisticated analytics look remarkably simple, which is exactly why the data foundation underneath it matters more, not less. A capable model can produce a convincing answer on top of contradictory definitions or missing integrations, which can make underlying problems harder to spot rather than easier.

Gartner reports over 75% of organizations are now prioritizing AI-ready data investment, while earlier research warned that 60% of AI projects unsupported by AI-ready data would be abandoned through 2026.

McKinsey’s June research reinforces the point: as pilots move toward scale, data, not model choice, is emerging as the real constraint. AI readiness, in other words, is enterprise information readiness: accessible, integrated, defined, governed and traceable data, not simply a bigger model.

Read the Full Article

4.WHY AI NEEDS CONTEXT, NOT JUST DATA

Retrieving data and understanding it are different problems, an AI asked why satisfaction declined still has to work out which queues were affected and what actually caused it, none of which lives in the raw numbers.

Database schemas were built for developers, not natural-language interpretation, which is why a 2026 benchmark found that giving AI models explicit business semantics alongside a schema improved accuracy by 17–23 percentage points across three frontier models. This is giving rise to “context engineering” as its own discipline and Gartner now recommends treating it as a core enterprise capability rather than an individual skill, a small number of large enterprises, including Adobe, have already created dedicated context engineering roles.

The section adds a newer lens too: knowing who’s asking.

A compliance officer, a CFO and a frontline manager can ask the identical question and each require a different, correct answer, so context engineering increasingly has to model role and persona, not only data and permissions.

Read the Full Article

5.WHEN THE SAME METRIC MEANS DIFFERENT THINGS

Some of the hardest data problems aren’t caused by bad data, they’re caused by different, equally correct versions of the same metric. Sales’ booked revenue, finance’s recognized revenue and marketing’s contact count can all be accurate and still disagree, because each answers a different underlying question; the trouble starts only when they’re combined without anyone flagging the difference.

Traditional dashboards hid this because analysts predefined the calculations; natural-language AI removes that buffer, so the same undeclared-context problem now happens instantly and at a scale no analyst can catch.

The same metric can also require a different answer depending on persona context, an abandonment rate means something different to a team leader, a workforce planner and an executive, even watching the same underlying number. The fix isn’t one universal definition; it’s making each definition explicit and pairing it with the role and decision behind the question.

Read the Full Article

6.REAL-TIME DATA IS NOT ENOUGH

Real-time data has become synonymous with better decision-making, but speed alone doesn’t create intelligence, a live number like “abandonment: 14.2%” is only useful once it’s connected to forecast, staffing, handle time and whether the same pattern has happened before.

The real objective isn’t real-time data; it’s the right data, the right context, at the right time, some decisions need millisecond precision, others need twelve months of history and good architecture has to know the difference. That means combining live state, historical context, business context and decision context, so an organization can move from monitoring events to actually understanding them.

This is the model ProDataIQ is built around, bringing those layers together inside a single operational data layer, with AskEmite letting a manager ask for that context directly in plain language and get the right-time answer, not just the real-time number.

Read the Full Article

Quick Read

7. GOVERNANCE IS PART OF THE CONTEXT

Context is not only about meaning.

It is also about authority.

Two employees can ask exactly the same question and legitimately receive different answers.

A frontline employee may be entitled to see the performance of their own team. A regional manager may see multiple teams. An executive may see the entire organization. A finance employee may access commercially sensitive information unavailable to an operations user.

The underlying question is identical.

The context surrounding the person asking it is different.

AI changes the governance challenge

Traditional analytics generally controls access at the application, dashboard or dataset level.

AI introduces more dynamic behavior.

An AI system may:

  • interpret a question,
  • determine which information is required,
  • retrieve information from several systems,
  • combine the results,
  • infer relationships,
  • generate an answer,
  • potentially recommend or initiate an action.

Governance therefore needs to travel through the entire process.

Permissions cannot simply determine whether someone can open the AI application.

They need to influence what information can be retrieved, combined, inferred and exposed.

That makes identity, role, purpose, source authority, lineage and policy part of the context of every query.

Governance is moving closer to the decision

This becomes particularly important as AI regulation matures.

On 2 August 2026, transparency obligations under Article 50 of the European Union AI Act began applying. Among other requirements, the rules address disclosure when people interact directly with AI and the identification of certain AI-generated or manipulated content.

The broader direction is clear.

Organizations will increasingly need to understand:

  • what AI accessed
  • why it accessed it
  • which information influenced an output
  • which permissions applied
  • whether information was generated or retrieved
  • who is accountable for the outcome

Governance is therefore moving from being something applied around data to something embedded inside intelligent workflows.

Trust becomes part of the architecture.

Donwload Understanding Data and AI Governance: A Practical Guide to Frameworks, Standards and Expectations

8. THE CONTEXT LAYER: CONNECTING DATA, MEANING AND DECISIONS

For enterprise architects, the emergence of context creates an important design question:

Where should business meaning live?

Traditionally, meaning has been distributed.

Some exists inside source applications. Some lives within ETL transformations. Some is embedded in BI calculations. Some exists in data catalogs. Some is stored in documentation. And a great deal exists only inside the experience of employees.

The result is often a fragmented semantic environment sitting on top of a fragmented technical environment.

A context layer changes the architecture

Conceptually, the modern enterprise data architecture is beginning to look less like:

Sources → Warehouse → Dashboard

and increasingly like:

Sources → Integration → Data Foundation → Context → Intelligence → Decision

The context layer does not necessarily need to be one physical platform.

It represents a set of capabilities that allow enterprise information to be understood consistently.

These may include:

  • Metadata — What information exists, where it originated and how it is structured.
  • Semantic definitions — What business concepts, measures and entities mean.
  • Relationships — How customers, transactions, employees, products, interactions and events connect.
  • Lineage — Where information came from and how it was transformed.
  • Identity and permissions — Who can access what information and under which circumstances.
  • Operational state — What is currently happening.
  • Historical context — What occurred previously and what constitutes normal behavior.
  • Policy — Which rules determine how information may be used.
  • Provenance — Which source or process contributed to a decision or answer.

Once these capabilities are available consistently, they can support multiple consumers: dashboards, analysts, applications, machine-learning models, natural-language interfaces, AI agents.

This is significant because organizations have historically rebuilt substantial portions of this logic for every new analytical experience.

A shared context foundation creates the possibility of defining meaning once and applying it across many forms of intelligence.

9.TECHNICAL DEEP DIVE: BUILDING AN ENTERPRISE CONTEXT ARCHITECTURE

There’s no single technology called an “enterprise context layer”, it’s better understood as a capability spanning seven layers, from source systems and integration through normalization, semantic context, governance, retrieval and intelligence and finally decision and action.

Each layer solves a distinct problem: entity resolution stops the same customer looking like five different people across systems; the semantic layer turns cryptic fields like tbl_int_245.status_cd into business language like “Interaction Outcome”; governance determines what’s appropriate, not just what’s technically possible; and retrieval assembles the minimum sufficient context an AI needs, rather than dumping the entire warehouse into a prompt.

This is exactly the architecture emite, ProDataIQ and AskEmite are built around, emite unifying operational and enterprise data, ProDataIQ providing the contextual intelligence foundation, AskEmite enabling natural-language insight and action on top of it.

The principle carries through every layer: the more autonomy AI is given, the stronger the foundation underneath it needs to be.

Read the Full Article

10. FROM DASHBOARD TO CONVERSATION

Business intelligence has traditionally required people to adapt themselves to the data.

Users learn which dashboard to open. Which filter to select. Which report contains the right measure. Which analyst understands the underlying model.

Natural-language analytics reverses that relationship.

Instead of learning how the data environment is organized, users can increasingly express the business question directly.

  • Which queues had the biggest abandonment problem this month — and what caused it?
  • Where did forecast staffing differ most from actual demand last week?
  • Which customer interactions contributed most to our decline in satisfaction?
  • Which teams are consistently exceeding service-level targets without increasing staffing?

Those are not dashboard questions.

They are business questions.

And answering them requires context.

This is where AskEmite changes the interaction model.

AskEmite allows users to interrogate enterprise information using natural language rather than navigating through predefined reports.

But the important shift is not simply the conversational interface.

The value comes from what sits beneath it.

emite brings information from operational environments together.

ProDataIQ provides the intelligence framework through which information can be structured, connected and interrogated including dashboards configured to specific role-based personas..

AskEmite provides the conversational layer through which business users can ask questions of that information.

The progression is therefore not:

Dashboard → chatbot.

It is:

Data → context → intelligence → conversation → decision.

That distinction matters.

Without the data foundation, conversation produces uncertainty.

Without context, AI must guess.

With both, natural-language analytics can become something much more significant: a new interface between people and enterprise intelligence.

Book a Demo

11.THE KEY TAKEAWAY

For years, competitive advantage came from having more information.

Then it came from analyzing information faster.

The next stage may be about understanding information more completely.

As organizations introduce generative AI, natural-language analytics and increasingly autonomous agents, they are discovering that the quality of an answer depends not simply on the amount of data available.

It depends on whether the system understands:

what the data represents,

how it relates,

which definition applies,

whether it can be trusted,

who is asking,

and what decision is being made.

That is the shift from data to context.

And it changes the question every data leader should be asking.

Not:

“Do we have the data?”

But:

“Do we have enough context to understand what it means?”

Because:

Data tells you what happened. Context tells you what it means — and what to do next.

The closing thought: the question every data leader should be asking isn’t “do we have the data?”

It’s “do we have enough context to understand what it means?”

Subscribe to the Data Advantage Book a Demo

12.SOURCES AND FURTHER READING

*Figures and dates are current as at the time of writing. This issue is not sponsored by, endorsed by, or produced in partnership with any organisation mentioned in this issue. This is not legal advice, organisations should confirm specific obligations with qualified counsel.

TECHNICAL DEEP DIVE: BUILDING AN ENTERPRISE CONTEXT ARCHITECTURE
emite
Powered by  GDPR Cookie Compliance
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.