The last blog in my ‘data decoded’ series looks at the common issues we see when data platforms are built and automatic ROI is expected.

Familiar Data Platform Challenges & Downfalls

• Data Silos

Department A builds a platform, Department B builds their own. Two platforms that don’t talk = fragmented view, no enterprise coherence.

Explore this further in: Should You Build or Buy Your Data Platform?

• Technology First Mindset

“Let’s build a lakehouse then figure out what we’ll do with it.” Wrong order.

• Lack of Business Alignment

If business users aren’t engaged, platforms sit idle.

• Ignoring Usage / Adoption

You may have dashboards, but no one uses them.

• No Clear ROI Tracking

If you cannot measure value, you cannot grow investment.

• Over-Engineering Early

Streaming, ML, fancy stuff before you’ve nailed decision-making and adoption.

Avoid these data challenges by starting small, focusing on decision-making, involving business users, measuring value, and building iteratively.

The Role of the Data Platform Consultant

As a data platform consultant, you’re not just building technology – you’re helping your client (or your business) become decision-driven. Your role is to:

• Ask ‘why?’ early and often: Why are we building this? What decision will it support?
• Translate business outcomes into data product features: What dataset, what transform, what serve layer?
• Build lean-first: Minimum viable platform that supports decisions, then scale.
• Embed metrics & governance: Usage analytics, data quality, cost control.
• Communicate value: Demonstrate ROI, show wins, secure funding for next phases.

In short, you’re a bridge between the business (who must decide) and the engineers (who build) – the ultimate solver of data challenges.

A Realistic Journey

Reporting → Streaming → ML

Let me walk you through a journey:

Phase 1: Reporting Store

Build a data warehouse or lakehouse. Ingest key business systems (CRM, ERP, contact centre data). Transform and serve basic dashboards (how many products, how much spend, where are we today).

Value: decision boundary-based actions; time saved.

Phase 2: Domain Expansion & Streaming

Add live data (fleet sensors, IoT, user-behaviour logs). Introduce real-time alerts. Business can act quicker (if latency > threshold, then reroute).

Value: faster reaction time, cost avoided.

Phase 3: Machine Learning / Advanced Analytics

Build models: churn prediction, pricing optimisation, recommendation engines. Embed decision-making logic into the platform (if predicted churn > X, trigger incentive).

Value: new opportunity, revenue gained, risk reduced.

Few companies make it to the full “data enterprise platform” stage. If you think you have, ask yourself: Are we still measuring decision outcomes? Are people using it daily? Are we still leaning on dashboards or spreadsheets?

How to Talk About Data Engineering ROI and Value to Executives

When you’re presenting to the board or senior leadership about the data programme, use their language: time, money, opportunity. Avoid tech-speak. Frame it in business outcomes.

Example one: “By centralising data into a single platform, we’re reducing reporting cycle time by 3 days, which frees 120 person-hours per quarter, equivalent to £X in cost savings.”

Example two: “By implementing real-time pricing analytics, we expect an uplift of Y% on margin, which translates to £Z additional annual revenue.”

Use frameworks like the ones from industry (ROI framework, ROI pyramid) to back-up your case. Learn more by reading: How to Deliver a Successful Data Strategy Presentation to the Board.

Three Big Takeaways from Data Decoded

So, wrapping up the hattrick of my data decoded articles and you should now understand that:

1. Decisions are everything. The only reason you need data is to make better decisions. Without decision-boundaries, you cannot declare you are data-driven.
2. Platform isn’t value. Building data infrastructure is necessary, but on its own it doesn’t yield ROI. You need business usage, adoption, measurable impact.
3. Value needs tracking relentlessly. Always link your data work to time saved, money made, or opportunity unlocked. Use simple frameworks, speak the business language, and secure buy-in.

Watch me talk about how to determine the value in a data platform here: Value of Data Platforms. And if you missed the first two articles in this Data Decoded series, catch up below.

Oakland’s Advice for Starting A New Data Programme

• Hold a workshop to list key decisions your organisation makes every day. For each decision, ask: what data, what threshold, what action?
• Build your platform in phases: ruling out the “acquire everything” approach.
• Engage business users early. Set metrics for adoption, usage, decision-impact.
• Celebrate wins, and communicate them. E.g. dashboards that work, models that deliver, cost savings, revenue gains.

And if you’re consulting or leading this work:

• Ensure every build has a why.
• Be the voice of value, not just tech.
• Don’t fall into the trap of “we built it, they’ll come”.
• Measure, iterate, expand.

The Final Word

In the world of data, everyone wants to talk about “big data platforms”, “machine learning”, “AI”, “data lakes”, “lakehouses”, “real-time streaming”. But none of that matters if what you build doesn’t change decisions or help you overcome data challenges. Because at the end of the day, you make decisions. Your organisation makes decisions. Your business lives or dies by them. Being data-driven means replacing guesswork with evidence, replacing intuition with insight and decision boundaries, and doing so consistently.

So when you hear the phrase “Data Decoded”, think less about the dazzling technology and more about the decoded decision. Think: what decision are you enabling today, with the data platform you have or will build? How will you know you’ve improved that decision? What value will result – in time saved, money made or opportunities unlocked? And when you bring in consultants or build your data team, make sure they understand: the platform is the enabler, the business decision is the destination.

That’s how you:

• Turn “we have a data platform” into “we are a data-driven business”
• See ROI in data programmes
• Get value from a data platform

And that’s how data consulting really pays off.

To speak to me further about any of the challenges we’ve touched on here – or for anything else to do with data platforms – please get in touch.

The post Common Data Challenges & How to Avoid Them appeared first on Oakland.

We hate to break it to you, but there isn’t any natural ROI in building a data platform. The ROI happens when decisions get made from that platform. As a data engineer or data platform consultant, you build around four key functions: Ingesting, storing, transforming, and serving the data. If your platform does these four things, you’re in business. But just because you can tick all four pillars of a data platform doesn’t mean you can tick one that says it provides ROI, too.

Here’s why.

The Four Pillars of a Data Platform Build

 We’ll start with the four functions in more detail:

1. Ingest – bringing data in from source systems (ERP, CRM, cloud apps, IoT, etc).
2. Store – persisting the data in a suitable repository (data lake, warehouse, lakehouse).
3. Transform – modeling, cleansing, structuring, and preparing the data for use.
4. Serve – providing access: dashboards, BI tools, ML models, APIs.

Many data engineering projects start with the diagram ingest → store → transform → serve. And that’s fine. But companies fall into the trap of believing: “we have dashboards, we have a lakehouse, we have streaming ingestion”, therefore “we’re data-driven.” But that’s wrong. If you haven’t tied it to decision boundaries and measurable outcomes, you are spinning wheels.

A Platform Doesn’t Equal Value

Data platforms have cost: Infrastructure, licences, engineers, maintenance. Without real usage and decision-making, you could have a negative ROI. For example, a data platform sitting idle still costs you £5-10k/month in cloud costs.

Jump into this further by reading: How to Manage Spiralling Cloud Costs.

What Does A Data Consulting Company Do?

A good data consultant doesn’t just build a stack. They ask: 

• What decisions will this support? 
• What value will it deliver? 
• How will we know? 

A data consultant helps you avoid the “acquisitive” trap and guide you to become value-driven.

Here’s what our data consultants often advise:

Define business outcomes first

Understand what business decisions you want from the data (pricing decisions, logistics optimisation, customer segmentation, risk reduction).

Set decision boundaries

For each key metric, define the threshold to drive action.

Align data product scope to value

Don’t build everything. Prioritise data products (dashboards, ML models, APIs) that map to time/money/opportunity.

Ensure adoption & governance

A platform is only useful if business users engage with it.

Track ROI continuously

Measure before-and-after, quantify value, build case studies. For example, when consultants deliver dashboards & platforms they embed tracking: “We reduced reporting cycle time by X hours, we reduced cost of chasing data by Y, we improved conversion by Z%.” 

Good consulting keeps the focus on action, not just architecture. Learn more: What is Data Platform Architecture?

How Do You Get Value from a Data Platform?

Let’s break down meeting the four pillars of a data platform into actionable steps:

Step 1: Start with Decisions

You must map which decisions your organisation needs to make routinely. For example:

In utilities: “If usage anomaly > 5%, then investigate the customer meter.”

In transportation: “If fleet-cost per km > X, then re-route fleet or renegotiate contract.”

In tech: “If process latency > X, trigger optimisation run.”

From there, you define the data that will inform that decision.

Step 2: Build Decision Boundaries

As noted earlier, define thresholds in advance. Know what numbers trigger what actions. This ensures your dashboards don’t just show “what happened”, they support “what we will do”.

Step 3: Build the Platform Lean

Instead of “ingest everything”, you prioritise the data sources that serve those decisions. Map ingestion, storage, transform, and serve accordingly. Then build the initial version – it may just be a reporting store – then grow from there: more domains, real-time streaming, ML models.

Step 4: Operationalise the Platform

You need adoption. Business users must use the dashboards, models must deliver insights, the thresholds must be monitored and acted upon. If your data platform is just “nice to have”, it won’t deliver ROI. Remember: value comes when decisions change behaviour. 

Step 5: Measure Value

You must link back to time, money or opportunity. Use frameworks to estimate value – for example:

Data ROI = (value from data initiative – cost of initiative) / cost of initiative

Track metrics such as: time saved, cost avoided, revenue gained, risk reduced. Get evidence around before / after.

Step 6: Iterate and Expand

Your first data product may be reporting. You then add real-time streaming, machine learning, and automated decisioning. The “lighthouse model” (start small, show value, expand) works better than big-bang POCs that die.

Two down, one to go! For more insight like this, please read the third and final blog in this data decoded series: Common Data Pitfalls & How to Avoid Them. You can also send your questions to me or speak with the rest of the Oakland consulting team by getting in touch – we’d love to hear from you.

The post The Four Pillars of a Data Platform appeared first on Oakland.

In today’s world, you don’t run a business unless you make decisions. It doesn’t matter what industry you’re in – utilities, transportation, computing, retail – you are in the business of making decisions. 

Whether you’re setting unit-rates of electricity in utilities, organising a fleet in transportation, or optimising a process in the tech/computer industry, the truth is that you are making decisions. And to compete and stay relevant, you try to be data-driven. Because philosophically that’s the only way humans really know how to operate, right?

The Purpose of Being Data-Driven

When I say “data-driven as a business”, I mean this: for many of the decisions we make as a company, as an employer, as a person, we stop trusting “vibes” or our “spidey sense” about a customer’s behaviour, or whether some process will succeed or fail. Instead, we lean on evidence. We use the data. We say, “This person matched profile A rather than B. Therefore, we think this person will buy A rather than B because we have evidence and data to support that.”

You could spend a lifetime writing books on the subtlety of evidence-driven vs bias-driven decisions. But we don’t have the luxury of endless pages. So, let’s unpack what it means to truly become a data-driven organisation.

Clear Decision Boundaries

First, if you’re going to drive value from data, you need to adopt some ground-rules. In particular, you need clear decision boundaries. 

I want you to reflect on the worst dashboards you’ve seen: the chart-junk, the yellow warning‐lights, the numbers splattered without interpretation. Those dashboards are not enabling decision-making. They’re confusing. 

If you’re going to be evidence-driven, you must know in advance:

• If this number > X, we’ll do Action A.
• If this number < X, we’ll do Action B.

At platform, engineering and analytics levels, this means knowing your decisions before you build the data asset. Because when you build without knowing the outcome, you’ll create a data platform, you’ll move numbers, you’ll host dashboards…

…but you won’t get value.

Data Decoded: Stop Acquisitive, Start Inquisitive

One of the biggest pitfalls in data programmes is what I call an ‘acquisitive data strategy’. That is: “Let’s take every dataset, grab every number, throw it into a warehouse, and then maybe something good will happen.” 

No. 

You need an inquisitive strategy. Ask yourself:

• What decisions will you support? 
• What thresholds matter? 
• What outcomes will change? 

Then drive your data platform accordingly.

When businesses implement a data platform, they often think: “Here’s our cloud budget, build everything, feed everything into the data warehouse, hook up dashboards, we’ll iterate.” But a platform alone doesn’t give ROI. Tools alone don’t deliver value. Without decision-boundaries, without usage, without an evidence base, you’re just moving data around from one expensive database to another expensive database.

What Does “Value” Really Mean?

So, let’s get into value. You may have heard the initialism “ROI” (return on investment). In the data world, ROI is sometimes misused or misunderstood. Value, when you’re building data platforms or data products, comes in (at least) three flavours: time, money and opportunity. Yes, they’re all about money if you dig into them – but they come at you in different ways.

Time saved 

If your data platform or analytics process saves hours, days, weeks of manual reporting or intervention, that frees people to do higher-value work.

Money made

If you can use your data to optimise pricing, reduce cost of goods sold, improve marketing performance, increase conversions, then you’re generating revenue (or preventing loss).

Opportunity unlocked

Sometimes you build a platform that enables new products, new customer segments, new business models – and all of these are valuable too, even if it’s future-facing.

If you’re working as a data engineer, a data platform consultant, or leading a data programme, you must ask: Which one of these am I addressing when building this feature, this platform, this dashboard? If you cannot answer that, you should reconsider your life choices…

…(well, maybe!).

For more insight on the value of data, please read: Why is Data Important for Business?

What We Do Day-to-Day

At the end of the day, what you and I do is we make decisions. Whether you’re in utilities, transportation, tech, retail, you decide. And your ambition is to decide based on evidence, not guesswork. When you embrace data, you move away from trusting your “gut” to saying: “Here’s the evidence, let’s act.”

To become data-driven means that for many of the decisions now, as a company, we’re going to stop using vibes, stop relying on “this customer might buy” because our spidey sense tells us so. Instead we say: “Look at the data. Based on this profile, we believe they will buy A, not B.” Because we have the evidence.

That’s the first of my Data Decoded series finished, but there’s plenty more like this in the second blog: The Four Pillars of a Data Platform. You can also send your questions to me or speak with the rest of the Oakland consulting team by getting in touch – we’d love to hear from you.

The post Data Decoded 2025: How to Get the Most from Data Consulting appeared first on Oakland.

But what is agentic AI? In its simplest form, the term describes semi-autonomous machine “agents” that can go beyond the chatbot and perform enterprise-related tasks without human guidance. 

These AI agents can set goals, plan, and adapt their actions based on continuous feedback from their environment. This level of autonomy allows agentic AI to take on various roles, from managing routine processes to addressing complex collaborative projects across different platforms and systems.

Whilst there is still a lot of AI hype, there is also the realisation that AI technology will impact every business 74% of CEOs surveyed by Gartner in 2024 believe this is the case, and agentic AI has incredible potential. Wherever you are on your enterprise AI journey, Oakland is a data consultancy ready to help. Book a free AI workshop or speak to one of our experts.

What Sets Agentic AI Apart?

Agentic AI stands out because it can perform semi-autonomous actions within a business environment. Unlike traditional AI tools that primarily function as enhanced chatbots, agentic AI brings a higher level of cognitive engagement, reasoning, planning, and executing tasks that typically require human intervention. The technology is built on large language models (LLMs), which are trained to process and understand vast amounts of text, enabling them to undertake complex business functions that go beyond mere data entry or routine customer interactions.

Process Logging

This is a hugely important point – every step taken by our agentic AI is meticulously logged. Actions, decision points, and outcomes are all recorded in structured logs. These logs are invaluable for human review and provide a foundation for ongoing improvements and learning within the system through careful analysis and supervised learning.

Natural Language Processing and Task Parsing

Our agentic AI systems use state-of-the-art language models to break down instructions into actionable tasks. This process transforms complex natural language into structured task blueprints, meticulously crafted through advanced prompt engineering and template matching.

Task Planning and Sequencing

Agentic AI excels in organising tasks efficiently. It plans and sequences tasks into logical, branching pathways using well-defined action schemas and goal-oriented frameworks. These pathways consider all necessary steps while accounting for dependencies and constraints, ensuring a coherent and strategic approach to task completion.

Agentic AI vs. Generative AI

The distinction between agentic AI and generative AI lies in their capabilities and intended applications. Generative AI, a hot topic in recent tech discussions, primarily focuses on creating content – text, images, or code – from existing data. It excels in tasks such as drafting emails, generating reports, or producing artistic content, based on patterns it has learned from large datasets.

In contrast, agentic AI takes this a step further by generating content, making decisions, and taking actions based on its understanding. While generative AI can suggest actions, agentic AI can actually execute these actions within certain parameters, thereby acting as a semi-autonomous agent within an enterprise. This ability to not only suggest but also enact makes agentic AI a powerful tool for businesses looking to automate more complex and decision-driven tasks.

The Evolution of Business Automation

For decades, businesses have aimed to offload as many tasks as possible to machines – from calculations done by software like Excel to customer service interactions handled by chatbots. However, these tools have always been limited to specific, predefined tasks. Agentic AI changes the game by applying its reasoning capabilities to a wider array of activities, thereby enhancing productivity and allowing human employees to focus on more critical, less mundane aspects of their work.

Integrating Agentic AI into Modern Workflows

The integration of agentic AI into business workflows is transformative, enabling not just automation but smart automation. By understanding and acting upon natural language, these AI agents can manage and sequence tasks effectively, leveraging existing tools within the enterprise to enhance workflow. This capability allows businesses to augment their workforce’s productivity without necessarily expanding it, a critical consideration in today’s economic climate where efficiency and cost reduction are paramount.

The Practical Applications of Agentic AI

One of the most compelling uses of agentic AI discussed involves its deployment within large IT sales organisations. AI agents interpret and organise sales data, translating messy, inconsistently recorded entries into structured, actionable information. This capability is crucial for companies that must accurately understand their sales patterns and inventory needs, helping drive informed business decisions and strategic planning.

Oakland Agentic AI Case Studies

With the AI hype, you could be forgiven for thinking you might get left behind. For many businesses, if you’re in an industry which isn’t yet being reinvented by AI don’t worry you can afford to go at a more steady pace.

So where do we expect businesses will get the biggest bang for their AI buck? Gartner research shows that an average enterprise expects 47% of AI benefits from employee productivity, 33% from process improvement, and 20% from business model innovation.

Oakland is seeing the biggest benefits in smoothing out clunky, routine business processes to allow employees to concentrate on more interesting and complex problems. From our recent research, knowledge management is a key area in which Agentic AI technology can really start to prove business value (link to our report) 

In a utility-sector example, frontline operatives must triage a high volume of alerts from their infrastructure network. Assessing the urgency of these alerts involves analysing complex, often unstructured data such as environmental factors, recent events, and historical infrastructure issues. Agentic AI steps in to streamline this process. By processing all inputs and applying reasoning capabilities, the AI can prioritise alerts and suggest actions, effectively replicating the decision-making process of human operatives. This saves time and ensures that critical issues are addressed promptly, enhancing operational efficiency and reducing risks.

Outside of knowledge management, we’ve found that large sales organisations often face challenges in managing extensive and fragmented product catalogues. With thousands of products listed, duplication and inconsistency in naming conventions, often due to decentralised data entry, can lead to significant inefficiencies. For example, the same product might be entered multiple times with vague or inconsistent descriptions, making it difficult to classify and report accurately. This lack of a unified taxonomy can create obstacles when aligning products with new organisational frameworks designed to improve financial reporting or operational insights.

Enter Agentic AI… Using Large Language Models (LLMs), businesses can process and understand product descriptions, cross-reference them with external data, and align them to standardised names and categories. These AI tools can manage vast datasets, accurately mapping products to a consistent taxonomy, even when descriptions are incomplete or unclear. Additionally, they can automate the classification process while incorporating exception-handling workflows for items requiring manual review. This approach enhances operational efficiency and provides more accurate and actionable data, enabling organisations to make informed decisions and deliver improved financial and operational insights.

These are just a few examples of the types of projects Oakland is currently working on, but there are hundreds of others.

Further Benefits of Implementing Agentic AI

Implementing agentic AI in business operations streamlines processes, significantly reducing the need for manual intervention and accelerating task completion. This technology enhances efficiency and boosts accuracy and consistency, minimising human errors and ensuring that tasks are executed reliably. 

With its scalable nature, agentic AI can manage expanding workloads without corresponding increases in human resources. Furthermore, it liberates human employees to concentrate on strategic and creative tasks, thereby adding substantial value beyond routine activities.

Here are the benefits of incorporating agentic AI:

Enterprise-wide System Integration

Agentic AI streamlines integration by interacting with existing information systems much like a human user but without requiring bespoke API development. These agents can navigate multiple interfaces, seamlessly transfer data between systems, and ensure consistency across your technology stack. For example, an agent could efficiently synchronise customer data across your CRM, billing system, and support platform, managing format conversions and validation checks with ease.

Autonomous Operation and Self-correction

Agentic AI operates with minimal human oversight, thanks to its advanced error detection and resolution capabilities. These systems autonomously adjust their strategies when encountering obstacles, reducing dependency on human intervention. For example, a document processing agent might automatically detect and correct formatting inconsistencies and log its actions for future review, ensuring accuracy and compliance.

Boosted Team Productivity and Strategic Focus

Multi-agent systems can take over routine operations and complex analytical tasks that traditionally consume significant human resources. This allows teams to focus on strategic initiatives and creative problem-solving. A research agent, for instance, could continuously analyse market trends, compile data, and generate insights, freeing analysts to focus on developing detailed strategic recommendations.

Dynamic Real-time Optimisation

AI agents excel at improving operational efficiency through real-time monitoring and adaptive decision-making. They process streaming data, identify emerging patterns, and implement timely improvements, which is invaluable for complex, multi-step processes. For example, a manufacturing agent might adjust machine parameters and coordinate maintenance schedules to optimise production metrics and improve output quality.

Scalable Learning and Knowledge Management

A standout feature of agentic AI is its ability to learn from diverse interactions and apply this knowledge across different scenarios. This ability enhances decision-making and problem-solving across the organisation. For instance, a customer service agent could identify effective resolution strategies from thousands of interactions, share these insights with the support team, and refine its approach based on feedback and emerging trends.

These benefits underscore how agentic AI can transform various aspects of an organisation, driving efficiency, adaptability, and strategic depth across operations. Companies can unlock substantial potential by adopting agentic AI, positioning themselves for success in an increasingly automated and data-driven business landscape.

The Challenges and Considerations

Despite its benefits, deploying agentic AI comes with challenges, particularly concerning governance. As these AI systems can perform actions semi-autonomously, businesses must implement robust governance frameworks to ensure that AI actions align with company policies and ethical standards. This includes setting up processes to review and approve AI-generated decisions and actions, ensuring they are appropriate before being enacted.

The Future of Agentic AI

As we look towards 2025 and beyond, the landscape of agentic AI is promising but requires careful navigation. These AI agents could take on increasingly complex tasks within the business sector, potentially leading to a new era of productivity and innovation. However, this will also require businesses to stay vigilant about ethical and practical implications of semi-autonomous systems.

Agentic AI represents a significant leap forward in business technology. By shifting the cognitive load from humans to machines, businesses can streamline operations, enhance employee productivity, and focus on innovation and strategic growth. As we advance, it will be crucial for companies to balance the capabilities of agentic AI with thoughtful governance and a clear understanding of its role within the broader business ecosystem. This balancing act will determine how profoundly agentic AI can transform business operations in the coming years.

Strategic Implementation of Agentic AI

To successfully implement agentic AI, organisations should identify areas where agentic AI can make the most significant impact, focusing on processes that benefit from automation and decision-making capabilities. Partnering with AI specialists who can provide insights and support throughout the implementation process is crucial. Regularly reviewing the performance of AI systems and making adjustments as necessary to optimise their functionality is essential for maintaining efficacy.

Read our guide below for further insight into crafting a data strategy that integrates with AI.

Agentic AI is reshaping how businesses operate, offering powerful tools for automation, decision-making, and process optimisation. By understanding and integrating agentic AI, organisations can achieve greater efficiency, improve accuracy, and better allocate their human resources towards growth and innovation.

Looking to enhance your organisation’s data capabilities? Discover how a data consultancy like Oakland can assist you. Dive into our AI services or visit our main knowledge management guide for deeper insights into our support offerings.

The post What is Agentic AI? appeared first on Oakland.

We start 2025 with a new term, but the same premise Agentic AI and AI Agents are where the market is heading. Autonomous systems or ‘AI agents’ perform all of those mundane (and increasingly not so mundane tasks that we never get around to doing). If 2024 was about figuring out how to take models into production, 2025 will be the year when we see agents emerging as first-class citizens in enterprise environments.

Where we started: The AI landscape at the beginning of 2024

At the outset of 2024, the AI world was still buzzing from the release of GPT-4, which solidified OpenAI’s leadership in the generative AI space. GPT-3.5 Turbo was being phased out as GPT-4 became the new standard. In the open-source arena, Meta’s LLaMA 2 was gaining traction, offering businesses an alternative to closed, proprietary models. Meanwhile, Anthropic continued to develop its Claude series, and Google’s Gemini 1.0 arrived, signalling serious competition in the race to dominate AI capabilities.

At the same time, Retrieval-Augmented Generation (RAG) started to gain serious traction in enterprise contexts. This approach—which combines large language models with external data retrieval—became the default architecture for deploying AI systems that required domain-specific knowledge. Tools like Azure OpenAI’s implementation of RAG and Google’s NotebookLM began offering practical solutions for integrating business data into AI systems. By the end of the year, deploying an AI solution without integrating it with your organisation’s data became unthinkable.

Production the key focus of 2024

While 2023 was the year of experimentation, 2024 was the year of making AI operational. Enterprises moved beyond proof-of-concepts (PoCs) to attempt production-grade implementations of LLMs. Yet, this process proved anything but smooth. The Gartner statistic that 85% of generative AI projects fail to make it into production rang true, illustrating the unique challenges that this technology presents.

The Production Problem

Deploying LLMs into production is fundamentally different from traditional software. In 2024, businesses grappled with challenges such as:

Cost considerations

Dedicated AI budgets, in the main, are few and far between. Data teams are being challenged to do more with less. Add to this the double whammy that most AI projects are far more expensive than initially projected. Gartner reported that many projects cost 300 to 500 times more than anticipated, due in large part to infrastructure, security, and integration costs.

Data integration

LLMs need access to enterprise data to add value. Whether via APIs, RAG pipelines, or custom integrations, aligning AI capabilities with business data became a priority.

Infrastructure limitations

AI requires massive compute resources. Many organisations found themselves constrained by global hardware shortages. For example, in the UK, models have to be deployed in Sweden because the necessary GPUs are simply not available locally.

Security and governance

The biggest challenge we see and one that the biggest tech vendors in the globe are wrestling with. Moving your enterprise data into AI workflows raises serious concerns around data sovereignty and governance. Even Microsoft’s Copilot could not guarantee that enterprise data would stay within a specific geographic region; an issue that created significant friction for regulated industries.

The SaaS boom

To address these production challenges, 2024 saw an explosion of Software-as-a-Service (SaaS) offerings targeting AI deployment. These tools aimed to simplify everything from RAG implementation to software development augmentation. For instance:

Copilot emerged as the dominant enterprise AI tool, thanks to Microsoft’s vertically integrated stack.

Third-party solutions like Cursor, Devin, and Windsurf tackled specific pain points for developers, automating workflows and improving coding efficiency.

Numerous RAG-focused SaaS platforms offered plug-and-play solutions for integrating enterprise data into LLMs, such as vector database providers like Pinecone and open-source alternatives like Weaviate.

The result? Companies could adopt AI solutions faster, but they also became increasingly dependent on these third-party tools to abstract away the complexity of production deployment.

The Physical Infrastructure Bottleneck

A major theme of 2024 was the physical infrastructure constraints facing AI. As global demand for compute power skyrocketed, manufacturers like TSMC and NVIDIA struggled to produce enough GPUs to keep up. Cloud providers like Azure, AWS, and Google Cloud were unable to provision sufficient hardware in all regions, forcing businesses to look for creative solutions.

This scarcity brought back challenges reminiscent of the 1990s on-premise era: organisations had to think carefully about where their workloads would run and how physical infrastructure limitations would impact their deployments. A system that worked seamlessly in Sweden, for example, might be impossible to deploy in the UK. As AI models become increasingly powerful, ensuring access to hardware remains a critical factor for enterprise success.

Enterprise Adoption Trends in 2024

Several trends emerged in the enterprise AI space this year:

1. Knowledge Management as the Leading Use Case

Our research found that knowledge management was the top enterprise use case for AI in 2024. Tools like RAG-based systems made it possible to extract insights from vast documentation repositories, improving workflows in industries like transportation, legal, and healthcare. For example, Oakland launched Network Rail’s first AI programme by deploying an AI-driven knowledge management system to surface lessons learned from past incidents, providing actionable insights for operations teams.

2. Custom Copilots

Enterprises increasingly embraced custom Copilots, integrating Microsoft’s Copilot platform with their own data to create tailored AI assistants. Rather than treating Copilot as a glorified spell-checker or summarisation tool, businesses began leveraging it for deeper productivity gains. Custom copilots became the new standard for enhancing workflows in applications like SharePoint, Teams, and Outlook. If you’ve seen the keynote speech at Microsoft Ignite you may be forgiven for thinking launching an AI agent or custom Copilot was as simple as creating a Power Point presentation. Sadly this is far from the reality.

3. The Imagination Gap

Despite these advancements, many organisations still struggled to identify where AI could deliver value. As one client insightfully put it: “I thought this was just a search bar.” Bridging the imagination gap remained a key challenge in 2024. Demonstrations of AI capabilities—particularly in workshops and proofs of concept—became essential for helping businesses visualise AI’s potential.

Preparing for 2025: The Rise of Agents

If 2024 was the year of LLM production, 2025 will be the year of AI agents. These systems will go beyond answering questions or summarising documents to autonomously take actions on behalf of users.

What is an AI Agent?

An AI agent combines a large language model with tools, workflows, and integrations that enable it to perform tasks independently. For example, a customer service agent might:

Retrieve information from a knowledge base (RAG).

Analyse a customer query.

Update a CRM system or initiate a workflow without human oversight.

These semi-autonomous systems will begin with routine tasks, such as scheduling, updating records, or generating reports. Over time, they will evolve to handle increasingly complex and non-routine business actions.

Risks of Agentic Systems

While agents hold immense promise, they also introduce new risks:

Capability gaps: Are LLMs capable enough to handle critical business tasks? The limitations of current models will be tested as organisations push agents to take on more responsibility.

Oversight and control: How do we ensure that agents operate within predefined guardrails? Governance frameworks will need to evolve to account for AI systems acting autonomously.

Trust and accountability: Who is responsible when an AI agent makes a mistake? Ensuring transparency and accountability in agentic systems will be critical.

Lessons Learned from 2024

As we look ahead to 2025, there are several lessons that enterprises can take from the past year:

Stay close to vendor roadmaps: The pace of AI development is unprecedented. Whether you’re using OpenAI, Microsoft, or Google tools, keeping a close eye on their roadmaps is essential to avoid obsolescence.

Build for flexibility: Given how quickly AI models evolve, organisations must build solutions that are adaptable to new technologies and frameworks.

Prioritise data integration: AI without access to enterprise data is just a toy. Ensuring seamless, secure integration between LLMs and business data sources is non-negotiable.

Invest in security and governance: Understanding where data is stored, how it moves, and what security wrappers are in place is essential for any AI project.

Start small, scale thoughtfully: AI projects are expensive, complex, and prone to failure. Begin with focused, high-impact use cases and scale gradually as you learn what works.

Conclusion

2024 has been a foundational year for AI. Enterprises have begun to move beyond experiments to deploy large language models into production, but challenges around infrastructure, security, and cost remain significant. The rise of SaaS solutions has made AI more accessible, yet the complexity of integration persists.

Looking ahead, 2025 will bring the next evolution: Agentic AI and AI agents that autonomously perform tasks and drive measurable business outcomes. As organisations prepare for this shift, the lessons of 2024 will serve as a valuable lesson.

Looking Ahead to 2025

As we screech into 2025, the role of AI and data within organisations will continue to accelerate and evolve. Several key trends and priorities are set to define the year ahead for data leaders, presenting both challenges and significant opportunities:

Operationalising AI at Scale

While 2024 saw a wave of AI experimentation and pilot projects, the focus in 2025 will shift toward scaling these initiatives to drive measurable ROI. Leaders will prioritise robust AI operationalisation frameworks, integrating models into production systems while ensuring governance, efficiency, and reliability. This will require alignment between data infrastructure, processes, and teams to ensure AI initiatives deliver tangible business value.

AI Governance and Ethical Frameworks

As regulatory environments mature and AI becomes increasingly embedded in decision-making, governance will move to the forefront of the AI agenda. Data and AI leaders must ensure their models adhere to ethical, transparent, and explainable practices, especially in sensitive industries such as finance, healthcare, and public services. Implementing strong governance policies that align with emerging regulations like the EU AI Act will be paramount.

Focus on Data Quality and Unified Platforms

Scaling AI is impossible without trusted, high-quality data. In 2025, data leaders will invest in modern data platforms that centralise, clean, and govern data across organisational silos. Unified platforms will become the backbone for scalable analytics and AI, enabling faster time-to-insight while reducing the friction between engineering and business teams.

Generative AI Maturity

The generative AI wave will mature further, with organisations moving from experimentation toward targeted, high-impact use cases. In particular, enterprise adoption of generative AI for content generation, code development, knowledge management, and customer experience will drive competitive differentiation. The demand for fine-tuning and customising foundational models to fit specific business needs will rise.

Investment in AI Talent and Upskilling

AI and data teams will continue to face talent gaps as demand outstrips supply. Organisations will double down on upskilling initiatives, developing AI fluency across the enterprise not just for technical teams but also for business leaders and frontline employees. Data and AI leaders will need to foster a culture of collaboration, pairing technical expertise with domain knowledge to drive innovation.

Measuring Value from AI and Data

In a challenging economic environment, data and AI leaders will face increasing pressure to prove ROI from their investments. Organisations will emphasise tracking clear KPIs, aligning AI and data initiatives with strategic business outcomes, and ensuring budgets are directed toward high-value priorities.

AI for Sustainability and Social Impact

2025 will also see AI playing a growing role in addressing sustainability and ESG (Environmental, Social, and Governance) goals. From optimising energy consumption and reducing waste to improving social equity, forward-thinking organisations will leverage AI to balance profit with purpose.

Our Closing Thoughts

As we move into 2025, the organisations that work hard to bridge the AI ‘imagination gap’ will thrive. Overcoming budget constraints and pressure to deliver a strong ROI will need data leaders with a strong stomach. Balancing innovation with governance, agility with resilience, and experimentation with measurable outcomes is no mean feat. But in the world of AI – who dares wins.

At Oakland, we provide pragmatic and actionable ways to get started with AI. In a rapidly evolving technology landscape, you need experts who can anticipate what’s next and align advanced AI—like Agentic AI—with your existing platforms, strategy, and governance frameworks. Oakland’s AI consultants can take the legwork out of putting AI to work across your organisation. Get in touch today or learn how we approach AI.

The post Agentic AI and AI Agents: Shaping the Market in 2025 appeared first on Oakland.

As generative  AI continues to advance at pace, more businesses are adopting it into their data management and day-to-day operations. Picking off-the-shelf tools for your AI is simple and straightforward. Still, these ready-made generative systems don’t work seamlessly with your business, thanks to their one-size-fits-all approach. An Intelligent Agent, on the other hand, creates a new and fresh spin on generative AI, functioning as a virtual worker and performing the exact tasks your company requires. 

In this guide, we’ll uncover all the benefits of an Intelligent Agent approach to your AI data strategy. Need more context? Explore our AI guide for a broader look at how we use artificial intelligence to drive insights from your data. 

What is an Intelligent Agent in AI?

Intelligent agents are a natural evolution of the LLMs we are currently seeing enter the market. They are layered on top of large language models (such as OpenAI’s GPT model) and provided with specific instructions. At the core of an agent is the instruction to create a plan in response to a given problem – and then execute that plan. The artificial agents are also provided with powerful tools and a memory store to allow them to actually DO things for your company. These tools can be as simple as the ability to respond in plain text – but can be as complex as writing results back to a database, alerting employees through text, or creating graphical executive summaries. 

Memory stores are also crucial to the work of an agent. They can be short term stores for use immediately, or a significantly larger body of corporate knowledge to be drawn on. 

By leveraging these resources, artificial agents function effectively as artificial workers in your organisation.

What Makes Intelligent Agents Special?

There are a few key traits that make Intelligent Agents so incredibly useful, including:

• Autonomous Decision Making: Intelligent Agents are programmed to make decisions about information without needing direct human intervention.
• Goal Orientation: Each Intelligent Agent is designed to achieve a specific goal, resulting in highly targeted and deliberate outputs.
• Adaptability: Intelligent Agents can reconfigure themselves continually to environmental changes, whereas more basic AI would require expensive and time-consuming retraining to handle similar changes.
• Interactivity: Intelligent Agents work with their environment. 

What is a Knowledge-Based Agent in Artificial Intelligence?

Knowledge-based agents are a type of Intelligent AI Agent with a knowledge base inference engine that applies logical rules to an existing knowledge base. For example, a knowledge based agent can read large quantities of documentation that your employees simply wouldn’t have time to work through in order to understand and respond to specific requests.

This makes a Knowledge-Based Agent highly adaptable to your company – an AI that’s truly ready for your specific needs. If knowledge management is a key issue for your company, discover how we’ve used Knowledge-Based Agents to automatically pull data for clients like Network Rail. 

What are some Examples of Uses for Intelligent AI Agents?

Intelligent Agents are so customisable they can be used in a variety of applications to make your workers’ lives easier and drive ROI. 

Project Analyst

Free up your human project analysts from repetitive analytic tasks using an Intelligent Agent with capabilities to interpret free text project documentation, interrogate supporting data and summarise insights. 

Asset Historian

An Intelligent Agent can work as an Asset Historian by reading and supporting asset records, interrogating supporting data from many different datasets and providing a single comprehensive view of the asset.  It can also answer questions about the asset from staff.

Resource Allocation

Assign tasks effectively to the right employees within your company using an Intelligent Agent. For instance, you could feed CVs into the AI, with which it could then build and populate a skills matrix used to optimise task allocation.

Data Science Co-Bot

Take some of the weight off of your data science team with an Intelligent Agent that can create and execute code to perform ad hoc data preparation, mastering routine task management.

Alert Triage

Read and filter telemetry alerts for your control engineers with an Intelligent Agent using control room policies, asset history and alarm description analysis.

Why Develop an AI Intelligent Agent with Oakland?

Oakland’s Intelligent Agent Framework allows for swift and simple creation and customisation of your generative AI tools so that it fits naturally into your workflow. 

It works with your existing data and systems, preventing you from having to take time-consuming detours. Plus, we can use proprietary managed services, open-source technologies or both for a high degree of flexibility and control in your customisation. 

We also prioritise your security so no data will ever leave your business or be pushed to public-facing AI models. Our years of experience working with security teams mean you can trust us. And we are also proudly compliant with industry standards, including CIS and ISO 27001. 

With forty years of experience, Oakland knows company pain points around data and how decisions are made, and we use that knowledge when creating the ecosystem of virtual agents that we can deploy into your business. 

Explore Oakland’s AI service offering to create an Intelligent Agent that will transform your data management, and contact us to learn more about what we can do for your data. 

The post Why We Recommend the Intelligent Agent Approach to Generative AI appeared first on Oakland.

Last week, I had the opportunity to attend the highly anticipated Big Data and AI World conference, an all-encompassing event that showcased the breadth and depth of the data industry. From the infrastructure backbone of data centers to the cutting-edge applications in machine learning (ML) and large language models (LLMs), the conference offered a comprehensive look at the current state and future directions of data technologies. 

Highlights 

The conference prominently featured a range of topics within the data pipeline, emphasizing well-established processes including DevOps, DevSecOps, and Cloud Security. While the ML and LLM segment of the conference was relatively modest, it provided a valuable glimpse into the evolving landscape of AI applications. 

Key Topics and Speakers 

One of the opening sessions, aimed at a broad audience, focused on how to use ChatGPT, indicating a general entry-level understanding of LLMs among attendees. Despite this, there were notable discussions and presentations that delved into more innovative applications of LLMs: 

• Pangeanic highlighted their work on leveraging LLMs over data, showcasing how AI can enhance data analysis and insights. 
• TAAP demonstrated their initiative in developing applications powered by LLMs, pointing towards practical uses of AI in software solutions. 
• A showcase involved a company that developed a full-stack LLM application, extending from software down to the hardware level, emphasizing the potential for integrated AI systems. 

Microsoft also made an appearance with their Copilot solution, underscoring the growing interest in AI-assisted development tools – as well as a few resellers of Copilot. 

Notable Exhibitors and Innovations 

Among the exhibitors, Compare the Market stood out for their advanced development work with LLMs. They shared a series of best practices that they’ve implemented, offering valuable insights into the practical challenges and solutions in deploying LLMs in a business context. 

Equally compelling was the announcement from Multiverse, a pioneering quantum computing company, which presented a revolutionary algorithm capable of significantly condensing the size of large language models (LLMs), such as LLAMA2. Their technology has achieved an astounding 85% reduction in the model’s size with a mere 5% loss in accuracy. This breakthrough paves the way for more sustainable and accessible LLM implementations, reducing computational demand and making it feasible to deploy sophisticated AI applications in more constrained environments. 

These exhibitors highlight the innovative spirit permeating the conference, showcasing not only the practical application of LLM technologies but also the forefront of AI and quantum computing research. 

Personal Takeaways 

The conference provided a fascinating overview of the data and AI industry’s current landscape, with a broad spectrum of technologies on display. However, one of my key observations was that the application of large language models (LLMs) among exhibitors was primarily at a superficial level. Despite the potential of LLMs to act as powerful reasoning engines, their utilization seemed to be confined to more basic tasks such as content generation, serving as advanced chatbots, or simplifying interactions with data. 

This suggests that there’s a significant gap between the capabilities of LLMs and their current usage in the industry. Many exhibitors are yet to explore the full maturity scale of these models, leveraging them beyond the initial layers of application to truly tap into their transformative potential. The reliance on LLMs for relatively straightforward tasks underscores a broader industry trend of cautious engagement with AI’s deeper functionalities. 

Reflecting on the conference, it’s clear that there’s a vast untapped potential for LLMs to revolutionize not just how we interact with data, but how we reason with it, use it to make decisions, and innovate. The journey of integrating LLMs more profoundly into our technological solutions is just beginning, and I’m excited to see how they will be pushed beyond their current boundaries to achieve their full transformative potential. 

Author: Mike Le Galloudec is an Innovation Lead at Oakland

The post Everything you need to know about Big Data & AI World appeared first on Oakland.