Steve Salvin, Founder & CEO of Aiimi – Interview Series

Steve Salvin is the founder and CEO of Aiimi, an AI platform which has been quietly scaling since 2013. Having bootstrapped the company since launch, Steve has grown Aiimi to 8-figure revenues and their tech is used by the likes of the FCA, PwC, and the UK government.

Steve has been working in tech since the 80s (even studying AI at university) and is a serial entrepreneur. He's a huge believer in building AI and technology that empowers users and gives them more control.

Aiimi's tech allows companies to find and make sense of their data – bringing together what can be a sprawling mass of data, documents, and digital information – helping teams access information instantly.

You initially studied AI over 30 years ago, what were you studying and what initially attracted you to the field?

I studied Computation during my Bachelor of Science at the University of Manchester in 1987. It was a natural choice; following a fascination for computers that started when I was just a child. I remember the day I got one of my own at the age of 12 – I spent hours teaching myself to code in my room. Since my school didn’t teach computing, I went on to convince my teachers to effectively rearrange the timetable, so that I could fit in lessons at a nearby college that did. This paved the way for me to study Computation at university, where I even studied a module in AI all those years ago. My career has been in this field ever since. Ultimately, it’s my passion for computers that first led me here and which has kept me here.

Could you share the journey behind your first startup that deployed next-generation document management and workflow technology, and what your key takeaways from this experience were?

I founded my first startup in 1996, having been inspired by a project I was working on in my previous role at PwC. The project opened my eyes to the challenges waiting to be solved through enterprise content management, and more innovative workflow technology. And so I set up APS. We built next-generation content management tools and worked with the likes of HBOS and Bupa. We grew quickly and were a 20-strong team within just two years. I learned a lot in these early days as a first-time founder, but the key takeaway was don’t be afraid to make decisions – if it turns out to be a bad decision you can make another one. You have to keep moving in a fast-moving startup environment.

Could you share the genesis story behind Aiimi? 

After I sold APS, I worked for OpenText for a few years. But I found that I missed being in the driver's seat and working closely with customers. Plus, my time in the industry up to this point had opened my eyes to a problem that wasn’t going away: the huge disconnect between transactional structured data and unstructured content within organisations.

It struck me how many businesses lacked insight into what data was being created, shared, and used across their teams, and how. The more data businesses accumulated and stored in different systems, the worse issues became. And, unable to find the information they needed to do their jobs, staff were being slowed down, doubling up on work, and making misguided decisions. This was the problem I set out to solve when I founded Aiimi in March 2007. Our mission is to help businesses find and make sense of their data, giving them the information they need to unlock efficiencies, spot opportunities, and de-risk their operations. Put simply, we connect people to insight.

Your technology is used by significant clients like the FCA, PwC, and the UK government. What makes Aiimi's AI platform stand out to such high-profile clients?

We’re extremely proud that our customers choose to work with us as a homegrown British AI business. Our approach is what helps to set us apart. We’re not just a software vendor; we’re an experienced team of data, digital, and AI experts who thrive on understanding the precise details of the challenges our customers face. It's our firm belief that technologies like AI should be used in an ethical way that gives users more, not less, control over their data. We also strongly believe there is no “one size fits all” approach to data management.

We invest a huge amount of time into getting to know our customers so that we can deploy the right technology solutions and tailor our services in line with the needs of individual organisations. And since we understand that these needs change over time, we continue to work closely alongside our customers to evolve our offering throughout our relationship with them.

Besides our industry-leading consulting services, the other thing that sets us apart is the cutting-edge technology that we’re able to offer our customers. Our continued investment in Aiimi’s IP, the Aiimi Insight Engine, alongside generative AI and emerging technologies, ensures that we’re able to serve the most novel use cases around, and solve even the thorniest of business challenges.

You're a proponent of building AI that empowers users and gives them more control. Can you elaborate on how Aiimi's tech achieves this and the impact it has on your clients' operations?

At Aiimi, we believe that AI should give users more, not less, control over their data. AI should be a driver of data quality and brand-new insights that genuinely help businesses make their most important decisions with confidence. That’s why we build AI tools that help organisations see their entire data picture, automate data governance, and enable them to get to the answers they need. Since well-governed data is behind any secure and successful AI application, our tools also put the power in organisations’ hands to adopt models more widely in a safe and controlled way – when we can get an organisation’s trickiest, most complex unstructured data into the right format and of a quality that can be used by AI models, that’s where they can unlock real business value.

Our platform also gives authorised users full visibility into our AI-powered answers. We use fully explainable approaches to AI , so that users with permission to do so can use the platform’s interactive dashboards to look “under the hood” and see exactly what data models are working with, what insights they’ve gleaned, and how they arrived at them. This gives our customers a comprehensive understanding and audit trail for how their data has been used to inform decision making; a hugely important step in making AI-powered answers usable and safe for enterprises.

With the Aiimi Insight Engine, you aim to solve the problem of underutilized data within enterprises. Could you explain how the engine works and the kind of insights it has unearthed for businesses?

A typical enterprise uses hundreds of different systems to store data. The problem is that these systems quickly become outdated and often don't speak the same language. As a result, the information within them is lost or forgotten about, and is impossible to find when employees need it. Recent Gartner research shows that 47% of digital workers struggle to find the information needed to effectively perform their jobs. The Aiimi Insight Engine addresses this disconnect by creating a data mesh layer on top of an organisation that connects these disparate sources.

The Aiimi Insight Engine discovers, enriches, and joins-up information so that it can be instantly accessed by those who need it – plus,brand-new insights are unlocked by combining previously disconnected datasets, like structured telemetry data and unstructured customer call transcripts. This helps teams realise efficiencies, and glean the insights needed to solve business challenges and spot opportunities. At the same time, the tool pinpoints and secures sensitive information, helping organisations de-risk their operations. Naturally, the data insights and possible use cases vary hugely from one organisation to the next. This is why we work closely alongside our customers, to help each one get the best out of the Aiimi Insight Engine.

Our recent work with a Government department offers a good example. Their team of analysts needed to access accurate information summaries from data from multiple sources, to supply timely and precise briefings to the government and public. They typically used open-source data, like trusted news outlets and websites. But with this volume of data ever-increasing, finding, retrieving, and collating all this information had become increasingly difficult. They required an AI-powered solution to streamline this process and enable them to create these briefings more efficiently and effectively. They chose the Aiimi Insight Engine for its ability to intelligently process large datasets – in this case, those news sources and websites – to find relevant information, before transforming this data into a consumable set of insights using secure Generative AI and Extractive AI models. With the help of our technology, they were able to increase their efficiency and enable more effective decision-making.

The risks of ‘shadow’ AI can be substantial for businesses. Could you define what shadow AI is and discuss the risks associated with it?

‘Shadow AI' refers to when employees bolt AI tools (like ChatGPT) onto their work systems for the sake of ease and efficiency, without their employer knowing or consenting to the technology. Employees may have good intentions. But shadow AI can pose serious data security risks.

Firstly, employees may be feeding AI models sensitive information without realising – and there’s no guarantee this data won’t find its way into the public domain. Generative AI tools that haven’t been vetted by IT leaders may also be unreliable and produce inaccurate results, particularly if used for inappropriate use cases. Inaccurate results that appear to be incredibly convincing to users, known as ‘hallucinations’, often go undetected. And the consequences of poor decisions that follow have the potential to be hugely costly for companies.

To avoid shadow AI, it’s important to educate teams on what safe and ethical AI practice looks like, and provide clear guidance on which AI tools can and can’t be used securely at work. I would advise avoiding public large language models altogether when it comes to your corporate data. Instead, invest in safe, reliable and robust AI tools that enable workers to do their jobs effectively and efficiently. This way, staff won’t need to resort to unauthorised, potentially insecure tools in the first place.

Implementing enterprise AI safely and securely is critical. What steps does Aiimi take to ensure the security and integrity of its AI solutions?

Security is baked into the Aiimi Insight Engine and all of our technology. Our built-in AI & Data Governance toolkit gives our customers complete control, so they know exactly where their personal or sensitive data lives (and can remediate any issues) and how our AI platform operates in their business. They can also “track and trace” every user interaction with their AI system and auto-verify the sources used for AI-generated answers with full traceability and data lineage. Because we use a range of AI models, each auto-selected by our enterprise AI platform for their reliability, security, and cost to perform a given task, we can tightly control factors like security, speed, and cost based on individual customer requirements. For example, ensuring AI models chosen are used in line with ISO 8000, or that they consider NCSC guidance, GDS, and/or DevSecOps principles.

Looking ahead, what future developments in AI and data management are you most excited about, and how is Aiimi preparing to integrate these advancements into its offerings?

The possibilities for AI-driven data insights are endless – particularly for organisations that get the fundamentals of data governance, data quality, and information retrieval right. At the moment, we are focusing on exploring the best use cases for GenAI in business and developing our product roadmap accordingly. For example, we’re helping customers identify where the most value can be realised from AI- such as by turning unstructured data into structured formats that can then be fed into BI reporting to support downstream decision-making. This is a great first use case for businesses just getting started with AI. It provides tangible benefits and has a fast, clear ROI. We’re also supporting our existing customers to take their next steps with AI and ensuring they get  the most out of the ever-evolving tech.

We are planning to expand our headcount over the coming year, too. We look forward to welcoming new members to our diverse and inclusive Aiimi community. This strengthened team will enable us to continue to grow, innovate, and enhance our offering in line with our customers' evolving needs.

Thank you for the great interview, readers who wish to learn more should visit Aiimi.