The Knowledge Before AI Framework addresses the information, governance and workflow problems preventing artificial intelligence projects from producing measurable business value
Digital transformation and knowledge management expert Dr Tori Reddy Dodla has launched a new framework designed to help organisations establish the knowledge, data and operational foundations required for successful artificial intelligence adoption.
The Knowledge Before AI Framework responds to growing evidence that businesses are investing heavily in AI tools while struggling to move promising pilots into production or achieve measurable financial returns.
Gartner forecast that worldwide spending on generative AI would reach approximately $644 billion in 2025, representing an increase of 76.4 per cent from 2024.
Yet AI project failure remains widespread. RAND Corporation reports that, by some estimates, more than 80 per cent of AI projects fail, approximately twice the failure rate of information technology projects that do not involve AI. RAND’s research identified unclear business objectives, inadequate data, poor infrastructure and a focus on fashionable technology rather than real operational problems among the principal causes.
Dr Dodla argues that many organisations are attempting to introduce AI before establishing control over the knowledge and information on which it depends.
“AI does not create organisational knowledge,” Dr Dodla said. “It reflects and amplifies the knowledge, structure and data an organisation already has.
“When the foundation is disorganised, undocumented or nobody’s clear responsibility, AI produces confident sounding output on top of a broken base. Giving the organisation a newer or more powerful model does not repair that foundation.”
A framework for building the foundation first
The Knowledge Before AI Framework provides organisations with a structured approach for preparing their knowledge, data, technology and working processes before making further investments in AI.
It covers five connected areas.
Define the business value
Organisations begin by identifying the specific operational or commercial problem AI is expected to solve.
This shifts the focus away from adopting AI for its own sake and towards measurable objectives such as reducing costs, improving decision making, accelerating access to information, increasing service capacity or supporting new products and revenue.
Capture critical organisational knowledge
The organisation identifies the knowledge required to support the selected process or decision.
This includes information held in documents, databases, emails, collaboration platforms, individual employees and undocumented working practices.
The purpose is to establish what the organisation knows, where that knowledge is located and which information is missing.
Establish governance and ownership
Reliable AI requires information that is current, controlled and clearly owned.
The framework helps organisations determine who is responsible for maintaining critical knowledge, which sources are authoritative and how information should be reviewed, approved, protected and updated.
Without this governance, an AI system may retrieve several conflicting versions of the same policy, process or answer.
Connect knowledge, technology and workflows
AI must operate within the organisation’s real systems and working practices.
The framework examines how knowledge moves between employees, Microsoft 365, SharePoint, Power Platform, business applications and other information sources.
It also considers whether the existing workflow should be redesigned before AI is introduced.
Measure operational and financial results
The final stage establishes how value will be measured.
Relevant measures may include time saved, costs reduced, errors prevented, decisions improved, enquiries resolved, risks controlled and revenue generated.
Dr Dodla said organisations frequently measure AI adoption through the number of tools purchased, licences issued, employees trained or pilots launched.
“These are measures of activity, not business value,” she said.
“An AI project should begin with a defined problem and an agreed result. The organisation can then identify the knowledge, data, governance and workflow required to achieve it.
“The model should be selected after this work has been completed, not before it begins.”
Why promising AI pilots fail to scale
A controlled proof of concept can work with carefully selected information and a small group of users.
Moving the same technology into normal operations introduces changing data, conflicting documents, unclear responsibilities, security requirements, legacy systems and working processes that may never have been formally documented.
RAND’s research found that AI projects frequently fail because organisations do not have the data or infrastructure required to train, manage and deploy them successfully. It also found that business leadership misunderstanding the problem to be solved was the most frequently reported cause of failure among the AI practitioners interviewed.
Dr Dodla believes knowledge management provides the organisational discipline needed to close the gap between an impressive demonstration and a dependable operational capability.
Knowledge management is the practice of capturing, organising, governing and making an organisation’s collective intelligence usable.
It determines what the organisation knows, where reliable information is stored, who owns it, how it flows and how employees or technology can retrieve it when required.
“Feed an AI system fragmented and ungoverned information and it will return fragmented and ungoverned answers faster and at a greater scale,” Dr Dodla said.
“When knowledge is properly managed, AI can retrieve reliable information, support better decisions and improve real working processes. The technology becomes a multiplier of an organised organisation instead of an accelerator of its existing confusion.”
Turning organisational data into measurable value
The framework also encourages leaders to treat knowledge and data as strategic business assets.
Dr Dodla describes data monetisation as the process of using data deliberately to generate revenue, reduce operating costs, improve services, support decisions and create lasting organisational value.
The value may be direct, through a data based product or commercial service, or indirect, through greater efficiency, reduced duplication and improved customer or employee experiences.
Knowledge management and data monetisation must work together, she argues.
An organisation cannot reliably extract value from information it does not understand, govern or maintain. Organising knowledge without connecting it to a commercial or operational outcome can also become an administrative exercise with no measurable return.
“The companies waiting for AI to pay off are rarely one model away from success,” Dr Dodla said.
“They are one discipline away. The missing piece behind AI was never more intelligence. It was knowing what the organisation already knows, governing it properly and understanding what it is worth.”
Building on established knowledge management research
The Knowledge Before AI Framework builds on Dr Dodla’s doctoral research into knowledge management systems and the practical methods presented in her book, Mastering Knowledge Management Using Microsoft Technologies.
Published by Apress in 2024, the book explains how organisations can use Microsoft 365, SharePoint, Power Platform and Microsoft Copilot to create effective knowledge management systems without automatically purchasing overlapping third party technology.
The book includes case studies, practical exercises and guidance on knowledge bases, information governance, Power Apps, Power BI, document management, knowledge sharing, analytics and AI integration. It is written for technology, information, digital and knowledge management leaders.
The Knowledge Before AI Framework can support executive briefings, AI readiness assessments, knowledge management reviews, technology planning and digital transformation programmes.
Further information about Dr Dodla’s work and book is available at torireddydodla.com/book-landing.
About Dr Tori Reddy Dodla
Dr Tori Reddy Dodla is a digital transformation, knowledge management and artificial intelligence specialist.
She is the author of Mastering Knowledge Management Using Microsoft Technologies, published by Apress. Her work focuses on helping organisations capture and govern knowledge, use existing Microsoft technologies more effectively and build stronger foundations for digital transformation and AI.
A former Chief of Digital Services in the United States federal government and a former US Army officer, Dr Dodla has 15 years of experience spanning technology leadership in public and private sector environments.
She holds a PhD in Information Technology, with doctoral research focused on knowledge management systems.
Dr Dodla is a graduate of Carnegie Mellon University’s Chief Data and Artificial Intelligence Officer Certificate Program and has completed executive education in AI implementation and healthcare strategy at Harvard Medical School. Carnegie Mellon describes its CDAIO programme as covering enterprise data management, data strategy, governance, responsible AI and the conversion of organisational data into business value.
Media contact
Tori Reddy Dodla, PhD
Email: [email protected]
Website: torireddydodla.com/book-landing
Media ContactCompany Name: WLW Press Release DistributionContact Person: David HollyEmail: Send EmailCountry: United KingdomWebsite: wlwfuture.com
