12
unstructured, like video files, images and PDFs, or redundant, obsolete or
trivial (ROT). Without usable data across the enterprise, most organisations
are limited to employing AI on specific siloes of structured information.
Instead of developing AI applications that can only perform niche tasks,
Microsoft is helping enterprises organise their structured and unstructured
data into a single, searchable knowledge base enriched with relevant external
data. AI deep learning capabilities run on top of data to understand the
content and infer relationships, turning data into knowledge and ensuring
that the user receives the most relevant information. Serving as a strategic
foundation for current and future AI initiatives, this knowledge base is also
understandable, consumable and searchable by AI models and can enable
enterprises to integrate AI into every aspect of their business.
From this centralised resource, employees can search, find and explore
previously unavailable enterprise information, enabling them to be more
efficient and impactful.
Enrich employee experiences with searchable
knowledge
Moreover, we are providing more user-friendly ways to democratise access
to this enterprise knowledge. With Microsoft AI, enterprises can build AI
experiences on top of their enterprise knowledge, either by infusing it with
existing productivity and business applications or by creating new
applications powered by AI, like conversational agents. For example, the next
generation of conversational agents can pull from customer knowledge and
world knowledge to answer questions specific to businesses like “Which
customers are within a 10-mile radius of the airport?” This knowledge base
democratises AI experiences and provides powerful new connections to
employees.
Consider the following illustration of the searchable knowledge base in
action: Lisa, a marketer, needs to find an advertising agency to use for an
upcoming campaign. Typically, identifying an agency with the right
capabilities would require relying on colleagues for recommendations and
spending hours reviewing past records, agency websites, web results and the
agency’s previous work. Lisa would have struggled to piece together this
information, manually combing through countless sources of data, never
certain if she had all the right information.
With an AI-infused knowledge base, Lisa can search “Find me advertising
agencies” and a ranked list of agencies her company used in the past will
be populated. Curious about her colleagues’ experiences, she selects two
agencies that her company has previously used and types a new query into
the search bar: “Find me Teams threads that mention these agencies.” Her
search yields dozens of Teams threads. Keen to quickly identify which agency
her colleagues have had the most positive sentiment about, she applies a
sentiment analysis AI model. She can now use more advanced queries like
“Show me Teams threads with negative sentiment for Agency A” or “Show
me the trend of sentiment for Agency B.”
Partnering with
Microsoft, Centrica was able to combine
their disparate data into searchable
knowledge, helping them improve
customer experiences, provide their
employees with more actionable customer
insights and expand the capabilities of their
AI-powered digital assistant, Wilbur.
In Dynamics 365 applications, the
Common Data Model (CDM) provides a
built-in shared data repository across all
applications – from Sales and Marketing, to
Customer Service and Field Service – as well
as external data sources, like SAP. Capturing
and synthesising data from across the
business makes it possible for enterprises to