Presentation at Digital Research 2013, University of Oxford. ExaMiner is a "research intelligence" tool that analyses the University of Bristol's research web sites and corporate databases to produce new ways of searching, accessing and visualising its research landscape. We discuss the rationale for and implementation of four proof-of-concept web applications that profile researchers' publications, web pages and non-confidential staff details before matching them to each other or against submitted text. We also report on the use of ExaMiner to support a multidisciplinary workshop between researchers from across the university and local health trusts. Historically at Bristol, as elsewhere, researchers and their projects are organised into departments, faculties, schools and so on, largely on the basis of their discipline. An unfortunate side-effect of this organisational structure is that researchers tend to know of other researchers and projects within their own branch of the organisational hierarchy. It is not untrue to say that they know more about researchers and projects elsewhere around the world than in their own institutions. This can leave researchers unaware of potentially relevant research going on elsewhere within their own university. Recognising that many important research areas now span multiple disciplines, the University Research Committee identified a need to produce a research intelligence tool to mine and map its research landscape. The aim being to provide its researchers and its Research and Enterprise Development group with new ways of searching, accessing and visualising connections between existing research -- irrespective of the organisational boundaries and structures. The resultant ExaMiner demonstrators, listed below, are e-Science/e-Research workflows built using the SubSift RESTful web services framework, augmented with secure web service wrappers around queries to corporate databases. In many cases, these tools give sensible results, but not always. SubSift compares textual documents using the vector space model from information retrieval, demoting the significance of words that occur in most of the documents and promoting words that occur rarely. In ExaMiner, because of the heterogeneity of staff homepage contents, rare but irrelevant words can take on exaggerated significance. Experimentation suggests that this may be countered by calculating significance within research groups rather than across the whole department or institution. Also, expanding the range of stopwords to be specific to academia will filter out many trivial connections (e.g. ignoring words like 'conference', 'international', 'workshop') regardless of their computed significance. Using ExaMiner to support a multidisciplinary workshop surfaced a new requirement. The organisers had conducted a pre-workshop survey of delegate's research interests and expectations for the event. Possible future work might scale-up ExaMiner to handle more data.