Cognitive Computing and the Education Sector


How would students enquire, explore, learn and be assessed ... and how would teachers prepare, deliver, assess and administer their courses ... and how would support teams and administrators carry out their work to support the needs of students if they all had access to a personal cognitive assistant or if the software applications that they all used took advantage of cognitive computing? I ask these questions because the answers or solutions that arise from the use of cognitive computing are set to transform the way schools, colleges and universities deliver education services to their local and distributed communities.

Cognitive computing will bring about many new products and services to the education sector. These will include the advent of the personal cognitive assistant for students, teachers and support teams; the automated compiling and delivery of online learning and assessment materials to students; the automation of routine tasks and activities such as the production of student report cards; the greater personalisation of student facing services and more. In the course of this article I will explore the advent of these services; focusing particular attention on the role of the cognitive assistant as it becomes the main online conduit between students, teachers, support teams and the cognitive services that they will all come to use on their campuses.

Cognitive computing in the education sector is set to become ubiquitous. Every service that is used by an institution's students, teachers, support teams and administrators will take advantage of one or more aspect of cognitive computing. Careful thought and design of cognitive services will enable schools, colleges and universities to deliver enhanced services that will benefit the students and communities that they serve.

How will students, teachers, support teams and administrators take advantage of cognitive computing?
The following section describes a broad range of scenarios were cognitive computing could be used by individuals in schools, colleges or universities. Potential use cases may run into the hundreds. Think of any service or software application that is used in an education setting. There is a high probability that it will be enhanced or transformed by cognitive computing. Click on the following links to learn more.

I have also detailed how the various components of cognitive computing are currently being used by students and colleagues at Bolton College.

Cognitive computing scenarios for students

Here is a list of scenarios were cognitive computing could be used to support students during their studies. Cognitive services can be used to support any element of the student life cycle.

Scenario 1: A student may ask the cognitive assistant the following question: "I want to go to university and I need 350 UCAS points to get accepted by my preferred university. What grades do I need to achieve on my remaining assignments to achieve the required number of points to get to university?" The cognitive assistant may provide the answer even before the student has asked the question. For example, the cognitive assistant will know that the student needs 350 UCAS points to get to his or her preferred university and will advise the student about the grades that need to be achieved when the assignments are posted on the institution's learning management system by the teachers on the course. The nature of these services will ease the workload of teachers because they can refer the query to their cognitive assistant. It will also reduce the number of instances were incorrect advice is inadvertently given to the student. The cognitive assistant can also provide students with supplementary information and advice relating to the UCAS points enquiry; such as forthcoming university open days or workshops on applying for university.

Scenario 2: A student is uncertain about which courses to progress onto after completing his or her current programme of study. The cognitive assistant is asked to provide some insight by the student. The cognitive assistant has a history of all the students who have graduated from the student's current course and what courses they all progressed onto. The cognitive assistant also knows that the student wishes to pursue a career in engineering. When responding to the student's enquiry the cognitive assistant will draw on this library of information before suggesting possible options to the student. The cognitive assistant will also inform the student support team so that further guidance and support can be offered to the student. As we will see later cognitive computing will also reduce the traction or friction that currently exists when students apply or pay for their courses.

Scenario 3: The learning and assessment materials that are presented to each student on an institution's learning management system are tailoured, contextualised and personalised to meet the needs of each student. Adaptive learning environments take advantage of cognitive computing as they learn and adapt to better support students with their studies. A significant advantage of this service is that each student receives a unique learning and assessment experience. Two students in a classroom could be sat next to one another. At first they both see the same title page for an online tutorial but thereafter they experience a very different journey through the tutorial. This is because the behaviour of the tutorial is informed by the student's unique dataset, previous assessment results, their current performance on the course, the targets and goals associated with the student and even their learning preferences. One of the challenges that will face teachers will be around the degree of agency that is assigned to the cognitive service within the institution's adaptive learning environment. If complete agency is provided to the cognitive service it alone will determine the learning and assessment materials that are presented to students as they start an online tutorial. The decisions made by the service could be so complex that they will would be unfathomable to the course team. Click here for more info.

Scenario 4: As we saw in the previous scenario, the merits of providing personalised learning and assessment materials cannot be understated. Nevertheless, it can be taken further when students have access to a cognitive assistant to support them as they progress through an online tutorial. Students can ask their cognitive assistant questions relating to the learning and assessment materials that are presented to them in an online tutorial. The responses are contextualised and personalised to meet the needs of each student. In this instance the cognitive assistant behaves as a virtual classroom assistant. Its behaviour, as it supports a student, is informed by the student's dataset, profile and the behavioural algorithms that have been authored by the teachers and support teams on the campus. Its behaviour and the support it provides will vary according to the student who is asking the question. And more importantly, the cognitive assistant amplifies the teacher's ability to work at scale; enabling the teacher to provide individual support to hundreds or thousands of students on a course. The service also provides insights to the teacher as it works pro-actively to support the students on the course. Click here for more info.

Scenario 5: Students who are new to a school, college or university always have lots of questions about their courses, the services that are available to support them with their studies; and they want answers to everyday questions such as what is my timetable, who is teaching me, where can I get a bite to eat, what are the semester dates, where is the exams office and more. Cognitive assistants are well placed to answer an array of questions from students as they start and progress with their studies. Cognitive computing can also enable teachers and student support teams to be informed about particular enquiries that may require face-to-face support. One of the advantages of these conversational services is that it enables students to access immediate responses to questions and services; even during out-of-office hours. The particular service has already demonstrated its value when it went live at Bolton College in April 2017. Click here for more info.

Scenario 6: A student is revising for a Maths exam. The cognitive service informs the student about the topics were he or she has a good grasp of knowledge and it also tells the student about the topics were extra practice is required. The recommendations that are presented to the student update automatically as the student prepares for his or her Maths exam. Practice test results are presented to teachers and they use these insights to lay on additional support classes for particular topics to help students as they prepare for their forthcoming Maths exam. Students are notified about planned support classes by their personal cognitive assistant.

Scenario 7: Cognitive computing or cognitive assistants will support many of the day-to-day routine tasks or jobs that students need to do as they progress with their studies. For instance, the cognitive assistant will support students as they make routine appointments with members of staff, help manage assignment deadline dates and the assistant will help the student to manage and schedule payments to support their studies. The cognitive assistant can also provide information to students about the services that are available around the campus to support their applications for further study or when applying for university. The cognitive assistant is context aware which means that the service is aware of the student's place in the student life cycle and it knows what to offer the student at each point in the student life cycle. Since this is a cognitive service it will work with teachers and support teams to improve the outcomes of all students at each and every stage in the student life cycle. Institutions will need to ensure that the assets required to inform the cognitive service are fit for purpose. These assets will need to be tagged to match the various points in the student life cycle, they will be assigned a weighting (which could change over time) and they need to stored in a manner that allows for speedy information retrieval.

Other scenarios - Here is a list of other scenarios where cognitive computing could be used to help and support students as they progress with their studies:

  • searching and applying for courses - courses are presented to students that match their profile. Historical data is also used to inform the algorithms before they present relevant courses to each student. Conversational services which take advantage of natural language processing will support students during the application process. Students can simply ask their cognitive assistant for help and support about suitable course choices. As time progresses the cognitive service will improve its ability to present appropriate courses to students as they accept course recommendations or as they complete the application process for these recommended courses.
  • managing payments for courses and other financial transactions on the campus - this traditional service will take advantage of online digital assistants that will handle day-to-day enquiries regarding student finances. As institutions start to use CRM systems that incorporate cognitive services we will see additional value being created for support teams and the students that they support. Cognitive services will guide, advise and support students and support teams throughout the entire course application process.
  • accessing information, advice and guidance pertaining to a student's studies - students will receive tailoured information to suit their needs and requirements. The service learns to present relevant and timely content to each student. The success of this service will depend on the use of natural language generation.
  • greater personalisation of news services - campus news and other information that could support the student with their studies whilst at the school, college or university will take advantage of cognitive computing to help support the delivery of personalised and contextualised information to each student. These services do not use deterministic algorithms or programming to inform the information that is presented to students. Previously news articles would have been authored and posted to a database by individuals around the campus; and the articles would be manually assigned to various student groups. However, with the advent of cognitive services and their use of probabilistic reasoning means that the news articles that students see on their home page are closely related to their interests. The service has lots of parallels with the course advisory service which recommends courses to students.
  • personalised analytics with advisory notes given to the student - the use of natural language generation services will enable schools, colleges and universities to deliver advice and guidance to students as well as progress graphs, charts or tables. The use of natural language generation is encouraging because information, advice, alerts, notifications or reminders can be presented autonomously to students in plain text. Cognitive analytics enables institutions to deliver highly personalised services to each student and it does it at scale.
  • a virtual librarian which supports students with reading lists, loans and reservations - this is an interesting concept because the cognitive assistant will recommend timely resources that could help the student with an assignment or a revision topic. A discovery service would act on behalf of the student as it searches the library of content on or off the campus. The virtual librarian will even make library reservations on behalf of the student. This could be seen as contentious by individuals because the service does part of the search and discovery on behalf of the student. However, as the volume of information grows it would be difficult to undertake effective research without a virtual librarian. Students can also ask their cognitive assistant for help as it has access to the entire library catalogue and to all the resources on the institution's learning management system.
  • a reward service which gives vouchers or credits to students - cognitive services can be utilised by schools, colleges and universities to encourage and support positive behaviour amongst their students. Institutions could use cognitive services to promote better attendance or punctuality to lessons, the timely submission of coursework by students or improved grades. Colleges or universities can use cognitive computing to offer incentives to students in order to raise enrollment rates for courses or to raise the number of early applications to courses.
  • a virtual essay checker which takes advantage of natural language understanding to advise students about how they could improve their work before submitting it to their tutors. This service has the potential to reduce the percentage of assignments that need to be resubmitted by students and it has the potential to raise the average grade profile for courses. Turnitin has successfully established such a service which takes advantage of the company's natural language understanding service that supports students to be better writers.
  • a discovery service which advises the student about literature or other sources of information that could support them with their current project
  • an academic review service which advises the student if grades are falling below par or if there is a risk of not achieving the required grades that are needed to progress onto an apprenticeship or university
  • an advisory service which suggests to students about how they could improve their current project in order to achieve a higher grade
  • a booking system that helps students to schedule appointments with staff around the campus - the service will be engage with students as they progress with their studies. The service would take advantage of natural language processing to support students and staff as they make their appointments. Once again the service is context aware. This enables the service to make recommendations to the student about making an appointment to see academic support or to see the careers teams to start the university application process.
  • a timetabling service which is context aware - it advises students about room changes, canceled lessons, things to do before scheduled classes and it provides students with class notes associated with forthcoming lessons
  • a campus alert service - the alert system would also incorporate external services such as local weather or local public transport services - the service could advise students about changes to bus or train times or the approach of poor weather such as snow.
  • a general concierge service which would handle day-to-day queries or tasks - cognitive services can be trained to instantaneously compare a student's educational records with an institution's policies and guidelines and historical student records. The service is fed huge amounts of information and is trained by teachers and support teams with previous student cases histories and their expertise and experience is used to improve the service's recommendations and insights. Students will also ask their digital assistants numerous questions about the campus and their studies. Over time, insight from the cognitive service will provide valuable information for teachers and support teams that will help them improve student welfare as they progress with their studies.

Cognitive computing scenarios for teachers

Here is a list of scenarios were cognitive computing can be used by teachers to support their day-to-day tasks.

Scenario 1: A course team meets regularly each week to discuss the progress that their students are making on the course. Members of the course team talk to their cognitive assistant to gain insights into each student on their course. The cognitive assistant is not only an important source of information it can also recommend actions that will support individual students. The assistant can also undertake some of the administrative duties that were previously undertaken by the course team such as making appointments for students with the support team, informing students about key calendar events on the course, producing timely management reports or keeping parents and other stakeholders informed. The cognitive assistant acts and functions as a supportive member of the course team. In a recent development Bolton College's ILT Team was pleased to announce that Ada, the College's digital assistant for students, teachers and support teams, acquired the ability to convert speech-to-text and text-to-speech. This means that students, teachers and support teams are able to talk to Ada and ask for assistance across a broad range of domains that relate to the College or to their work. Likewise, Ada can also provide a verbal response to enquiries. Click here for more info.

Scenario 2: Cognitive computing will automate many of the key processes in a school, college and university. The production of the traditional termly or yearly student report for parents has always been a challenge for teachers and managers. For many years now parents have had access to student dashboards which provide graphical and tabular information to them; but teachers are still expected to either type comments or select from a set of predefined statements about each student's progress which is prone to errors or omissions. Their could also be delays in producing the reports. The advent of cognitive computing will automate the production of student reports at scale. Cognitive computing will enable institutions to present hourly or daily reports to students, teachers and parents. The advent of the cognitive assistant will mean that students, teachers and parents can ask for up-to-date information at any time in the day and at anytime in the year. Please note that these reports are produced by cognitive computing services so they have nuances that are missing from the traditional report card. For example, they give real time advice and guidance to the student and to all parties involved in supporting that student with his or her studies; and they do this accurately, regularly and at scale. Please note that the use of cognitive computing to produce real time student report cards cannot be realised if the school, college or university is struggling to manage the basic governance arrangements of its data. If cognitive computing was introduced to an institution were data governance was not mature the outputs produced by the service could be not be trusted and relied upon. Click here for more info.

Scenario 3: In a classroom setting, good teachers can differentiate, contextualise and personalise the way they structure and present topics to their students. Good teachers make this task look effortless and they do it time and time again. When the class gets larger, the task of differentiating gets increasingly difficult and eventually becomes impossible; especially when the class has a wide range set of students. However, the advent of cognitive computing and cognitive assistants simplifies the task and enables course teams to effectively differentiate, contextualise and personalise at scale. This means that institutions could deliver personalised teaching, learning and assessment to hundreds or thousands of students on wholly online or part online courses. When courses operate at scale they would not be able to operate effectively without the support of cognitive services. The same holds true for all other services that are provided by larger colleges and universities.

Scenario 4: Cognitive computing is ideally suited to help teachers provide informed feedback to each students. The most obvious area were feedback is provided is when students progress through an online tutorial or assessment activity. At the present moment in time student interaction and the responses submitted by the student to questions during an online tutorial deliver pre-programmed responses from the teacher; and the same feedback is provided to all students as they correctly or incorrectly respond to questions or activities. The advent of natural language generation enables teachers to deliver unique, personalised and contextualised feedback to each student within the online tutorial. Teachers and support teams can also deliver personalised feedback and messages to the student across multiple channels. The feedback can be supportive, it can be advisory or it can be affirmative to each student. Teachers can condition the cognitive service to behave in a particular manner as it supports different groups of students around the campus. Once the parameters are set, the cognitive service will automatically provide feedback and guidance to each student as they progress with their studies. As the service develops, Bolton College's cognitive assistant will provide feedback and guidance without being prompted by the student. For example, a student may ask Ada about the approaching half-term break. Ada will respond with the appropriate answer but in addition, the service will acknowledge a recent top grade that was achieved by the student and provide the student with an affirmative and encouraging message from the course team. Please note that no member of the course team authors this message - it is automatically generated by the institution's cognitive service on behalf of the course team. The service will learn about the messages that work or don't work for particular students and adapt its behaviour accordingly.

Other scenarios - Here is a list of other scenarios where cognitive computing could be used to help and support teachers:

  • I like the following service because it would help me with my terrible memory. As a teacher enters the classroom and opens the online services which enable him or her to commence the class there is an opportunity for the cognitive assistant to present the teacher with reminders, prompts or calls to action. For example, the service could prompt the teacher to remind a particular student about paying for a course outing or to prompt a set of students about their overdue assignments. This particular service could also be presented to students, support teams, administrators and parents across multiple platforms. Prescriptive, contextualised and pervasive analytics will be more common place in schools, colleges or universities where cognitive services are used.
  • The use of learning analytics within the education sector is well embedded in our schools, colleges and universities. It enables teachers and support teams to easily identify the status of all students within the campus through a simple RAG rating system. A typical RAG rating system will identify three groups of students. Firstly, a red rating identifies those students who are not performing well, were there is significant risk of them not completing their studies or risks have been identified which will interrupt study. Course withdrawl is highly probable. In more sophisticated models the red rating will also be associated with those students who will complete their studies but whose academic progress falls below anticipated levels. A red rating will also trigger immediate action by individuals who are in a position to support the student. Secondly, an amber rating identifies those students who have started to fall behind schedule, their performance on the course has deteriorated or risks have been identified which could interrupt their studies. Once again, action is required by the course team and the wider support team to support the student. Thirdly, a green rating identifies those students whose planned progress on a course matches expected or predicted progress; and were no risks have been idenitfied which could interrupt their studies. There is a common misconception that students who have been rated as green on a RAG rating system require no intervention from the course team and from the support teams across the campus. A RAG rating system which makes use congitive services will provide prescriptive analytics to course teams and support teams for all three groups of students. The use of learning analytics and predictive analytics enables teachers and support teams to get an early insight into each student which will enable improved intervention and support measures to be put into place. These insights are improved when institutions start to utilise historical data to RAG rate their students. The use of cognitive services and cognitive assistants will improve the support arrangements for each student and it will enable a greater proportion of students to successfully complete their studies. Institutions will need to show caution when they use predictive modelling because they could inadvertently be using biased algorithms. Educational institutions must be careful in their use of algorithms that take advantage of cognitive services. As algorithms become more prevalent, previously unforeseen risks could begin to emerge. For instance, a perfectly well-intentioned algorithm may inadvertently generate biased conclusions that discriminate against a particular group of students. Input bias could occur when the main student dataset is biased because it misses key information, is not representative or reflects historical biases. Please note that bias could also exist in institutions were computer modelling is not used.

Cognitive computing scenarios for support teams

Here is a list of scenarios were cognitive computing can be used by support teams as they deliver information, advice and guidance to students around the campus.

Scenario 1: Colleges and universities have teams that are dedicated to providing specific support services for their students; such as careers advise, student finance, academic support, exams support and so on. These teams take advantage of the institution's main dataset to retrieve and interrogate the information on any given student that they are supporting. With the advent of cognitive services much of the retrieval work and the decision making work will be done by the algorithms that have been created by the support teams which will allow them to better support the students in their care. So how would this technology manifest itself? Let's take a traditional example were a student calls into the student finance office to enquire about making an application for a student bursary. The student wants to know if they are eligible for a bursary. In this scenario the support team would typically ask the student to answer a set of questions to determine if they are eligible to apply for the bursary. If the student meets the eligibility criteria the application process can continue. The student may be asked to complete an online form or in the worse can scenario the student is presented with a paper form to complete. Once the details are taken from the student there may be a waiting period before the student learns about the outcome of his or her application. Now let's view how the student could access the student bursary if they could take advantage of cognitive services. The cognitive service could be given the task of maximising the total number of students who benefit from the institution's bursary scheme. If the institution has collected a large sample of data from each student as they enrolled for their courses the information could be used to automatically determine student eligibility to the bursary scheme for all students across the campus. Eligible students would receive notification via multiple channels including the student home page, via the institution's mobile app, social media or via SMS. If there is enough data held on individuals, students may not even have to apply for the bursary. The time taken to realise the institution's goal for its bursary scheme could be measured in seconds.

Scenario 2: Student retention is a major issue that is faced by many colleges and universities around the globe; and the technologies that make up cognitive computing are ideally suited to address the problem. Probabilistic programming will enable support teams to combine anecdotal reasoning with more reliable statistical approaches. This will improve the support team's ability to identify those students who are at risk of falling behind or leaving their studies. Probabilistic programs can be modularised; enabling support teams to improve the accuracy across particular problem domains. As these services develop and learn, their accuracy and reliability will improve as they assimilate a larger and larger dataset. Institution's need to be mindful of the ethical issues surrounding the use of cognitive services that help teams manage student retention. For instance, what statistical approaches are used to profile students, what traits are assigned to student profiles and how do institutions avoid bias in their statistical models? Colleges and universities will also need to consider if and how probabilistic programs will be used to select students to their courses.

Scenario 3: A careers advice service - in most schools, colleges or universities the careers guidance team is typically small in size and they are unable to provide a truly personalised service to every student on the campus. The advent of cognitive computing provides the careers advisors with the tools and the means to deliver a very personalised and contextualised service to every student. Cognitive computing amplifies the work of these support teams and it enables them to engage with a much larger student group. With regard to the careers advice service, the decisions made by the cognitive service will be supported by both deterministic and probabilistic programming. If it is used well the work of the cognitive service will support colleges and universities to improve their student retention and completion rates; and it will lead to improved progression pathways for their students. For example, if a student is performing well with his or her studies, the cognitive careers advise service could suggest new opportunities that open up if the student achieves a higher grade on the course. The cognitive service would also take advantage of the institution's campus news feed to suggest recommended events that support the changing goals and aspirations of the student. Bolton College is currently using a deterministic programming model which clearly defines the information, advice and guidance that is presented to students as they progress with their studies; and as they make preparations to graduate from the College. The model has proven to be very successful, with a doubling of students who apply to university. Colleagues at Bolton College are also utilising the Ask Ada service, the College's cognitive assistant for students, teachers and support teams. The service uses natural language to respond to day-to-day enquiries and questions from students about career choices and progression to further studies or into employment.

Scenario 4: Support teams on a campus will organise a myriad of events during the course of an academic year. A calendar service which took advantage of cognitive services would provide reminders and highlight opportunities to join events in and outside the campus. The reminders would include hand-in-dates for coursework, appointments, event reminders and more. The cognitive element of the calendar service would enable it to advise teachers to set different deadline dates for assignments if it discovered that different course team members were all setting deadline dates that were closely spaced together. The service would have access to every student's dataset which would include structured data that highlighted their interests, progression paths or their career goals. The service would also have access to unstructured data - the most common being all communication between students and their teachers or support staff. These datasets would enable the cognitive service to offer students events and services; and to timetable them into each student's calendar. Since the service uses cognitive computing it would avoid events that clashed with timetabled lessons. The calendar service would also advise staff about the best time to schedule events on the campus. This is possible because the service has access to every student's timetable and it has knows of every other event that is happening around the campus.

Other scenarios - Here is a list of other scenarios where cognitive computing could be used to help support teams:

  • An interesting area of development has been the use of cognitive services to drive certain behaviours or to realise certain goals and targets amongst the students on the campus. In a recent project at Bolton College colleagues used deterministic algorithms and student profiling to raise the number of students who submitted university applications. The project more than doubled the number of university applications. When colleagues at Bolton College widen the use of cognitive services we would expect similar results for other projects and services. For example, it could be used to raise average attendance levels, improve retention figures on courses and increase the number of students who apply and progress onto further study or employment. Over a period of time cognitive services will learn and adapt their behaviour to drive forward improvements across a number of key performance indicators for the institution.
  • Cognitive services are ideal when they help support teams to deliver timely and targeted information, advice and guidance to students at various points in an academic year. The use of a cognitive assistant that is context aware; the use of prescriptive and contextual analytics; and the use of natural language generation services will combine in a unique way to deliver a comprehensive set of services which are tailoured to each student on the campus.

Cognitive computing scenarios for administrators and managers

Here is a list of scenarios were cognitive computing can be used by administrators as they support students, teachers and support teams across the campus.

Scenario 1: Cognitive computing will also ease the burden on support administrators and managers when they produce their monthly, termly or yearly business reports for individual courses, departments and for the institution as a whole. At the present moment in time individuals may spend alot of their time in chasing up information for their reports, they may find it difficult to read or interpret the data once they have it in their possession or they may find it difficult to identify key insights in a large and complex dataset. The automation of reports and the ability to gain insights at anytime in the day may means that many of the reports that are produced by colleagues may not be necessary anymore; especially when administrators or managers can simply ask their cognitive assistant for information or insights. I am particularly interested in seeing how colleagues will start to use Bolton College's conversational service when they seek out information or insights to support their work. Will they continue to use traditional software applications to pull off reports or will they start to ask Ada for information?

Scenario 2: Even though administrators have access to large quantities of data they may find it difficult to take advantage of it especially when they are faced with a torrent of structured and unstructured data. It may be the case that schools, colleges and universities actually analyse very little of the data that they have collected. As cognitive systems are introduced to the education sector these systems will start to learn from the institution's structured and unstructured dataset, they will discover correlations in that data, make hypotheses for these correlations and they will suggest actions that lead to better outcomes for the school, college or university. Administrator's will be advised by their cognitive assistant to either commence, expand or reduce provision for courses and course teams will be advised about accepting or rejecting a student's application to a course.

Scenario 3: When school, college or university leadership teams gather for their weekly meetings they will also invite their cognitive assistant to these meetings. Cognitive systems will enable leadership teams to run hypotheses, make informed decisions and make better decisions. In these meetings individuals will speak to their cognitive assistant and ask for information, run hypotheses, make projections or seek advice on any aspect of their school, college or university. In fact the organisation's cognitive assistant will be present at all meetings across the campus; enabling the cognitive assistant to foster greater insight and collaboration across the organisation.

When writing about the use of cognitive assistants in schools, colleges and universities I have focused my attention on how they complement, rather than replace, human capabilities. The design and development of Ada, Bolton College's cognitive assistant, is firmly based on how the service can enhance the capabilities of students, teachers, support teams and administrators around the campus. One of the premises that was used in writing this article acknowledges that educational services can be improved if teachers, support teams and administrators use cognitive services to enable them to better support their students.