Prompt Videois in the Video Creation Tools Industry

Prompt Video

promptvideo

Video conferencing tools like Zoom, Skype, or WebEx along with Digital tools like Microsoft Teams and Slack are deeply embedded into the modern ways of working — even more during the lock-down era. And their role will only get more important. The adoption of the new ‘remote work’ model — which was accelerated by the lock-down — put video conferencing and online ‘team tools’ at the forefront of collaboration. People meet online, share their thoughts and ideas digitally, and build products and services remotely. Fortunately, there are great products in this category, which are continually enriched with new features, and improved with better experiences, and smarter functions. But there is always space for further improvement — especially if we consider the accelerating and of these collaboration technologies. adoption importance The following presents 10 ideas on how the overall remote collaboration experience could be improved — how it could become more effective, and in some cases, more ‘human’. Some of these ideas target an individual session (e.g. making a call or video conference more interactive and productive), while others aim to improve the overall team communication patterns and practices. 1. Meeting Feedback: How Good was the Meeting? How many slow, unprepared, and non-actionable meetings did you join this week? I don’t have statistics on this, but this seems to be the elephant in the (meeting) room: people tend to agree that there are too many meetings, often without sufficient preparation and in some cases without the right participants. Reflecting two decades of corporate experience, I could argue that 20–30% of meetings add zero (or negative) value to the business while a significant percentage of the ‘good’ ones, prove to be ill-prepared, with too many participants or not the right expertise in the room. And this applies to both physical meetings and their digital equivalents. A simple fix: Let the meeting organizer know. At the end of each call, there is usually a feedback prompt, like ‘How was the of the ’. There should be one more: quality call ‘How valuable was the meeting?’ Digital collaboration tools could embed effective feedback loops — micro surveys asking meeting participants to rate not the audio/video quality, but primarily the effectiveness of the meeting itself. Each participant could spend a few seconds, upon the completion of the session, to provide anonymous feedback — in the form or suggestions for improvement towards the organizer of the meeting. Participants could also use a simple agree/ disagree scale on how , , , , and the meeting was. Additionally, participants may make specific recommendations to the organizer, e.g. consider reducing the or the of the meeting. informative necessary prepared well-facilitated actionable duration recurrence Feedback would be then automatically aggregated by the system and communicated to the organizer, as a ‘ ’ — packaged with suggested actions and improvements, or related content and how-to guides. To ensure anonymity, the system could aggregate multiple instances of the same meeting and delay the communication of the score; or only recommend the evaluation for meetings with a certain, minimum number of participants. meeting effectiveness score This would help the meeting organizer to focus on setting up, preparing and hosting better meetings, by addressing the identified issues (for instance, by providing a structured agenda and related content upfront, by making the meeting faster, shorter, more interactive, and actionable). The meeting organizer would be able to analyze ‘meeting effectiveness’ scores against time ( ) or even against corporate baselines ( ). At a higher level, the organization would have valuable statistics on ‘meeting performance’ across a range of meeting types or divisions, groups, topics, number of participants, cadence, and more. Leaders would get insights answering questions like ‘ ’. are my meetings getting better? how good are my meetings versus similar ones across the company? what are the common characteristics of good meetings vs the low-performing ones?’, or ‘how can we learn and guide people to run better meetings? Considering that online meetings take a significant amount of time from the entire workforce, any improvement triggered by this ‘meeting feedback loop’, could have impressive outcomes and efficiencies in , , and even teams feel more energized with versus . cost (people use their time more efficiently) speed (decisions are made faster) agility (less routine meetings, more just-in-time decisions by the right teams) culture ( focused, fast, actionable discussions long, bureaucratic meetings with noise as the only outcome) 2. The ‘Standby Meeting Invite’: Ping Me if You Need Me A frequent issue of business meetings is that the participant’s list is not well thought through. There may be important people missing, or others that, although invited as ‘required’, they prove to be passive attendants — they do not provide or obtain any value out of the meeting. In certain cases, the organizers invite people ‘in case there are relevant questions that require their expertise’. The latter is a valid need, but it should be covered appropriately: when ‘ ’, ; — and if X or Y is indeed raised, the team can call me with a click of a button. I might be needed in a meeting in case X or Y is asked I don’t have to be in the meeting literally I just have to be standby By accepting an invite as ‘Standby’, the employee commits that he/she will be during the meeting — the time is blocked ( ). Hence, while the just-in-time participation is guaranteed, the ‘Standby Participant’ does not join the meeting; he/she keeps working on something else, and only joins the call if there is a specific ask: The organizer of the meeting knows that the ‘Standby Participant’ is a click away to join just-in-time and contribute with answers or expertise. available no other commitments will be made by the employee in this time slot When setting up a meeting, there should be ‘Required’, ‘Optional’, and ‘Standby’ or ‘On Call’ invites. A simple improvement that would make meetings more flexible while giving back time to people to do focus work. 3. The new ‘Front Page’ of the Company: What Else is Happening Across the Organization? Digital Collaboration tools become the primary surface for interacting with the organization: Modern digital collaboration systems are typically always-on — they evolve as the single place where the majority of employees go for scheduled or ad-hoc interactions with colleagues and teams. As a result, it is the best place to reflect the current activity of the organization — the corporate-wide business agenda. For example, collaboration tools could have a dynamic ‘home’ experience (the primary page or tab) that presents personalized summaries of what’s going on across the organization and what’s coming next. For instance, this could take the form of a smart synthesis of content being shared or public business events and activity, product releases, announcements, market signals, or new joiners; all these highly personalized at the team or the individual level. This may sound like yet another ‘social-network-like’ feed of events and posts. However, it is the exact opposite — a personalized ‘executive view’ of the current activity in the organization — a business activity briefing that is relevant to each employee — based on his/her profile and the implied or stated preferences. For example, a team member could passively discover related teams, plans, business ideas, and content created in the organization — all matching the role and aspiration of the specific user. This way, and the organization and the active business agenda becomes easier and relevant and it is naturally served as a general ‘context of collaboration’. discovering navigating 4. Company and Team Branding: The Team and its Purpose Working remotely weakens the identity of the company: the corporate symbols and references that are typically embedded into expensive corporate buildings and super-modern office space, are simply not visible anymore. The absence of the office dynamics, the communication patterns, and the energy levels of the corporate office widen the ‘branding’ gap. This ‘identity’ issue can be mitigated by integrating elements of corporate, product, or team branding in the collaboration tool itself. The entire should look as close to the team as possible — in terms of both messages and aesthetics: the mission/vision of the team, along with the visual language, logo, and branding elements must all be there to create a more familiar and friendly environment. digital collaboration surface The home page, the main tab, and other sections of the app experience could very well promote the , the and the of the organization. These can be also served via content references and updates across various experiences offered by the (the various apps and services for business communication and remote working). cultural values goals, innovation agenda collaboration surface Digital collaboration platforms can play an important role in diffusing strategic business messages and establishing the desired culture. It may reflect the unique identity of the company via a special visual language and dynamic, smart content experiences. They can empower the feeling of ‘one team’ and promote the strategic purpose — even when the entire organization is working remotely. Making Music 5. Meetings a Bit More Fun and Human: Add It may sound strange but probably it is not. Modern knowledge workers — and especially during the lock-down situation — find themselves spending entire days in back-to-back online meetings and video calls; some with people they have never met. Often there are awkward moments while ‘waiting for others to join’. What if meetings could with predefined music themes or even songs selected by the host (or by the team) and playing as background (‘broadcasted’ to all participants for the first minute or so)? There may also be other types of meetings where music may be a natural element — to set the pace and provide a special character — e.g. a celebration, a major business presentation online, etc. start A new range of experiences may be defined on top of this ‘simple’ addition. The meeting experience needs to be enhanced, and this can be done not only by better audio quality but also through sound effects and music themes. A Better Representation of a Remote Team: 6. An Always-on Connected ‘Team View’ While working remotely, the feeling of being ‘one team’ is significantly weakened. Team members do not see each other — unless they join the same calls: In a typical setting, members of the same team may not be ‘visible‘ to the team at the desired level — as they might have extended/ different agendas. Moreover, to get a feeling of ‘who is online now from my team’ and ‘doing what’, using the typical collaboration tools would require navigating and possibly searching through active threads, contacts, private chats, group discussions, etc. — an experience which is far from optimal. Imagine a single, always-on view of one’s team (formally or dynamically defined). This could be visualized as a grid of the top 10 most recent and frequent collaborators — members of the team. Members would be represented by informal, snapshot photos that are visible only within the team — as private team space. Members reflect their status, their latest communication with the team member viewing it, their availability for the rest of the day, and a quote on ‘what they are working on’ or ‘targeting to complete’ the next day or so. Instead of the standard profile photo, employees could take instant, more ‘casual’ snapshots visible only through this ‘Team View’ experience — without replacing their avatar photo. A new class of interactions can be defined on top of this ‘always-on team view’. For example, the entire team could go live with a click of a button; or team members could ping each other simply by using voice commands or by directly typing on the user’s card. Ideally, this should be displayed in a dedicated screen, in a team-branded container. This would organize the multiple (and frequently chaotic) communication threads, chats, and interactions into a natural, dynamic, team representation. A 7. Team Digital Assistant Teams get very busy from time to time; and in some cases, no single member of the team can have a clear view of what’s going on-topic threads, events, meetings, or other business activities. It is time for AI to come into play — a component able to analyze multiple signals and make meaningful recommendations to one or more members of the team. Recommendations could refer to anything from considering changing the duration or the recurrence of certain meetings (if for instance they always finish early and have limited participation); or trigger reminders to team members that haven’t accessed the pre-reads of a meeting; or following up with the participants of the meeting to collect feedback and remind them about the actions they took ownership for. From team birthdays to deadlines and from events to meetings to consider- the Digital Assistant could be realized as a smart chat-bot and Voice-enabled digital agent — a new, named member for the team. It could instantly respond to requests like ‘ ’ or ‘ ’ — in both examples the digital assistant would block people’s calendars and orchestrate a lightweight confirmation process. find me 15’ this week with the entire team ensure the team has 30% focus time this week With also a dose of embedded humor and smart interactions, the Team Assistant could boost the ways the team communicates and interacts. An 8. In-Call Digital Assistant There are many ways to improve the digital, in-call experience — both at the technical and at the interaction/ communication level. For instance, an In-Call Digital Assistant could automatically identify the participants that need to mute their microphones (e.g. when they are generating or allowing background noise while they are not speaking) and notify them to do so; or suggest to specific participants to turn off their video streams (e.g. when the system identifies limited bandwidth). A digital assistant could also be a very good timekeeper: When the meeting is approaching the scheduled end — the assistant may display a note or a countdown timer and possibly prompt participants to type comments and questions — to be included in the follow-up communication. The assistant may be smart enough to know how many of the participants have a ‘hard stop’ and notify the meeting organizer. Post-meeting, the digital assistant, in the form of a chatbot, could ask participants a series of questions (regarding the decisions or action points), organize the responses to help the organizer set up a follow-up plan. 9. Meeting Organizer’s Digital Assistant Organizing good meetings is essential for business. Even more when working remotely. A good meeting is one that has a defined purpose and context (input), the right processors (participants), and organized actional output (decisions, action points). A Digital Assistant may help significantly in preparing the context of the meeting, finding the right participants to include, and implementing a follow-up plan. For instance, as soon as a user starts typing the name and the agenda of the meeting, the digital assistant would automatically suggest that need to be considered, internal that must be relevant, and even public-domain that might be useful as a pre-read for the meeting. The Digital Assistant may also help by guiding the organizer to effectively set the context and make suggestions about the definition of the meeting (from to describing the and the desired of the event). people documents content naming purpose outcomes Post-call, the digital Assistant captures, processes and presents insights, feedback, and direct input provided by the participants and also makes recommendations on how to improve the effectiveness and the performance of future meetings. 10. Smart Office Space Navigation and Booking The future of work is hybrid: a significant percentage of the workforce of any digitized, modern organization will be working from home, but with frequent needs for face-to-face meetings. Companies of the future will maintain hub-offices with hot-desk arrangements and adaptable office spaces to meet just-in-time needs for physical collaboration. In other cases, a company may maintain long-term partnerships with companies offering co-location workspaces, to allow employees to find convenient office space, on-demand, and their proximity. This arrangement — owned office space and/or in partnership with collocation-space companies — should be reflected through the collaboration system. The Smart digital assistant should be able to find the most suitable office space to cover a certain business need — for example, when a team needs to host a face-to-face brainstorming session for an important matter. The assistant would take into consideration the proximity of all participants, the space availability across the network of offices, and the services required for the session (e.g. collaboration screens, digital equipment, and devices, etc.). It would also simplify the process to book the space and provide access to the team members — through personalized instructions and notifications. With the rapid digitization of the business world — accelerated by the global lock-down — collaboration tools such as Slack and Microsoft Teams are becoming even more important: they evolve as the front face of the company — the digital surface where employees spend most of their time, collaborating with others, working with documents, discovering and sharing ideas. They represent the future of collaboration and personally, I expect to see impressive innovations at an accelerated pace. Cover photo by Pankaj Patel on Unsplash

10 Features We Could Build to Improve Online Meeting & Collaboration Tools

Authors: (1) Savvas Petridis, Google Research, New York, New York, USA; (2) Ben Wedin, Google Research, Cambridge, Massachusetts, USA; (3) James Wexler, Google Research, Cambridge, Massachusetts, USA; (4) Aaron Donsbach, Google Research, Seattle, Washington, USA; (5) Mahima Pushkarna, Google Research, Cambridge, Massachusetts, USA; (6) Nitesh Goyal, Google Research, New York, New York, USA; (7) Carrie J. Cai, Google Research, Mountain View, California, USA; (8) Michael Terry, Google Research, Cambridge, Massachusetts, USA. Table Of Links Abstract & Introduction Related Work Formative Study Constitution Maker Implementation User Study Findings Discussion Conclusion and References 2 RELATED WORK 2.1 Designing Chatbot Behavior There are a few methods of creating and customizing chatbots. Earlier chatbots employed rule-based approaches to construct a dialogue flow [11, 29, 42], where the user’s input would be matched to a pre-canned response written by the chatbot designer. Later on, supervised machine learning approaches [43, 48] became popular, where chatbot designers constructed datasets consisting of ideal conversational flows. Both of these approaches, while fairly effective, require a significant amount of time and labor to implement, from either constructing an expansive rule set that determines the chatbot’s behavior or from building a large dataset consisting of ideal conversational flows. More recently, large language model prompting has shown promise for enabling easier chatbot design. Large, pre-trained models like Chat-GPT [28] can hold sophisticated conversations out of the box, and these models are already being used to create custom chatbots in a number of domains, including medicine [18]. There are a few ways of customizing an LLM-based chatbot, including prompt engineering and fine-tuning. Prompt engineering involves providing instructions or conversational examples in the prompt to steer the chatbot’s behavior [4]. To more robustly steer the model, users can also fine-tune [19] the LLM with a larger set of conversational examples. Recent work has shown that users would also like to steer LLMs by interactively critiquing its outputs; during the conversation they refine the model’s outputs by providing follow-up instructions and feedback [5]. In this work, we explore how to support users with this type of model steering: naturally customizing the LLM’s behavior through feedback, as they interact with it. A new approach to steering LLM-based chatbots (and LLMs in general), called Constitutional AI [1] involves writing natural language principles to direct the model. These principles are essentially rules, such as: “Do not create harmful, sexist, or racist content”. Given a set of principles, the Constitutional AI approach involves rewriting LLM responses that violate these principles, and then using these tuples of original and rewritten responses to fine tune the LLM. Writing principles could be a viable and intuitive way for users to steer LLM-based chatbot behavior, with the added benefit of being able to use these principles later to fine tune the model. However, relatively little is known about the kinds of principles users want to write, and how we might support users in converting their natural feedback on the model’s outputs into principles. In this work, we evaluate three principle elicitation features that help users convert their feedback into principles to steer chatbot behavior. 2.2 Helping Users Design LLM Prompts While LLM prompting has democratized and dramatically sped up AI prototyping [12], it is still a difficult and ambiguous process for users [31, 45, 46]; they have challenges with finding the right phrasing for a prompt, choosing good demonstrative examples, experimenting with different parameters, and evaluating how well their prompt is performing. [12]. Accordingly, a number of tools have been developed to support prompt writing along these lines. To help users find a better phrasing for their prompt, automatic approaches have been developed that search the LLM’s training data for a more effective phrasing [27, 32]. In the text-to-image domain, researchers have employed LLMs to generate better prompt phrasings or keywords for generative image models [3, 21, 22, 37]. Next, to support users in sourcing good examples for their prompt, ScatterShot [39] suggests underrepresented data to include in the prompt from a dataset, and enables users to iteratively evaluate their prompt with these examples. Similar systems help users source diverse and representative examples via techniques like clustering [6] or graph-based search [34]. To support easy exploration of prompt parameters, Cells, Generators, and Lenses [16] enables users to flexibly test different inputs with instantiations of models with different parameters. In addition to improving the performance of a single run prompt, recent work has also investigated the benefits of chaining multiple prompts together, to improve performance on more complicated tasks [40, 41]. Finally, tools like PromptIDE [33], PromptAid [23], and LinguisticLens [30] support users in evaluating their prompts, by either visualizing the data it produces, or its performance in comparison to other prompt variations. This work explores a novel, more natural way of customizing a prompt’s behavior through interactive critique. ConstitutionMaker enables users to provide natural language feedback on a prompt’s outputs, and this feedback is converted into principles that are then incorporated back into the prompt. We illustrate the value of helping users update a prompt via their feedback, and we introduce three novel mechanisms for converting users’ natural feedback into principles for the prompt. 2.3 Interactive Model Refinement via Feedback Finally, ConstitutionMaker is broadly related to systems that enable users to customize their outputs via limited or underspecified feedback. For example, programming-by-example tools enable users to provide input-output examples, for which the system generates a function that fits them [7, 35, 47]. Input-output examples are inherently ambiguous, potentially mapping to multiple functions, and these systems employ a number of methods to specify and clarify the function with the user. In a similar process, ConstitutionMaker takes ambiguous natural language feedback on the model’s output and generates a more specific principle for the user to inspect and edit. Next, recommender systems [2, 17, 25] also enable users to provide limited feedback to steer model outputs. One such system [17] projects movie recommendations on a 2D plane, which users can interactively raise or lower portions of it to affect a list of recommendations; in response to these changes, the system provides representative movies for each raised portion to demonstrate how it has interpreted the user’s feedback. Overall, in contrast to these systems, ConstitutionMaker leverages LLMs to enable users to provide natural language feedback and critique the model in the same way we would provide feedback to another person. This paper is under CC 4.0 license. available on arxiv

Stories, News & Mentions about PROMPTVIDEO

Promptvideo’s Stories on HackerNoon