The conversational maxims (also known as “ Gricean maxims") are a framework for explaining how conversation proceeds rationally and efficiently. For conversational speech applications, they provide a reasonable though incomplete set of communication guidelines. Following them will facilitate straightforward communication; violating them will cause difficulties - “Failure to adhere to the maxims…. opens the door to miscommunication with and mistrust of the application” (Lewis, 2011, p. 34).

The conversational maxims are:

• Quantity: Be as informative as is required, but not more informative.
• Quality: Do not make statements that you believe are false or for which you lack evidence.
• Relation: Be relevant.
• Manner: Avoid obscure expression, avoid ambiguity, be brief, and be orderly.

In addition to their obvious application to VUI design, the maxims have been used as the basis for standardized questionnaires to assess of the quality of conversational systems and IVRs (R. Bloom et al., 1999; Polkosky, 2002, 2005a, 2005b, 2008).

Note that an underlying theme that runs across the maxims is one of conversational efficiency – provide the necessary information but no more; don't lie or provide potentially misleading information; be relevant, clear, brief, and orderly.

Ensure that necessary information is given
This follows from the first half of the maxim of quantity. To achieve this goal, it's necessary to engage in careful task analysis during the early stages of design.

Provide only necessary information
This is the second half of the maxim of quantity. Like the first half, it requires careful task analysis. It also requires the designer to keep the caller's goals in mind. Do they really want or need to hear this information? Or is it the business that wants the information presented? (Think cross-sell as an example.) Depending on the situation, the answer could be either.

This principle is also associated with the maxim of Manner, specifically, “be brief.”

Only ask for what you’re going to use
This is consistent with the maxim of Quantity and the general theme of efficiency. There's simply no point in taking the time to ask for information unless it's going to be useful.

Applied Example
Extra verbiage, such as “Please listen carefully, as our options may have changed.” (Who ever thought, “Oh, good! Thanks for letting me know”?)

Ensure cross-channel consistency
Due to the emerging use of different channels to get information (e.g., IVR, Web, multimodal), it's important to ensure from the early stages of design that the cross-channel architecture will support access to the same data regardless of channel. Otherwise, you run the risk of violating the quality maxim.

In situations where that's just not possible due to whatever business constraints exist, it's not a good idea to dumb down one channel to match the other, but be aware the customers will note the differences, or rather, the shortcomings of the one channel, which could affect their acceptance of the automation and even the company. It is definitely not safe to assume they won't notice. Consider what noticing might do to their frame of mind or attitude and use that to help guide decisions.

Prompt for what’s on the caller’s agenda, not the business’s
It's important to address common caller goals, regardless of what the business wants to do with them. Say a certain common task must be done on the website. It should still be included in the IVR menus, with an explanation of what the caller needs to do. Omitting it just means the caller will take a roundabout way to an agent who probably can't help them.

Organize information clearly
The information architecture (IA) of an IVR application is important, especially when requiring the navigation of hierarchical menus. In these types of applications, typical callers have limited options with regard to skipping around, so it's critical to provide options in an optimal order.

For menu options, place specific before general, then frequent before infrequent
When determining the order of menu options, first place specific options before more general ones to ensure that callers don't mistakenly select the general option when the specific one is a better match to their needs (“specific before general”). Then, to as great an extent as possible, place frequently chosen options before more rarely chosen ones (“frequent before infrequent”). For more detail, see Chapter 5.

Organize the IVR's information architecture from a user rather than an enterprise point-of-view
Organize your IVR according to the caller’s anticipated target tasks and not according to the company’s structure. Also, avoid jargon that the caller is not likely to know, e.g., when choosing the words to use as options in menus.

Don't ask users for the same information twice
This is consistent with the maxims of both Quantity and Manner. It is well known that callers find it very irritating to answer a question in the IVR, only to have to provide that same information again following transfer to an agent. The conversational maxims and their underlying focus on conversational efficiency help to explain this irritation. For this specific example, the solution is to ensure that the system includes computer-telephony integration (CTI – also see Partial Automation vs. Full Automation), and that agents are trained to use it.

There are other maxims or guiding principles not directly related to conversational maxims.

Format information for multiple audiences
For the sake of efficiency, if you expect multiple audiences to consume the information provided by your IVR, you need ensure early in the design process that the information format(s) will support the expected range of use and users.

Minimize data dip latency
This is a general guideline not necessarily implied by the conversational maxims, other than the underlying theme of efficiency. A common industry goal is for 95% of system responses to occur within two seconds of the end of a caller’s input (Balentine & Morgan, 2001; Fried & Edmondson, 2006). Toward reaching this goal, it is important to minimize the data dip latency – the amount of time it takes to get data from a database.

When it will take longer than 2 seconds for an application to respond to a caller's input, it is necessary to manage the caller experience during the wait (Lewis, 2011). Often, simply saying something like, “Just a moment while I look that up” serves this purpose. Several researchers (e.g., Boyce, 2008; Fröhlich, 2005) have published data that support guidelines for the management of processing time.

Use the following table to guide the management of processing time in conversational systems.

Balentine, B., & Morgan, D. P. (2001). How to build a speech recognition application: A style guide for telephony dialogues, 2nd edition. San Ramon, CA: EIG Press.
Bloom, R., Pick, L., Borod, J., Rorie, K., Andelman, F., Obler, L., Sliwinski, M., Campbell, A., Tweedy, J., & Welkowitz, J. (1999). Psychometric aspects of verbal pragmatic ratings. Brain and Language, 68, 553–565.
Boyce, S. J. (2008). User interface design for natural language systems: From research to reality. In D. Gardner-Bonneau & H. E. Blanchard (Eds.), Human factors and voice interactive systems (2nd ed.) (pp. 43–80). New York, NY: Springer.
Fried, J., & Edmondson, R. (2006). How customer perceived latency measures success in voice self-service. Business Communications Review, 36(3), 26–32.
Fröhlich, P. (2005). Dealing with system response times in interactive speech applications. In Proceedings of CHI 2005 (pp. 1379–1382). Portland, OR: ACM.
Lewis, J. R. (2011). Practical speech user interface design. Boca Raton, FL: CRC Press, Taylor & Francis Group.
Polkosky, M. D. (2002). Initial psychometric evaluation of the Pragmatic Rating Scale for Dialogues (Tech. Report 29.3634). Boca Raton, FL: IBM.
Pearl, C. (2017). Designing Voice User Interfaces: Principles of Conversational Experiences. Boston, MA: O'Reilly Media, Inc.
Polkosky, M. D. (2005a). Toward a social-cognitive psychology of speech technology: Affective responses to speech-based e-service. Unpublished doctoral dissertation, University of South Florida.
Polkosky, M. D. (2005b). What is speech usability, anyway? Speech Technology, 10(9), 22–25.
Polkosky, M. D. (2008). Machines as mediators: The challenge of technology for interpersonal communication theory and research. In E. Konjin (Ed.), Mediated interpersonal communication (pp. 34–57). New York, NY: Routledge.