Common questions. Answered.

Dilog is a customer simulation panel built directly from 1:1 interviews from customers (in Australia and New Zealand currently). It enables individuals and teams to ask questions of their customer panel quickly, accurately, and much more economically than previously possible. Feedback earlier and at scale leads to better results with less wasted time.

An ARP (Anonymised Reflective Profile) is a digital version of one real customer, built from a deep 1:1 interview. It keeps the attitudes, behaviours, rationalisations and cognitive biases that drive that person's decisions. Like a focus group, we recruit and interview your users, then turn each interview into one ARP in your Dilog platform.

It's a bottom up approach with ARPs vs. a top down approach with synthetic audiences. Synthetic audiences are large-scale simulations of real people, built from companies' customer data. ARPs are built from deep background interviews with human customers, where each interview corresponds to one agent (the ARP). ARPs are aggregated up into a panel for your organisation, for questions answered in seconds.

We work with organisational clients to define target customers and conduct interviews of those groups. All interviews create ARPs that mimic customer responses with high accuracy, and ARPs are assembled into panels. Each user can then filter and create any customer segmentation they wish when running a scenario/query against the panel.

On average 86%, measured by putting the same questions to each ARP and to the real customer it's based on. This testing is continuous for ongoing quality in responses, and greater context collection as well as our focus on operational improvement increases that average.

Usually under a minute from submission to the report being presented. It does depend on the number of ARPs that you ask however, so it can take up to a few minutes for larger panels used.

Any type of question that you would ask a customer that may or may not be that familiar with your business. UX, CX, Product, Marketing, CRO, and questions on B2C preferences are all great examples. Asking the panel about a process that is internal to the business, or with jargon that a normal customer wouldn't understand, and the accuracy of the response will decrease.

Customer data from your CDP or databases are optional for the functioning of a Dilog customer simulation panel. They are nice to have as they can enhance the specific responses for questions pertaining to your particular field. However, since we base the foundational ARP models off deep interviews with the individual customers, that context and the context gained with continual interviews is the basis for a functioning panel.

ARPs will respond as a consumer would, as we interview the consumers to create them. As part of that interview process, we will work with you to gather detailed information of each person's attitudes and behaviours in your vertical. For instance, questions about style, aesthetic preference, colour choice, lifestyle, etc for a client that is a home goods retailer. In addition, we ask further questions that would be business specific, around responses to your brand and your competitors. All of this becomes context to a tailored ARP for your business.

Dilog is grounded in real customer interviews; ChatGPT isn't. LLMs like ChatGPT, Claude and other AI tools generate plausible-sounding answers based on general training data. Because that training data is what is available on the internet and context about your customer that you could put in, the results produced are an average of what those LLMs think a person in that cohort should sound like. And an average opinion turns out to be an opinion of nobody in particular. Those results tend to be bland, missing the human-ness that comes from being grounded in real interviews, updated quarterly.

We continually go back to each person interviewed for an ARP and ask follow up questions. In addition, there is automated input around geopolitical, brand and market shifts. That updated context, as well as new questions on specific areas, keeps each ARP current and increases the amount of what each one "knows".

Individual tiers cost less, and are meant for individuals and small teams. If you are on this tier you are only able to access more generalised ARPs that do not have vertical or business specific context. So there is a tradeoff in accuracy of some more specific questions. In contrast, Organisation tiers have a separate setup, where Dilog will work with you to create a bespoke panel tailored to your needs and target customers. We interview for you, certify the panel accuracy, and continually update the panel for greater depth of context in your particular field.

Each person interviewed signs a consent form and agreement. Upon completion of the interview, any personally identifiable information is immediately removed or fully anonymised before an ARP is created. All personal data is kept in a separate location and never used for operational purposes. Anyone that has been interviewed can request that we permanently delete any data associated with them at any time, which deletes the ARP.

Since we base all panels first on the customer, customer data from any organisation isn't needed for full functionality. If desired, we can work with you to layer further customer information onto the panel for better accuracy, but that isn't a necessity for accurate responses.

Yes. Your whole panel is private, and since there is no actual human interaction with the questions that are being asked, there is no risk of those lines of enquiry being out in the general population.

For questions that are asked, each account is a private panel for that account, where input into that panel is held securely at an enterprise level. We follow GDPR principles and are working toward full compliance, and are currently in the process of being ISO 27001 and SOC2 certified.

Organisations will have full control of the duration input data is held. During setup we will set an agreed data retention rate for query input, and an admin user can adjust this at any time. Information shall be classified according to the information security needs of the organisation based on confidentiality, integrity, availability and relevant interested party requirements.

Any B2C type of question or scenario that you would usually put to a user research panel. Questions on UX/CX design, new product features, marketing, CRO, promotions, product launches, and competitor comparisons can all be presented.

It depends somewhat on what you are using the panel for. Quick directional feedback on UX design or marketing campaign optimisation questions can be as little as 12-15 ARPs per cohort. More in depth questions are best done with 100-120 ARPs.

Depending on the size of the overall panel desired, and subsequently the number of interviews needed, a bespoke panel setup takes 6-12 weeks.

With our free 14-day trial you can sign up and be asking your first questions within a few minutes.

When you are logged into your panel, you can create a cohort within the User section or in the Cohort section. First filter to the users you would like, select them and create a cohort to use when creating a scenario. Those can be reused for any future queries, and can be edited within the Cohort section.

Of course. If the background context is present to allow us to create a filter, we can do so. Please send requests to hello@dilog.ai.

You can currently upload images or a series of images into a scenario for a panel to look through. If a file type that is needed is currently not present, please let us know what you need (hello@dilog.ai) and we can add it in the submission page.

Pricing is competitive for both individuals that want to use the platform with more general customer panels, and for bespoke panels created to an organisation's customer research requirements.

Billing is per month for whatever tier you are on. If you go over your allotted credits for the month there is an option to purchase more at the same overall rate for that tier.

If you believe some credits were used on a question where there was a problem, please reach out to us directly at hello@dilog.ai to resolve the issue.