Future of Software

Site Outline

Working Notes

Future of Software Project

Design Notes

Questions about Capabilities

Phase II

Exploration / Bugs / Ideas

Testing

Combined Project

Explore

Phase II

User Experience/Design

While it looks good, there are a number of issues that scream that this isn’t an enterprise-grade platform. We need to make a number of edits to the platform to offer users with a more consistent user experience.

[Cellstrat] The “thinking” process should be default closed. If you don’t know what I mean by that, ask questions to clarify. While it was nice to have it open, the consistent issue of opening and closing and re-trying make it look like the platform is broken most of the time. Closing with make is seem like it’s taking longer, but at least it won’t look broken.

[Cellstrat] The thinking process needs to have a clear visual clue to communicate that something is going on in the back end. I’ve provided examples of the text shimmer. If you don’t think you have the required information to proceed, ask. See OpenAI’s shimmer process for o3-mini

[Cellstrat] The initial comments I received about the platform included a concern about the ‘loading’. To the user, this seems that every new question includes a load step. To the user it seem repetitive since this might have been the 5th task. “Why does anything need to load again? Didn’t it load last time”. Please suggest language that is more in line with the backend process that doesn’t make it seem like the app is broken.

[Cellstrat] Do not continue to share with the users when there is an issue and you need to take another approach. Again, this looks like this is an error to the user. Every other platform keeps these reworks more private. Please think like a user. Retrys should not have their own “thinking” section.

[Cellstrat] The initial welcome messages are not left aligned with the other text and operate at a much slower pace. This is supposed to be exactly like the agent response so the typewriter speed should approximate the response speed and the text should be aligned.

[Cellstrat] The right panel need to follow the requested design.

[Cellstrat] The right panel should include the data file(s) used for the data set.

[Cellstrat] Please provide a rough style-guide outline so that we know what everything should be ahead of time and if we create new section, we have a common style guide for instructions.

[Cellstrat] If you have any questions about the users flow, follow what you see in sana.ai.\

[Cellstrat] Fix the broken process for whenever the users asks for an Edit. It current throws an error and crashed the app.

Reviews

Sandbox access to the review API, pulling in all key review details for discovery and analysis. You already have access to the the account using the same credentials, here is the API call:

https://reputationmanager.io/api/getReviews

{

"from": 0,

"size": 500,

"profileId": "9be8687429dc971a4b4ba222943391bc",

"filters": {"dateRange": { "from": "2025-01-01","to": "2025-01-26" } },

"readMask": "storeCode,title,masterAddress,locationKey,entityTags,createTime,locationId,detectedLanguage,source,timeZone,isReviewFromPubSub,reviewer,starRating,comment,commentWordCount,sentimentAnalysis,websiteUrl,lastUpdateTime"

}

Milestones

[Cellstrat] Access the profile review data set for the last 90 days up to the last 500 reviews.

[Cellstrat] Home page navigation button triggers the correct file download and presents the already providing greeting conversation to initiate.

[Cellstrat] Provide [LC] with the outline of how we need to provide and store help data. A very simple set of data that can answer questions about what the user can do with this section.

[LC] Create sets of Tasks for Reviews Analysis that provide insights and reporting not available from the platform.

[Cellstrat] Review the task prompts to identify language suggestions to improve user experience.

[Cellstrat] Provide LC with instructions on how to provide a systemic markdown so that the review analysis outputs look polished.

[LC] Different card background style for Tasks from different Data sets.

API Updates

LocalClarity APIs gain new fields that are incorporated into existing workflows.

[LC] Additional Categories

[LC] Link to listing edit detail page

Imagery

Accessing image files to power more colorful and engaging responses.

[LC] Update the listing and review APIs to include imagery.

[Cellstrat] Build single response review markdown that includes presentation of images.

Google Edits

Writing the json output that is to be sent to Google for all desired edits.

Milestones

Sample code sets established in the app documentation for all major edit categories.

Strong confirmation chat workflow developed to assure users of the steps.

Simple code generation to push edits to Google for 1 listing.

Simple code generation to push edits to Google for 10-20 listings.

Simple code generation to push edits to Google for 100 listings stress test.

Agent Tasks

Milestones

[Cellstrat] Build interface to schedule Task execution (following OpenAI model design within the framework of Lia’s design)

[Cellstrat] [LC] Build email template outline — Simple email structure that is also similar to OpenAI but one that includes the content within the email template.

[Cellstrat] Trigger the execution of a Task and delivery an email.

[Cellstrat] Build interface to add/edit/delete existing Tasks.

Static Data Sets

Milestones

[LC] Provide a handful of data sets to be used in combination with listings or reviews.

[Cellstrat] Include fixed data sets of Google categories and services.

[Cellstrat] Add additional sets of distinct news and knowledge.

[Cellstrat] Build interface to allow the LocalClarity business team to add more

Platform Direct Integration

Milestones

[LC] Business team access to the interface directly from within the LocalClarity platform as SuperAdmins.

[LC] Purge routines to make certain no cross-profile data transfer.

[LC] Access control to SuperAdmins of other domains.

[LC] Access control at the Tier level in platform.

API Updates

Milestones

Google 2nd API into the platform to look for differences in data sets

Multi-source API conditional responses. “Will you be working with just Google or multiple listing sources?”

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.