Today's Plan / Patch Notes / Migration Guides

Okay, sometime today / tonight - v2.1 be pushed, afterwards the message and call filter will be pushed to stop rate limit and filter for quality traffic.

There will be an migration guide for those to switch over from Stripe Express to Stripe Oauth. And a minute-to-wallet interrogation guide as well. Before reaching out to support, see if your concern is addressed in the patch notes. We'll be standing by closely all day today / tomorrow to ensure clean and easy transition.

Here are the patch notes:

## Platform / infrastructure:

We now have a pre-processing env for messages and calls to filter out duplicate messages caused from GHL double installs - messages will only come through twice
We now process only active, paid workspace sub-accounts - all other data will be filtered to clean up rate limiting and processing - should make everything generally faster and more reliable (messages not showing etc)
Autopilot settings now active - sleep timeout, max responses etc, keep an eye out for stops based on these metrics
Call handling and re-checking - we now do a "double-take" on calls to ensure timestamps and information is accurate and accurately collected and displayed
Billing - billing a "double-take" to ensure information and limits are accurate because they are more heavily enforced with this update, example if sub-account limit is 3 but there are 4, we have time and creation settings to disallow this but will require action on your part before limits are automatically placed, calls will not process if there is no wallet balance / wallet auto-pay (post-v2.1 push to ensure easy transition)
Minutes -> wallet conversion, we will buy back minutes (from direct accounts) to fill your wallet initially so there is no down time
More error handling on the GHL and LLM side
.... and more

## Agency / Subaccount Management:

Edit permissions (tab viewing) has been patched and updated to hide and automatically re-route home page for sub-account (defaults to all)
Edit account (new) edit the information pulled via oauth like location name, address, website etc, we use this to build a profile for assistant creation (defaults to oauth information)
Manage limits (patch & new) has been patched and expanded upon - Allow for unlimited or define a hardset limit for assistants, messages, calling time (in mins), and phone number purchasing (defaults to unlimited on all)
Access control & security (patch and new) has been expanded upon ensuring data is safe, secure and only accessible by people you want it to be - invite members (clients or other users) to the sub-account and manage access control (remove) ---- *** For initial push, this will be defaulted to allow all, in the future this will be defaulted to only invited members
Connection options (new) create and manage connection to sub-accounts with the connection menu - Create / manage connection at the agency level or sub-account level, allow external sub-accounts to be added an connection to be reset at the account level (defaults to agency connection)
Re-billing (patched & new) connection switched from stripe express to stripe oauth meaning you will connect your own stripe account rather than creating a stripe account through express. Rebilling & subscription management are managed separately for wallet billing, a client can be rebilled without a subscription, a client can have a subscription without rebilling and either / or / both

## Agency / Settings:

Branding / White label (patched & new) will now show a preview of the client's view of your branding and allows for a color selector + hex code input
Domains (patched) now have further verification and dns checks and proper interface to relay information for domain connection and stability (later patch will allow for customization to favicon and meta information)
Agency (new) allows you to create default settings for account creation like a snapshot, set enabled permissions and account limits as a default
Access control & security (patch and new) has been expanded upon ensuring data is safe, secure and only accessible by people you want it to be - uninvited members get routed immediately
Wallet (new) replaces the minutes infrastructure and allows for easy account and billing management - phone numbers, phone calls, voice knowledge base storage, and more will be billed to the agency via a wallet system (like GHL), allows you to rebill for all available options to the client

## Subaccount / knowledge:

Voice enablement (new - post-v2.1 push, ~24hr later) allows you to enable a voice knowledge base / rag and 800ms processing speed, feeding vector data to voice without compromising quality ($8 / mo per enabled knowledge base - rebillable to client - can be shared between all assistants)
All file type uploads (new) allows you to upload any file type to be vector stored for knowledge (even images) and embedded text will be displayed to view
Knowledge base testing (new) allows you to run a embedded model to test your knowledge base information. This allows you to see where you are missing information and what information is not clear enough to the AI model - this will display a match score and the embeddings that were matched - you can use this to re-embed / add knowledge right from the testing portal
Message / testing message upload (new) allows you to embed knowledge quickly either through the knowledge testing portal or in live conversations with users - simply click on the thumbs up / thumbs down / plus button and the output will be displayed - edit if needed, select your knowledge base, then upload

## Subaccount / assistants:

Copy by ID (patched) allows you to take IDs from other builds and import them into your accounts (does not include tools calls at the moment, that will be in a near patch) - imported assistants are locked so someone cant steal your IP and sell it / claim as their own - you can also lock to prevent copying
Duplicate (new) duplicate assistants in your account exactly as is
Archive access / deletion protocol (new) allows you to retrieve deleted assistants from the past and current - simply navigate to archived and unarchive them - unarchived assistants will not count toward account limits or be shown in linked tools
Creation menu (new) allows you to create assistants either by generation, blank or copy by id

## Subaccount / assistants / tools:

"Bad Request" on add (patched) has been fixed and tool calls are now processed on the server side so tool limits are removed
transfer call (patched & new) now allows for warm transfer with a warm transfer prompt to use when tying the transferrer and transferee
Tool call names (patched) allows you to change the name for any tool type (tags, values, etc) so that you can put and value in the tag field and it not effect the data
Auto-correct names (new) prevents bad formatting on tool names and parameters - when creating / editing a tool, the name and parameter names will be auto-corrected to match proper formatting before allowing you to save to ensure no bad request come through (be weary of this when editing current tools that may have mapped params that are contactFirst - they will be formatted to contact_first)
Extraction (patched) allows for cleaner execution of value filling - input the custom field that you need and the description of the data that needs to be filled
add / remove tag (patched) prompt injected to help proper tool calls
Tool strict output (new) allows for us to force openai to output your arguments exactly as you have given, making tool call mapping and calling more reliable
Name and information associated (new) has been cleaned to be "pretty format" for better visuals
Examples (new) will inject example tool calls for you to view proper cadence and output
Tool output params (new) will include access token for ghl endpoint access in 3rd party platforms and more information to use in processing
Send email and self-schedule are patched and match the last conversed channel (defaults to SMS if none)

## Subaccount / AI voice

Call Center (new) Get an "at-a-glance" view of campaigns and voice metrics including juxtaposing call time vs volume, sentiment analysis of your account, end reason mapping, and status - search history by to, from and call id and filter by status, end reason and contact sentiment
Post-call webhook (updated) includes start and end timestamps, call type, call time in ms | seconds | mins, disposition and wallet metrics - we will also allow you to look for your own variables and map them (new - post-v2.1 push, ~24hr later) for data analysis on post-call processing
Contact tagging (new) will allow contacts to be tagged with disposition and call metrics (appt book, answered, no answer, etc) so workflows and disposition can be collected without webhook trigger
Transcripts to messages (new) will allow call information to be added to the overall conversation for contextual awareness between both models - summaries will be fed to voice along with direction context (inbound and outbound)
Orb will have appointment booking capability (new - post-v2.1 push, ~24hr later) and attribution reporting via contacts tab - view how many times someone has tried your orb, booked appointments, and injected context by past summaries - other widgets will be deprecated to be rebuilt and added-to
Internal contact system (new) allows us to process voice and general call data faster without waiting for integration sources (GHL) to send us information so we can process inbound faster leading to more stable uses

## Agency / Reseller / workspace management:

New workspaces will require a subscription to ensure a price floor for reselling to upkeep quality control and intrinsic value (we are don julio 1942, not jose cuervo) - this allows us to gain exposure to invest into resources, and also prevents "racing to the bottom" on pricing
Any user can create new workspaces at face value - resellers can create workspaces at a discount to resell for margin (requires reseller license)

... and more - will post more later

45 comments