Ask AI is the most amazing innovation yet from GHL. It shouldn't be just 'Ask and Have It' , instead, it must be 'Ask and Get It Done' and 'Proactively Done for You' . So please rename it to something fancier. Ask AI shouldn't need an MCP setup for GHL. It's internal, and so establish the connection automatically, with role-based access control built in. Core ideas: Ask AI must allow the creation of recipes or flows (define user journeys, that are a bunch of actions related to a certain business process — such as onboarding, quotes, scheduling, upgrades, manage inventory, running promotions, cancellations, etc.) Clients wouldn't always remember the names of workflows, pipelines, or contacts. So, Ask AI must be able to understand context and proactively prompt for and act on what they would like to get done and help them do more consistently. Use predictive analytics for competitive advantage. -> Is a super-intelligent assistant (like Jarvis!) -> Hook up Ask AI with Voice AI - so clients can call in and talk to their assistant. -> Seamlessly identifies and authenticates clients using a pre-recorded voice and not allow anyone who has access to the device to issue commands at will. Banks in the UK use voice-based auth for customer phone calls, so it isn't impossible. -> Clients interact with their assistant hands-free (mainly on mobile), or even chat when in meetings or at the doctor's, but need to get something done. -> Provides a summary of the day at the start, say 8 am. -> Provides insights or highlights with KPIs at the end of every day. -> When a client has to attend to something urgent and can't join the next call, Ask AI instantly shoots a message to the contact with a link to reschedule or, even better, initiates a conversation with the contact and guides them through rescheduling. -> Sends an automatic LC App notification checking in a few mins later after every 'confirmed' appointment, asks for a status, and updates show/no-show status, meeting summary and contact fields with relevant info, such as quoted price, etc., and moving opportunity stages, adding the contact to workflows, send an estimate or invoice, payment links, or even a text message, add tags etc. -> Proactively monitor OKRs and KPIs of the business and push the owner on areas where there's a slowdown. Simple Business Use Case: Voice AI books a job for a plumbing client, who receives the daily schedule in his mobile app, visits the customer's property, completes the job, gets back in the van to drive back his place, activates the 'Ask AI' assistant on mobile, confirms job status, estimated hours on the job and cost of parts required to fix a leak, and if an amount was quoted for the job incl VAT - and that's all. Ask AI then sends out an estimate with the required items, follow up until the quote's accepted, moves the opp to the quoted stage, sends out a text/email to the customer with ways to book another appointment, or uses the proposed project calendar to suggest a few slots optimising routes but prioritising urgency, agrees the date/time and schedules a follow-up appointment, moves the opp to the follow-up appointment stage, adds to the client's itenary for the day, tags as relevant, and sends reminders as needed. The plumber then visits the property, fixes the leak, and heads back to his place. Again sharing a quick update to the Ask AI assistant on his mobile, who then send out an invoice to the client, follows up for payment, and sends a review request post payment, working with opps and tags as relevant throughout. A lot of this would seem like it can be done using workflows and it's what we said at the start - creating recipes or flows for the AI (defining customer journeys), so everything falls in place, and gets done seamlessly in the backend - removing the need to build complex workflows, or involving wait steps, if/else logic, handling errors by intelligently deciding what the next steps are, all while acting like a true assistant. There'd be many more use cases for this to come.
