What's your opinion on using third‑party tools to automate Microsoft Office tasks that require syncing with Google Calendar, given the frequent errors?
I recommend avoiding Microsoft Office automation altogether. Instead, move to Google Workspace, which is more scalable, simpler, and less permission‑heavy. Making that switch alone can give you about a 30% productivity lift. After you're on Google, then you can worry about automating. People who make the transition tend to do really well, though it takes time to adjust and deal with the unnecessary complexity in Microsoft 365. For automations that need a calendar, the approach is the same on most no‑code platforms—you just use the built‑in calendar functionality.
Two ways to do this without needing a client's actual login. One: have the client create a Google Workspace account for you (or your automation account) and get invited as a user on all events automatically, so that account's calendar mirrors all the relevant events. Two: use Google Calendar's native sharing feature — the client shares their calendar with you (or specific events only, if some are private), and your automation runs off that shared calendar. Both let you build the automation without ever touching the client's login credentials directly. That said, I don't think a VAPI outbound calling agent is a very high-value approach to build your ads agency around in the first place — it's worth reconsidering the bigger picture, not just the calendar-access mechanics.
First, don’t jump straight to automation—build a solid, money‑making process first. Only automate a step if the process already works and there’s clear value in automating it. Look at each part of your workflow: outreach, nurturing, closing, onboarding, fulfillment. Closing is tough to automate without advanced AI, so focus on onboarding and fulfillment where automation is realistic. Sketch out your process (e.g., with a Whimsical mind map) before touching any automation tools. For example, a simple onboarding flow might start with a payment trigger (Stripe), then create a ClickUp folder, add a list, insert standardized tasks, send an onboarding email with a calendar link, update the client status when a call is booked, run the call, fill out an onboarding form, send a creative brief to AI to pre‑generate assets, turn those into a Google Doc and slides, and add a tracking doc. Once you have that map, see which steps can be hooked up via APIs or webhooks—like using a payment‑intent‑succeeded webhook to automatically create the ClickUp folder and list. Steps that can’t be automated via API (e.g., creating the folder or list) would instead trigger a manual action, such as a Slack notification to your project manager, which then starts the next automatable segment. In short, optimize the process first, make sure it’s profitable, then automate as much as you can.
Basically, Autobuild guy is not offering a specific solution; instead, he asks what you need help with, offers to do an audit, and then provides a hyper‑fine‑tuned automation solution. In theory it's a great idea, but in practice it's not very good because a detailed workflow business audit isn't as sexy as promising 20 booked appointments in 60 days with a money‑back guarantee. Everybody sells the latter because it's easier: you guess the problem, present the problem and solution with a guarantee, and skip the audit. You don't need to ask what the client needs; just assume they have a common problem and sell a solution for it. Agencies often struggle with lead generation at the top of the funnel, project management, client notifications, and client comps. I solve those with systems, create many campaign variants, test them at scale, and pick the best performer. That's what I do, and I don't recommend the audit approach. I'm still experimenting with audits for larger deals (e.g., leftclick) to see how it works, but generally productizing your service is the way to go, and I'd recommend that. We probably won't be doing the audit approach for long.