What was said
What this says
Current to 10 Oct 26Tobias Renwick walked Craig Paterson through configuring and loading the call recorder connector. It connected, the first backfill began, and Craig agreed an evening change window.
What happened: On the 12 Apr 2023 onboarding call Craig created the connection with calls and transcripts as the only scope. The status went green and the first backfill started, with records indexed climbing.
Customer view: Craig was relaxed about the rate limited warning and said he would rather it was slow and polite. His worry is a connector that stays in retry, and he will raise any fault with a screenshot and log excerpt.
Commercial: Next come the helpdesk connector and then the accounting system, which needs the finance manager's approval through Imogen. These widen the brain's roots in the account.
Watch: Tobias asked Craig to report how the load looks the next morning. Flat records indexed for several days would mean a stuck connector or a changed source.
Themes
Current to 9 Oct 26Who was there
BB-Demo
Them
Follow-up
Next steps
Current to 9 Oct 26- Tobias Renwick writes the evening change window into the plan and sends it to Craig and Daniel.
- Tobias sends the scope note for the accounting system that week.
- Craig checks the console the next morning and reports how the load looks.
Risks
Current to 9 Oct 26- The backfill showed a rate limited warning and would be a fault if it stayed in retry with no growth.
- The accounting connector depends on the finance manager's approval coming through Imogen.
Opportunities
Current to 9 Oct 26- The helpdesk connector is next and the accounting system follows, deepening use of the brain.
Transcript
- Tobias Renwick NoneAfternoon, Craig. Can you hear me? I've got you on the screen but I'm not certain the sound is working.
- Craig Paterson NoneAfternoon. Aye, I can hear you fine. I've had a long day, so you'll have to forgive me if I'm slow.
- Tobias Renwick NoneNot at all. Before we start, what do you need from this session? It's your system, so I'd rather work to your list than mine.
- Craig Paterson NoneSimple enough. I want the call recorder connected, I want to see it load, and I want to know what to watch afterwards. I'll start by reading out what the console says. Workspace is up, credential is created, connector status is not yet configured.
- Tobias Renwick NoneGood. Then let's configure it. In the console, choose add connector, pick the call recorder, and paste the credential you created. I'll talk you through each field as you go, so stop me if I'm going too quickly.
- Craig Paterson NoneRight. Pasting. And it's asking for a scope. I can see calls and transcripts, and a couple of other options I don't recognise.
- Tobias Renwick NoneTo be precise, choose calls and transcripts only. The others are for participant metadata and for things you may not want in the brain at all. We can always widen the scope later, and narrowing it later is more awkward, so start small.
- Craig Paterson NoneDone. Right, it's saying validating. Hold on. Connected. Status has gone to green.
- Tobias Renwick NoneExcellent. That's the connection. The load itself is a separate step, and it's the first backfill that takes time. It reads the history and indexes it, and for a first load that's measured in hours rather than minutes.
- Craig Paterson NoneUnderstood. I'll start it. Actually, hang on. It's running and it's reporting records indexed climbing, which is what I want to see.
- Tobias Renwick NoneThat's the number I'd watch. Records indexed should rise steadily during the backfill and then grow by a small amount every night. If it goes flat for several days, the source has probably changed, or the connector is stuck.
- Craig Paterson NoneNow it's showing a warning. It says rate limited, retrying. That's the connector timing out again, I think, or something very like it. I've seen the same with another tool.
- Tobias Renwick NoneIt's not quite a timeout, so let me be careful. The source system limits how fast we may read, and the connector backs off and tries again. It's the trade-off of a large backfill: slower, but we stay within the limits that your source system sets. It will recover on its own.
- Craig Paterson NoneThat's fair enough. I'd rather it was slow and polite. What would worry me is if it stayed in retry.
- Tobias Renwick NoneThen the line to draw is this. If it's still retrying after several hours with no growth, that's a fault and not a limit, and you should tell us. If it's retrying and records are still creeping up, leave it.
- Craig Paterson NoneAye. And if I do need to raise it, I'll send a screenshot and the log excerpt. I won't raise anything without them.
- Tobias Renwick NoneThat's exactly what helps us. Thank you. Now, while that runs, shall we agree the change window we talked about? I'd suggest evenings, after your business day closes.
- Craig Paterson NoneEvenings suit. Anything that moves should do so after the working day and well before seven the next morning. Just to flag, I'd want it written down so that if I'm off, someone else knows.
- Tobias Renwick NoneI'll write it into the plan and send it to you and to Daniel. And the nightly refresh itself will be complete by seven on business days, so the window and the refresh don't compete.
- Craig Paterson NoneGood. I'll check the console first thing tomorrow and see whether it's finished, and I'll keep an eye on it through the day.
- Tobias Renwick NoneThen the next step is the helpdesk connector, which we'll do the same way, on a call next week. Before that, the finance manager's approval for the accounting system should come through Imogen. I'll send the scope note this week.
- Craig Paterson NoneAye, that works. Right, I'll let you go. Thanks, Tobias. Regards.
- Tobias Renwick NoneThank you, Craig. Best, and do tell me in the morning how the load looks.