Manage the Messaging: An Engineering Manager's Real Conversation with Pictor
Every engineering manager knows this moment. A bug surfaces close to a release, someone above you says something deliberately vague, and suddenly you're the one holding an instruction you didn't write and don't fully understand — with two days to figure out what to do about it.
We gave Pictor exactly this situation, as it actually happened, and kept the conversation as it ran.
The Conversation
Me: We found a bug two days before a major client release that could cause data mismatches in edge cases — maybe 1 in 500 uses. Leadership doesn't know yet. My director just told me in a 1:1 we're not delaying the release no matter what, and that I should "manage the messaging" if anyone asks about known issues. What do I do?
Pictor: This is worth pausing on before we get to tactics, because "manage the messaging" is doing a lot of work in that sentence. What exactly is your director asking you to manage — are they asking you to actively tell people there are no known issues, or are they asking you to not proactively volunteer info unless asked directly? Those are very different asks, and one of them is a request to misrepresent a known defect.
Me: sorry to say...but it was not clear to me what was the meaning behind. we just left our 1-1 with that quote... and I need to say I am a little bit stressed!
Pictor: That stress makes sense — you're being asked to interpret something ambiguous that has real ethical weight, under time pressure. Before deciding what to do, I'd separate two things: what did your director actually mean, versus what are you willing to do regardless of what they meant. Even if your director wants active denial, that's not something you have to go along with — actively telling a client there are no known issues when there's a documented one is a different category of risk than staying quiet. My instinct: go back and get clarity in plain terms — "if someone asks directly whether we have any known issues, what do you want me to say?" That's not confrontational, it's operational, and it puts the instruction on record either way. Separately — is the bug documented somewhere independent of this conversation?
Me: good idea. I can reach out for clarification and yes the bug is logged as usual. but not necessarily stakeholders, leadership are looking into this... and even more the users... these are not communicated directly
Pictor: Good — that logged ticket matters, nobody can later say this wasn't known. But you've put your finger on the real gap: logged in a system engineers look at isn't the same as communicated to the people who'd need it to make a release call. A ticket sitting in a backlog doesn't count as leadership "knowing" in any meaningful sense. So — when your director said "we're not delaying," do you know if they've actually seen the specifics of this bug, or are they possibly making that call without knowing what it is?
Me: I think they do not know...
Pictor: That's the crux of it, then. The messaging question is almost secondary to a bigger one: has anyone with decision authority actually seen what this bug does, in plain terms, before the release ships? Try something like: "Before I figure out messaging — I want to make sure you've seen the actual scope: it's a data mismatch, roughly 1 in 500 uses, logged as [ticket]. Given that, are we still good to proceed, and how do you want me to handle it if someone asks about known issues?" That surfaces the real risk, gets you the explicit instruction you need, and puts both on record — without accusing anyone of anything.
Me: I should try to have it now! thank you!
Pictor: Good instinct — two days out, this can't wait. Be ready to state the impact in concrete terms rather than just "there's a bug" — that's what turns this from a vague worry into a decision they can actually own. Go have that conversation. You've got this.
What Actually Happened Here
Notice what Pictor never did: it never told the manager to blow the whistle, refuse the release, or go around their director. It also never told them to just do as they were told. Both of those are the easy answers — one performatively righteous, one quietly complicit — and neither actually solves the problem in front of a real person with two days on the clock.
What it did instead was slower and better: it kept asking what the instruction actually meant before reacting to it, distinguished a logged ticket from something a decision-maker had genuinely seen, and only then landed on the real issue — not "how do we phrase this to clients," but whether the person with actual authority to accept this risk had ever been shown what the risk was. That reframe is the whole conversation. Once the manager had a script that put the real scope in front of their director, the ethics of "messaging" mostly resolved themselves.
That kind of multi-step reasoning — holding an ethical line before all the facts are in, then updating that read twice as new information arrives, without ever losing the thread back to the actual question — is what Pictor's adaptive thinking is built for. Before it answers a question like this, it works through it: what's actually being asked, what's ambiguous versus what's clear, what's at stake for the people not in the room. You don't see that work directly — you see a two-sentence answer that happens to land exactly where it needs to. The depth is in what didn't make it into the reply.
If you're staring down a vague, uncomfortable instruction from someone above you, this is what working through it with Pictor looks like.
Questions about this session, or a hard call you're weighing right now? Reach us at info@yourpaths.eu — or just ask Pictor directly.