# NovaPay

Quick bit of context for anyone new here: NovaPay has been live with us since January 2025, processing £12M/month, and it's been a genuinely smooth integration. Tariq sees real-time payments as a strategic differentiator for their instant payout product — he wants NovaPay to be first in line when Atlas goes to pilot. Really excited about this one.

sounds good. I've been working on the atlas API spec — happy to walk through the technical requirements with you so you can share with Tariq's team

That would be amazing, Tom. Can we do that early next week? I want to go back to Tariq with something concrete.

Had a great call with Tariq this morning — he's really engaged on the Atlas pilot. His main questions: (1) what does the integration look like — new API endpoints or an extension of v1? (2) latency expectations for real-time confirmation, (3) rough timeline to pilot. I said I'd come back with answers this week.

it's new endpoints, part of API v2. the real-time confirmation target is sub-second for payment initiation — settlement is the harder part technically. re: timeline, I need to check with David on the broader atlas schedule but from my side the API design is in good shape

Sub-second confirmation is exactly what Tariq wants — his product team is already designing their UI around it! What does the integration effort look like on NovaPay's side?

if they're already on our v1 API it should be fairly straightforward. new auth scope, new endpoints, webhook format changes. maybe 2-3 weeks of integration work on their side once we're ready for them

Perfect. I'll put together a pilot brief for Tariq — scope, timeline, what we need from them. @david-kimura — are you able to confirm the Atlas timeline? Tariq is asking about a pilot start date and I want to give him something solid.

Alex — Atlas is on track for a Q1 beta. I'll share a more detailed timeline at the engineering all-hands on the 25th. Happy to have a quick sync before then if you need something for NovaPay.

That would be great, thanks David. I'll reach out to schedule something.

Just shared the preliminary pilot brief with Tariq. He's sending it to his engineering lead. The energy from their side is really positive — NovaPay could be a fantastic reference case for Atlas when we launch more widely!

Tariq came back with detailed technical questions from his engineering lead. Sharing here so Tom can take a look. Questions are around: webhook delivery guarantees, idempotency for real-time payments, and how Atlas handles payment failures differently from the current async model.

good questions. webhook delivery is at-least-once with retry, same as v1. idempotency works the same way — key per request. the failure model is different though: with real-time you get synchronous confirmation so there's no ambiguous 'pending' state the way there is now. I'll write up a proper technical FAQ, give me a couple of days

Thanks Tom, really appreciate it. Tariq also asked about batch payments — they do a daily payout run of ~2,000 individual payments and they'd love a batch submission endpoint rather than hitting the API 2,000 times sequentially. Is that on the roadmap?

yeah, batch submission is part of API v2. it's on my list. not sure exactly when it ships though — depends on how the atlas work sequences out

fyi Alex — I just pulled NovaPay's webhook stats for the last 90 days. their failure rate is basically zero. really clean integration.

Good to know, thanks Lena! That's consistent with what Tariq says — they take their engineering seriously.

Spoke to David — he confirmed Atlas pilot target is late March / early April for initial API access. I've told Tariq we're looking at April for pilot kickoff. He's really pleased. Going to set up a joint technical session for the first week of April — Tom, I'll want you on that call if you're available.

let me know if his team has follow-up questions. happy to jump on a call with their engineers if that would help move things along

Tariq loved the technical doc. His exact words were 'this is the kind of detail that makes me confident in the partnership.' He's provisionally blocked out his engineering team for an April integration sprint. This is going so well.

Also — Tariq mentioned that at their last board meeting he presented Acme's real-time capability as a strategic differentiator for NovaPay. His board is excited about it. No pressure or anything! 😅

Quick update — I was going to confirm the April pilot kickoff with Tariq today but wanted to check in first. @tom-brennan @david-kimura — are we still on track for initial API access this week?

hey alex — slight issue. the pennington settlement incident last week pulled me and lena off atlas for most of the week. we're behind on the pilot environment setup. it's not going to be ready this week

OK. What's the revised timeline?

realistically? late april. I've got some incident follow-up work to finish and then I can get back on it properly. sorry

That's frustrating but I get it — these things happen. I'll talk to Tariq, he'll understand. @david-kimura — flagging that we're slipping on the NovaPay pilot. Tariq has his engineering team ready to go and we're pushing back by three weeks.

Thanks for flagging Alex. The Pennington incident was unavoidable but Atlas remains a priority. I'll work with Tom on getting the pilot environment fast-tracked once the incident work wraps up.

Spoke to Tariq. He was understanding — said 'these things happen, just keep me in the loop.' I've told him we're targeting late April. He's adjusted his sprint planning accordingly.

Any update on the Atlas pilot environment? I told Tariq late April and I want to make sure we're still tracking to that.

I'm back on atlas this week. the API endpoints are solid but the pilot environment needs the settlement piece wired up. working on it

Thanks Tom. Tariq's been asking about the batch payment endpoint separately — is there any chance that could ship independently of Atlas? Even without real-time, a batch endpoint would be a meaningful win for them right now.

it's architecturally tied to the v2 API which is built on the atlas data model. I can't really decouple it without doing the work twice. let me think about it though

Team — I need an honest update for Tariq. Are we going to have the Atlas pilot ready by end of April? He's asking and I don't want to give him another date I can't keep.

I'm going to be straight with you — atlas isn't going to be ready this month. David pulled me onto the vertex API response time investigation last week and I've barely touched atlas since. the pilot environment is maybe 40% done

When can we realistically have it ready?

honestly? I don't want to give you a date I can't hit. the atlas OKR has been moved to Q2 and the team keeps getting pulled onto incident work. if I get uninterrupted time, maybe 4-6 weeks. but I haven't had uninterrupted time since January

Alex — I appreciate the pressure from NovaPay's side. Atlas remains a strategic priority and the team is committed to delivering a quality pilot experience. We've been stretched across some critical incident response work but I expect capacity to free up as we move into May. I'll have a clearer timeline for you by end of next week.

David — I need something more concrete than 'end of next week.' Tariq has blocked engineering resource twice now based on timelines we've given him. Every time we push back, I'm spending credibility I can't get back.

Understood. Let me regroup with the team and come back to you.

Called Tariq. Told him we're looking at Q2 now rather than April. He asked me directly: 'Is Acme actually going to deliver this, or is it vapourware?' I reassured him but it's a fair question. He's still on board but the tone has changed.

Tariq asked on our call today whether there are any API v2 features we could give them independently of Atlas. Specifically the batch payment endpoint — he said 'even if Atlas slips further, batch payments would solve a real problem for us right now.' @tom-brennan — is there any way to make that happen?

I looked into this after you asked before. the batch endpoint uses the v2 request schema which is built on the atlas data model. pulling it out would mean building a compatibility layer between v1 and the new batch logic. it's not impossible but it's probably 3-4 weeks of work and I'd need to pause other things

Would it be worth it? NovaPay is doing 2,000 individual API calls for their daily payout run. A batch endpoint would genuinely improve their experience — and it's something we can actually ship.

I think it's worth exploring. but I'd need David to sign off on the prioritisation. right now I'm supposed to be on vertex remediation

@david-kimura — thoughts on Tom pulling forward the batch payment endpoint for NovaPay? It's independent value even without Atlas and it would give Tariq something tangible while we work toward the pilot.

Let me look at the capacity plan. We've got the Vertex remediation and some other commitments. I don't want to make promises we can't keep.

Heads up everyone — had my regular check-in with Tariq this morning. A few things to share.

First — he's still interested in Atlas and still wants to be the pilot customer. He asked about the timeline and I said we're targeting Q2 but couldn't give a specific date. He accepted that.

Second — he mentioned that Clearpay has approached NovaPay about their real-time payment capability. He was very clear that he's not evaluating them and he's not making a threat. He just wanted to let me know that the market is moving and NovaPay can't wait indefinitely.

that's not great. did he say what clearpay is offering?

He didn't go into detail. He was very matter-of-fact about it. But the subtext was clear — if we can't give him a timeline, he has other options.

for what it's worth, clearpay's real-time stuff is pretty basic from what I've heard. but that doesn't matter if we can't ship ours

Exactly. I need to go back to him with something concrete soon. Even if it's not Atlas, we need to show momentum. The batch endpoint, a firm beta timeline, something he can take back to his board.

I've been thinking about what we can realistically offer NovaPay in the short term. A few ideas and I'd really value input from the team:

1. Pull forward the batch payment endpoint — even if it means a v1-compatible version rather than waiting for v2. 2. Give Tariq's team early access to the Atlas API spec so they can start integration planning now, before the pilot environment is ready. 3. Commit to a firm Atlas pilot date with David's sign-off, even if that date is July.

option 2 is easy — I can share the API spec tomorrow. option 1 is more work but I think I could build a v1-compatible batch endpoint in 2-3 weeks if I actually get the time. option 3 — I genuinely don't know. I don't want to commit to july and miss again

on the batch endpoint — NovaPay's current integration is consistent enough that a v1-compatible batch wrapper would work cleanly. their payload structure is uniform across all their calls. I can help with the spec work if Tom wants to build it

that would really help, thanks lena

This is really helpful. I want to put together a proper proposal for Tariq — not just 'Atlas is coming, trust us,' but a concrete package of things we can deliver now and a roadmap for the rest. Something he can take to his board.

Alex — I think that's the right approach. Let me know if you need me on a call with Tariq. Happy to represent the engineering commitment directly.

Update — I've drafted a proposal for Tariq. It includes: (1) API spec access this week so his team can start integration planning, (2) v1-compatible batch payment endpoint targeting end of May, (3) Atlas pilot access in July. I want to get this right before I send it. Tom — does the July target hold?

the batch endpoint timeline is tight but doable if lena helps with the spec. for atlas — I think july is possible but not guaranteed. maybe say 'targeting july, will confirm by mid-june' rather than committing hard?

Fair. I'll soften the Atlas language. The batch endpoint is the headline — it's something tangible we can actually deliver.

Shared the Atlas API spec with Tariq's engineering lead today. He was impressed with the design — said it's 'well thought through.' Small win but I'll take it right now.

Working on the final version of the proposal today. Meeting Tariq on Thursday. If anyone has anything to add or anything I should know before I go in, now's the time.