@SAXEMBERG 

Thank you for the feedback. We want to respond to both points directly and give more context on why these items were included and how the work fits into the long term stack.

1. On the concern that Technical Upgrades were already covered under the yearly maintenance referendum

We understand why this question came up. The key distinction is:

The yearly referendum covers platform maintenance and minor improvements.
The work billed here is multi-chain restructuring and infra rebuild that was not part of that scope.

Specifically:

• The Asset Hub migration required a fundamental rewrite of our indexer, schema, cache layers and chain-aware routing.

• The GCP migration, PAPI integration, and the re-architecture around multi-chain governance were not included in the yearly maintenance contract and were not planned at the time of Ref. 1463.

• These were completed at our own cost because they were urgent and ecosystem-critical. We are only requesting retroactive coverage now that they are delivered and stable.

• This is also why the budget is kept strictly retroactive and narrow. We did not include any of the future-facing features to avoid speculations till the roadmap for future is cleared out by the Foundation and Parity.

2. On the utility and current state of Klara

We understand the skepticism, and it is valid to ask whether current AI performance justifies investment. Two clarifications are important:

A) Klara today is intentionally limited, because it is not a wrapper

The early release only exposes a small, curated query surface, which is stated clearly in the articles 1 & 2.

This is by design:

• Klara is not a GPT wrapper.
• It is built to run on top of a DKG-backed knowledge base, where each knowledge asset must be structured, validated and added progressively.
• The scope expands layer by layer, not all at once, because the goal is verifiable, provenance-backed answers, not broad, uncontrolled output.

We are actively building this layer.

The current version available is only an MVP to evaluate need , assess queries and get feedback on sources needed.

Despite that :

• Klara has already passed 2500+ answered governance queries in its first month.
• These include questions users regularly struggle with: treasury flow, bounties, identity concepts, and how to participate in OpenGov.

As each new knowledge asset is added, the range of valid queries will expand.
Klara v2 (which the proposal references) includes the richer context, full provenance and the decentralised backend layer.

B) Klara is infrastructure for future Polkadot apps, not just a chatbot

As Polkadot moves into the product era, hundreds of parachain apps will need:

• onboarding bots
• governance explainers
• chain-specific assistants
• yield or portfolio explainers
• identity and PoP guidance
• compliant referral, reward or routing logic

Klara’s shared vector index, the DKG provenance layer, and plugin architecture are designed for these use cases.This avoids each team spending 50–150k to build their own AI pipeline from scratch.This is why the Treasury ROI is high even if you only measure avoided duplication.

We understand how Klara today invites questions and we have tried to answer all those here : FAQ on Klara vs Others , Klara as an AI Layer

3. On “users need analytics, integrations, and missing features from JS”

We agree  these are important, and most are already delivered or in progress:

• Treasury dashboard (balance, inflow, outflow, liabilities)
• Delegation stats
• Judgement explorer
• Bounty analytics
• Identity + proxy mapping
• Multichain indexer for Relay + AHM
• Stable RPC surface via PAPI

The AI request in this referendum is only one part of the proposal.
The majority of the budget is tied to exactly the kind of features you mention: analytics, identity, multi-chain infra, and a stable governance layer the ecosystem can continue to depend on.

On the tooling gap left by Polkadot JS:

We have already started discussion work on raw extrinsics, transaction review surfaces and others but we intentionally did not include them in the ask because we restricted the scope to shipped work till further clarity on direction is provided by the Foundation

TL;DR

• Klara v1 is limited on purpose because it is not a wrapper.
• The proposal funds the underlying AI and governance infrastructure, not speculative features.
• The analytics and infra upgrades address exactly the gaps the community has asked for.
• The costs billed were not included in the prior yearly referendum and represent delivered work the ecosystem is already using.

We remain fully open to answering any more concerns and to have a continued discussion

Thanks again for the comment and feedback.

Team Polkassembly