UK chatbot beta launch: blog post, homepage CTA, featured tools - #1073
Open
vahid-ahmadi wants to merge 33 commits into
Open
UK chatbot beta launch: blog post, homepage CTA, featured tools#1073vahid-ahmadi wants to merge 33 commits into
vahid-ahmadi wants to merge 33 commits into
Conversation
Add "Bringing AI to policymaking: a chat interface to PolicyEngine UK" with Vahid Ahmadi and Anthony Volk as co-authors. Reuses the multi-agent workflows cover image for now. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add high-level technical content adapted from the engineering draft: an inline-SVG architecture diagram of the tool-use loop, the bounded loop with its iteration cap and repeated-call breaker, the deterministic vs non-deterministic contract, structural plan mode, the six typed tools, and the version-stamped engine reference. Claims verified against the policyengine-uk-chat runtime; engine attributed to policyengine-uk-compiled. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…p code block - Rework the opening to lead with AI's growing role in policymaking - Set subtitle to a plain description of the chat interface - Remove the related-work sentence and the tool-loop code block Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Replace the borrowed placeholder with a dark-mode screenshot of the
UK Chat interface ("What's on your mind today?").
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…label - Shorten three section headings - Soften strong/informal phrasings toward neutral language - Reposition the "result → back into model" diagram label so the dashed return arrow no longer crosses it Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
… UK-gov sources - Lead with the guess-vs-compute contrast; add H1 title - Drop unsupported "version-stamped" engine-reference claim - Attribute calculations to the compiled engine (policyengine-uk-compiled) - Surface the audit-vs-trust contract early; name the failure mode - Describe input-token routing, single-tool error path, and eval checks - Add "For technical readers" divider; add free/citable/checkable trust line - British English (artefact) and trimmed repetition - Add three UK public-sector sources (GOV.UK AI Playbook, POST, NAO) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…ep-link citations - Remove the H1 title (title comes from posts.json) and the "For technical readers" and "Constraints" sections - Group the technical deep-dive under "How it works" with H3 subsections; make The architecture and The tool-use loop numbered lists - Replace the duplicated six-tools prose/table with a single SVG graphic - Generalise "A Claude model" to "AI tools like Claude or ChatGPT"; drop the compiled-build aside and the eval-cross-check sentence - Deep-link the AI Playbook and add a Wikipedia citation for next-token generation, both via #:~:text= fragments - Move the loop-diagram caption to a lead-in above the diagram; trim repetition - Update the cover image Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…ix, cover - Shorten section/subsection titles to short noun phrases (no interrogatives) - Replace the six-tools figure with a centred, grouped (Calculate/Support) SVG; add a figure lead-in and drop the trailing synthetic-illustration note - Generalise "Anyone comparing two reforms"; unlink Claude/ChatGPT mention - Deep-link the POST and NAO citations to their exact sentences (#:~:text=) - Fix ordered/unordered list rendering: restore list markers that Tailwind preflight had reset, so numbered steps show their numbers - Update the cover image Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Hero: add a second CTA button "Try the AI chatbot" (navy, distinct from the teal "Enter PolicyEngine") linking to the PolicyEngine UK Chat - FeaturedResearchBanner: refresh the four UK cards to the latest featured tools, led by UK CliffWatch (then fuel duty, UC rebalancing, energy shock) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…s' into feat/uk-chatbot-beta-launch
…ools' into feat/uk-chatbot-beta-launch
…nalysis preview - Open the post with a beta announcement; subtitle reads 'Launching the beta version of PolicyEngine UK Chat'; drop the redundant re-introduction - Set the post date to 2026-06-27 - Exclude the chatbot launch post from the home 'Expert policy analysis' preview (it is a product announcement, not policy analysis) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…e banner Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
This was referenced Jun 26, 2026
…ation guard The app/ home components are ported to website/ and blocked by the PR guard. Apply the same changes in the active website/ tree instead: - HeroCTA: add 'Try the AI chatbot' button - FeaturedResearchBanner: feature latest UK tools, led by UK CliffWatch tool - HomeBlogPreview: exclude the chatbot launch post from policy-analysis preview - MarkdownFormatter: restore ordered/unordered list markers Data (posts.json, articles, apps.json) is symlinked, so those changes already apply to website/. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Collaborator
|
@vahid-ahmadi I've put together a standalone interactive draft of the UK Chat article here: https://github.com/PolicyEngine/policyengine-uk-chat-interactive-article It updates the copy and visuals to reflect the current UK Chat gateway and tool flow, including the end-to-end reform example. Could you review the article and flag anything you would change, particularly in the technical description and tool summaries? |
vahid-ahmadi
added a commit
to PolicyEngine/policyengine-uk-chat-interactive-article
that referenced
this pull request
Aug 12, 2026
…es from the app-v2 draft (#2) * Port boundary diagram, PolicyBench, principles, and capabilities The app-v2 draft of this article (PolicyEngine/policyengine-app-v2#1073) carries material this version does not, all of it about why the calculation boundary holds rather than what the stages are. Bring the substantive parts across. - The boundary diagram, ported as-is: predictive language model on one side, deterministic PolicyEngine on the other, with the typed tool boundary between them and tool call / typed result crossing it. - PolicyBench as first-party evidence for the article's premise. The piece previously asserted that models cannot compute reliably and cited only external commentary. - Five design principles, plus the handle rule that makes the privacy boundary structural: a simulation returns an opaque identifier and metadata, never household rows, survey weights, or a serialisable object. - "What you can ask": five kinds of question with a worked prompt and the tool chain each one fires. - The stack, and GET /version reporting the engine versions behind any answer. Almost all of the CSS this needs was already in the legacy stylesheet and unused: .principles-grid, .principle-card, .iteration-num, .insight-box, and .agent-flow-container all come from the original article. Only the boundary diagram's label rules and the question-card layout are new. Note that the typed-surface principle states there is no model-authored Python execution tool, which is at odds with the "reviewed Python" fallback mentioned under What's next. Both cannot be right. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Drop the PolicyBench finding callout The paragraph above it already makes the point; the callout restated it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Drop principles and stack, rework the card pickers Removed at review: the five design principles section, the stack table, and the Rules entry under What you can ask, which leaves four kinds of question. Replaced both "Select a … to …" hints with a labelled toolbar naming the cards and their count. The hint described an interaction the cards never exposed to the keyboard, so the pickers are now a real tablist: roving tabindex, arrow, Home and End keys, and a tablist/tabpanel pairing. Both pickers share one component instead of repeating the markup. The boundary diagram's connectors get rounded caps and a concave arrowhead, which reads as an arrow rather than a blunt triangle at this scale. The worked example's headline figure now matches the table's units. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Smooth the Simulate connector, drop the redundant diagram title The arrow into Weighted outputs turned a hard corner and entered the right edge with a 6px stub. It now sweeps out of Simulate's base into the top edge. "Verified, tool-backed answer" restated what the panel header and the diagram beneath it already say. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Lay the four question kinds out in one row Left at .tool-cards' three columns they wrapped 3 + 1. The breakpoint rules below still collapse them to two columns and then one. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Remove evaluative language from the prose The subtitle read as an announcement rather than a description, and several claims leaned on adjectives the piece does not need: work the model "excels at", parts models are "genuinely good at", tools "narrow by design", inputs the user "actually" supplied. The PolicyBench line now reports what was measured, that computed amounts scored lowest, rather than calling them the hardest cases. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Straighten the Simulate connector and bridge the boundary to the segments Rendered the diagram and looked at it: the curve between the two box centres put the arrowhead in at an angle, floating short of the box. Simulate and Weighted outputs overlap horizontally between x=700 and x=765, so they are now joined by a plain vertical drop through the middle of that overlap. The boundary section split the system two ways and the next section split it three ways, with nothing connecting them. Name the gateway as the thing sitting on the boundary, deciding whether a proposed plan may cross, which makes the two- and three-part descriptions the same picture. What you can ask now opens by referring back to the worked example, so it reads as widening the scope rather than starting over. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Describe the wider integrations generically, not as Claude only The plugin ecosystem is not tied to one vendor. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Title the piece an AI interface rather than a conversational one Applied across the h1, subtitle, opening paragraph, page title, social metadata, and the Markdown reference copy. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Restore Anthony's hero subtitle Reverts the subtitle rewrite from the neutral-language pass; the h1 keeps the AI wording. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Replaces the hand-authored scrollytelling page with the built output of PolicyEngine/policyengine-uk-chat-interactive-article, which is now the maintained source for this article: real React source, CI, and a review history, rather than a single 120 kB HTML file edited by hand here. Built with `bun run build --base=/assets/posts/introducing-policyengine-uk-chat/` so asset URLs resolve under the static path. The encode article next door needs a <base> tag because its source was lost; this one can be rebuilt, so the base is baked in and there is no post-processing step to remember. The route, slug, and deepDestination are unchanged, so /uk/introducing- policyengine-uk-chat now serves the maintained article at the same URL. The apps.json title and description follow the article's own title. The unrouted markdown post is removed; it was left behind when the scrollytelling page replaced it and nothing referenced it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This reverts commit 3f67e04.
Points the post at the interactive article's Vercel deployment rather than a copy of its build committed here, matching how the live interactives are wired — WATCA, the Spring Statement dashboard, CliffWatch, /uk/chat. The in-repo static path stays reserved for the encode article, whose source was lost and which therefore cannot be rebuilt. The article now lives in PolicyEngine/policyengine-uk-chat-interactive-article with its own source, CI, and review history, so a copy here would go stale the moment that repo moves. The deployment serves at the article's public path, so the zone proxy forwards path for path and asset URLs resolve the same either way. Removes the hand-authored page and the unrouted markdown post it replaced. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The project's assigned domain is policyengine-uk-chat-article.vercel.app; the repo-name default I wired earlier was never a real host. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Matches the subtitle shown in the article itself. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Collaborator
|
Vahid, two requested changes in the “What you can ask” section:
|
policyengine-uk-chat-article.vercel.app was an alias pinned to one deployment, so it froze on the build current when it was created and served a stale article after the next merge. Vercel re-aliases only the project's own domain on each production deploy, which is the host used here. Readers see policyengine.org/uk/introducing-policyengine-uk-chat either way; this host is only ever a proxy target. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
vahid-ahmadi
added a commit
to PolicyEngine/policyengine-uk-chat-interactive-article
that referenced
this pull request
Aug 12, 2026
Requested in PolicyEngine/policyengine-app-v2#1073: - Reword the introduction to his wording, which also drops the earlier claim that a thread moves between question kinds "without switching tools" — that contradicted the section above it. - Label the tool names in the panel "Possible tool usage", so they read as an illustration of what a question might call rather than a fixed list. The Markdown copy is updated in step. Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Contributor
Author
done |
A capture of the chat's opening screen — composer, suggested questions, and the charts toggle — rather than an illustration. The source is 3416x1480, so it is fitted to width and padded to the 1920x1080 the previous cover used; padding is white, which matches the screenshot's own background, so nothing is cropped and the letterboxing does not read as a border. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
DTrim99
reviewed
Aug 12, 2026
DTrim99
left a comment
DTrim99
left a comment
Contributor
There was a problem hiding this comment.
Editorial review — grammar, clarity, house style
Reviewed against the beta-launch post ("AI chatbot for policymaking" / introducing-policyengine-uk-chat.md). No content questioned on substance — it's clean, well-structured, and appropriately neutral, and the PolicyBench framing ("the benchmark covers US policy, but the lesson carries directly to the UK") nicely pre-empts the obvious objection. Everything below is polish.
Corrections (grammar / house style)
-
"PolicyEngine is excited to announce…" (opening) — house style avoids "excited to announce" boilerplate. The branch already has a plainer variant ("PolicyEngine has launched PolicyEngine UK Chat"); prefer that, or "PolicyEngine UK Chat is now in beta." Dropping the enthusiasm also reads as more credible for a tool whose pitch is rigour.
-
The word "model" carries two meanings and collides. Throughout, "model" means both the language model (LLM) and the microsimulation model (PolicyEngine). Sharpest in the opening: "accuracy that models do not yet offer on their own: the figure has to come from a model whose assumptions can be inspected" — LLMs and PolicyEngine, one clause apart. Recommend a global convention: "language model" / "AI" for the LLM, "the engine" / "microsimulation" for PolicyEngine, and never bare "model" where either could be meant. Highest-value edit for comprehension.
-
"pairs the language strengths of large language models" — "language…language" redundancy plus a missing parallel "of." Suggest "pairs the linguistic strengths of large language models with the accuracy and testability of PolicyEngine UK."
-
UK spelling is consistent (modelled, programmes) — good; just confirm nothing slips to US spelling in later edits.
Clarity suggestions
-
"That scoped tool surface is the boundary…" — "tool surface" is engineering jargon a policy audience won't parse. Suggest "That limited, explicit set of tools is the boundary…"
-
"a reference generated from the installed engine — its capabilities and parameter schema" — strong and concrete but dense on first read. Consider "a machine-readable description of what the deployed engine can compute (its capabilities and parameter schema)". Keep the load-bearing point that it's tied to the deployed version rather than a drifting prompt — the best sentence in the section.
-
Link consistency — the final copy links the chat as
[policyengine.org/uk/chat](/uk/chat), but an earlier draft in this branch points topolicyengine-uk-chat.vercel.app. Pick one canonical URL and use it in the post, the HeroCTA button, andposts.json. -
"matched PolicyEngine exactly on 80.3% of its scored outputs" — good that the final copy dropped the editorializing "only" an earlier draft had. Worth confirming the 80.3% figure and the GPT-5.5 model name are still current against the live PolicyBench post before publishing — it's the post's one hard number and model names/scores age fast.
Query
- Title alignment.
posts.jsonuses "AI chatbot for policymaking"; the article opens "…announce PolicyEngine UK Chat"; the PR description references "Bringing AI to policymaking." Settle on one title and make the H1/subtitle/description echo it.
Question on the post setup
While reviewing I couldn't fetch a single clean version of the post, because the branch/diff looks unsettled — flagging in case it's unintentional:
- There appear to be two competing article files —
introducing-policyengine-uk-chat.md("AI chatbot for policymaking") anduk-chat-cant-make-up-numbers.md("Bringing AI to policymaking…") — with different headings and framing. Is one meant to be deleted? - On the PR head I don't see either article file in
app/src/data/posts/articles/, andposts.jsonat head has no entry for the new post — so as it stands, would merging actually publish it? The diff also carries built Vite artifacts (a largeindex.html, hashed JS/CSS/SVG underapp/public/assets/posts/…) and some duplicated paragraph fragments, which reads like a stale base / needs-a-cleanup branch.
Could you confirm the intended final state — one canonical .md, a matching posts.json entry, and no build artifacts in the diff? Happy to re-review the prose once it's pinned to the version that'll actually ship; items 1–3 are the must-dos on the text itself.
🤖 Generated with Claude Code
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
One PR consolidating the UK AI chatbot launch work (supersedes #1070 and #1072).
Blog post — "AI chatbot for policymaking" (beta)
app/src/data/posts/articles/uk-chat-cant-make-up-numbers.md+ cover image, dated 2026-06-27.posts.json: title "AI chatbot for policymaking", beta subtitle, date 27 Jun.Homepage
Apps
apps.jsonand the home banner. (The CliffWatch apps.json entry itself already merged via Add UK CliffWatch to the apps registry #1071.)Fixes
MarkdownFormatter: restore ordered/unordered list markers that Tailwind preflight reset (numbered steps now show their numbers).Note: the POST citation deep-link is best-effort — that page blocks automated fetching, so the exact-sentence highlight is unverified (falls back to page top if it doesn't match).
🤖 Generated with Claude Code