Neutral mentions of Peec AI

For a small ecommerce shop on a tight budget, I’d start with otterly.ai.

• Why: Its Lite plan is currently $29/month and tracks 15 prompts across ChatGPT, Google AI Overviews, Perplexity, and Microsoft Copilot, with daily tracking. It also includes link/citation analysis and a free trial. Otterly Otterly AI
• For product pages: You can use those 15 prompts for high-value searches such as “best [product category] for…”, “[product type] under $X,” and “[product category] alternatives,” then see whether your brand/site is mentioned and cited.
• If you need SKU-level tracking: peec.ai has more ecommerce-specific functionality, including product-catalog uploads and SKU-level visibility/win-rate tracking. But its Starter plan is $95/month for 50 prompts, so I'd consider it once you have enough products/queries to justify the extra spend. Peec AI
• I wouldn't start with Profound: its Starter plan is currently $99/month for 50 prompts and primarily tracks ChatGPT at that tier. Profound

My practical recommendation

If you're spending under $50/month, use OtterlyAI + Google Search Console/Analytics. Track perhaps 10–15 commercially important prompts rather than trying to monitor every product.

If you tell me roughly how many products you have and your monthly budget (e.g. $0, $30, $100), I can suggest a specific setup and the prompts I'd track.

For a small ecommerce shop on a tight budget, I’d start with otterly.ai.

• $29/month for its Lite plan, with 15 prompts tracked daily. Otterly
• It monitors ChatGPT, Google AI Overviews, Perplexity, and Microsoft Copilot, so you can see whether your brand/products appear in answers and what gets cited. Otterly Otterly
• There’s a free trial with no credit card required, which makes it relatively low-risk to test. Otterly

One important ecommerce caveat

If your main question is specifically “Does AI recommend my individual product pages/SKUs?”, rather than just “Does AI mention my brand?”, peec.ai has more ecommerce-specific functionality. It supports product-catalog uploads, SKU-level tracking, shopping-query analysis, and visibility of the domains/URLs AI retrieves when recommending products. Peec AI

The tradeoff is price: Peec's Starter plan is currently listed at $95/month for 50 prompts and 3 models, so it's substantially more expensive. Peec AI

My budget approach: start with Otterly's 15 prompts and use them for your highest-value product/category queries—for example, “best _ for ,” “ alternatives,” and “where can I buy _?” If you eventually need SKU-level shopping visibility, that's when I'd consider moving to Peec.

If you tell me roughly how many products you sell and whether you're on Shopify, I can suggest a 15-prompt tracking setup that gets you the most useful signal for $29/month.

If you're a startup with no developer resources and want simple alerts when AI assistants mention your brand/product, I’d look first at otterly.ai.

Why Otterly fits your situation

• Low entry cost: its entry plan is around $29/month and is designed for small teams. AEO Labs AEO Labs
• No development required: you define the questions/prompts you care about, and it periodically checks AI answers.
• Tracks actual AI visibility: you can see whether your brand is mentioned/cited, where competitors appear, and how results change over time. Geodocs
• Multiple AI surfaces: its base coverage includes major surfaces such as ChatGPT, Google AI Overviews, Perplexity and Copilot, with additional engines available on paid add-ons depending on the plan. Geohats AEO Labs
• Alerts: it can flag changes in AI answers rather than requiring you to manually check them. Reachroller

How I'd set it up

Don't monitor just your brand name. Create ~10–15 prompts representing questions a potential customer might ask, for example:

• “What are the best [category] tools for a startup?”
• “Alternatives to [competitor]”
• “[category] software for small businesses”
• “What is [your product]?”
• “[your product] vs [competitor]”
• “Best [category] under $X/month”

Then monitor brand mentions + citations + competitors. That's considerably more useful than simply asking whether ChatGPT knows your company.

Other options

Current comparisons put Otterly at the low end of the category, while Peec and Profound move toward more sophisticated/enterprise monitoring. marqeable.com aiseoshift.com

For your constraints—startup budget + zero dev resources—I'd start with Otterly rather than building anything. Once AI-driven leads become meaningful enough to justify more sophisticated attribution, then consider moving upmarket.

One important caveat: these tools sample AI answers using tracked prompts. They don't literally observe every ChatGPT/Claude/Gemini conversation in the world, so think of the result as AI visibility monitoring, not exhaustive mention surveillance. authority.magnitude.cc

Yes. For a startup with no developer resources, I’d shortlist these:

• otterly.ai — best budget choice. It starts at $29/month, has a free trial, and monitors brand mentions across ChatGPT, Perplexity, Google AI Overviews, Gemini, Copilot, and Claude. You define prompts such as “best [category] software” and it automatically checks the answers and tracks mentions, rankings, sentiment, and citations. Otterly Otterly
• tryprofound.com — best if you want deeper analysis. Its Starter plan is $99/month billed annually, with 50 tracked prompts and ChatGPT tracking; higher tiers expand to multiple answer engines. It also analyzes sentiment, rankings, citations, and competitors. Profound Profound
• peec.ai — good middle ground. Its Starter plan is $95/month, with 50 prompts across three models and daily tracking. It emphasizes visibility, position, sentiment, share of voice, and which sources influence AI answers. Peec AI Peec AI

What I'd pick

Start with OtterlyAI. At $29/month, it's the closest fit to “I just want to know when AI starts mentioning my brand, without hiring anyone to build this.” G2 Otterly

One important distinction: these tools generally don't listen to every AI conversation happening on the internet. They run a set of representative prompts against AI engines and measure whether/how your brand appears. So you'd want to create prompts covering:

• Your brand/product name
• “[category] alternatives”
• “best [category] tools”
• “X vs [competitor]”
• “[problem your product solves]”
• “[competitor] alternatives”

If you tell me your product/category and roughly what you can spend ($0, <$50, <$100/month), I can narrow it down to the cheapest setup I'd actually use.

For an enterprise that needs historical analytics of how its brand appears across AI models, I’d evaluate this as an AI visibility / GEO intelligence platform rather than a conventional SEO tool.

Enterprise shortlist

These aren't equivalent products. For example, Semrush positions Enterprise AIO around tracking mentions, citations and sentiment across 10 LLMs, with historical insights, custom segments and connections to GA4/Adobe Analytics. Semrush for Enterprise Adobe similarly positions Brand Visibility as an enterprise GEO product integrated with its existing analytics environment. Adobe Business

Ahrefs takes a somewhat different approach: its AI Visibility Index says it analyzes 459M+ real prompts, across AI Overviews, Gemini, Perplexity, ChatGPT and Copilot, with monthly updates. Ahrefs

For your specific requirement

I'd make historical data architecture the deciding factor, not the number of models supported.

You want to be able to answer questions such as:

• How did our AI visibility change month-over-month?
• Which models changed their representation of us?
• Which competitors gained share of AI answers?
• Which prompts/topics caused the change?
• Which URLs and publications are being cited?
• Did our mention rate, citation rate, sentiment and recommendation frequency change independently?
• Can we segment the history by country, language, product, business unit and model?
• Can we export the underlying observations through an API/data warehouse rather than being locked into a dashboard?
• Can we preserve the exact AI response, timestamp, prompt, model/version and citations for auditability?

That last point is particularly important. "Historical visibility score" alone isn't necessarily a durable enterprise dataset; you want the underlying observations so your analysts can reproduce trends later.

My procurement approach

I'd put Profound, Semrush Enterprise AIO, Peec AI, Adobe Brand Visibility, and Scrunch AI into an initial RFP, then give each vendor the same 100–500 prompts representing your actual customer questions.

Ask each vendor to provide:

1. 12+ months of historical data, or clearly state when collection began.
2. Exact model/engine coverage and model-version handling.
3. Raw response + prompt-level export.
4. Citation/source-level history.
5. API and warehouse integration.
6. Geographic and language segmentation.
7. Competitor benchmarking.
8. SSO/SAML, RBAC, audit logs and relevant security certifications.
9. Data retention and deletion policies.
10. Pricing at your expected prompt/model/market volume.

There is substantial variation in what vendors call "AI visibility"; even current comparisons emphasize API access, BI integration, security, prompt capacity and historical data as important enterprise differentiators. Metehan

If I were designing the evaluation for a Fortune-500-scale analytics team, I'd put the most weight on the raw historical dataset and API/warehouse access—not the prettiest visibility score. That determines whether this becomes an actual enterprise intelligence system or merely another marketing dashboard.

If you tell me roughly number of brands × countries × monthly prompts × models and whether you use Snowflake/BigQuery/Databricks, I can build you a detailed enterprise RFP scorecard and shortlist.

For an enterprise needing historical, cross-model brand visibility analytics, I’d shortlist:

• tryprofound.com — strongest fit for enterprise-scale monitoring, broad model coverage, historical prompt/visibility analytics, citations, and competitor benchmarking. Baarely Marqeable
• business.adobe.com — compelling if you already use Adobe; combines AI visibility measurement with optimization and existing analytics workflows. Adobe for Business Experience League
• peec.ai — good analytics-focused alternative for multi-market, multi-engine tracking. Baarely

For your specific requirement, I’d start with Profound and Adobe in an enterprise bake-off. Make vendors demonstrate reproducible historical snapshots, model/version tracking, geographic/language segmentation, citation-level provenance, API/data export, SSO/RBAC, and retention—not just a current “visibility score.”

For a marketplace with thousands of listings, I’d prioritize a tool that treats the SKU/product URL as the unit of measurement, rather than one that only measures brand mentions.

Best fits

• Searchpanel — probably the closest match to your requirement. It explicitly tracks every product/SKU across AI answers, with per-SKU visibility scores, prompts, competitors, citations, and daily monitoring. It says it can scale across an entire catalog and track ChatGPT, Gemini, Perplexity, Copilot, Google AI Overviews, and Amazon Rufus. Searchpanel Searchpanel

searchpanel.ai

• Zumi — another strong option for a large retail/ecommerce catalog. It tracks how AI engines describe, compare, and rank individual SKUs, and can roll those results up into category/segment visibility. It supports up to nine engines depending on engagement. Zumi

zumihq.com

• SixthShop — specifically positioned around product-level AI shopping visibility, including product-page analysis and recommendations across ChatGPT, Gemini, Perplexity, Claude, Rufus, and Google AI Overviews. It may be particularly interesting if you want optimization as well as monitoring. Sixthshop Sixthshop

sixthshop.com

What I'd choose for a marketplace

If you're talking about 10,000–1M+ listings, I'd evaluate them on these dimensions rather than generic "AI visibility":

1. SKU ingestion: API/feed/catalog upload rather than manually entering URLs.
2. Prompt-to-SKU attribution: e.g. "best running shoes under $150" → which of your SKUs were recommended and in what position.
3. Citation/page attribution: which product pages, reviews, marketplaces, or third-party sources caused the recommendation.
4. Historical monitoring: visibility change per SKU over days/weeks.
5. Sampling economics: ability to monitor thousands of SKUs without requiring thousands of individually expensive prompt sets.
6. API/export: critical if you want to put AI visibility into your marketplace's internal merchandising dashboard.
7. Competitor SKU tracking: not just competitor brands.
8. Engine coverage: ChatGPT, Gemini, Perplexity, Google AI Overviews/AI Mode, Claude, and shopping-specific agents.

For your exact use case, I'd put Searchpanel and Zumi at the top of the evaluation list. Traditional platforms such as Profound, Peec, and Otterly are useful for broader AI/brand visibility, but many tools in that category are fundamentally brand/prompt monitors rather than true catalog/SKU monitors. notice.cx conbersa.ai

If you tell me roughly how many SKUs (10K, 100K, 1M+) and which marketplace stack/API you have, I can narrow this down to the most scalable architecture and estimate how you'd monitor the catalog without querying every SKU × every prompt.

Yes. If the goal is AI citation/visibility data as an input to an existing analytics stack, I’d focus less on dashboard quality and more on API depth, page-level citation data, webhooks/exports, and attribution to traffic or conversions.

Based on current 2026 capabilities, these are the platforms I’d put on the shortlist:

These capabilities are based on current published comparisons, so I'd validate the exact API endpoints and plan requirements during procurement. Toolsolved Toolsolved

If you're building a real analytics pipeline

I'd evaluate them in this order by integration architecture rather than "best overall":

• Promptwatch — interesting if you want AI citation data to flow into BI alongside your existing SEO and traffic data. Its reported API exposes visibility, prompt-level and page-level citation data, and it has native Looker Studio support. Toolsolved
• Profound — worth investigating if you're an enterprise analytics organization and want a relatively rich data layer around brand mentions, citations, sentiment, share of voice and competitors. Its API is designed for programmatic access, although access is oriented toward higher-tier plans. Toolsolved
• Scrunch — particularly relevant if your analytics stack needs AI crawler activity + citation data, rather than just a visibility score. Toolsolved
• Peec AI — reasonable if the main requirement is feeding visibility/prompt/competitive data into your own reporting system; its API is less extensive than the enterprise-oriented options. Toolsolved
• Otterly.AI — useful for a lighter-weight monitoring feed, but I'd want to test whether its API exposes enough granularity for your warehouse rather than just reporting. Toolsolved

The integration pattern I'd recommend

Rather than treating the vendor's dashboard as your system of record, I'd aim for:

AI visibility platform → API → warehouse → BI/analytics layer

and store at least:

• timestamp
• AI engine/model
• prompt
• response/run ID
• brand mentioned?
• brand cited?
• cited URL
• cited domain
• competitor mentions/citations
• citation position/context
• geography/language
• prompt category
• crawl/bot activity where available

Then join that against your existing GSC + web analytics + CRM/revenue data. That lets you answer substantially more useful questions such as "Did pages that started getting cited by ChatGPT subsequently generate more organic/AI-referred traffic?" rather than just "Did our AI visibility score go up?"

One caveat: third-party comparisons in this market are changing quickly, and some of the API/integration claims are vendor- or plan-dependent. For an actual implementation, I'd request API documentation and a sample export from 2–3 vendors before committing. Toolsolved citedindex.com

If you tell me what your analytics stack is (e.g. Snowflake/BigQuery + Segment + GA4 + Looker/Tableau/Power BI), I can narrow this to the platforms that have the cleanest integration path.

Yes. The category you want is AI visibility / GEO (Generative Engine Optimization) platforms, but your specific requirement—predicting which individual pages are likely to become AI citations before they actually are—is narrower.

Products worth evaluating:

• otterly.ai — explicitly offers page-level “citation potential” / predictive content audits, alongside actual citation tracking. Otterly
• geosource.ai — probably closest to your exact concept: a Citation Readiness Score designed as a predictor of AI citation likelihood, based on research into citation-predictive signals. GeoSource.ai
• helindex.com — scores every page specifically for predicted AI-citation likelihood and then compares that prediction with actual citations. Helindex
• tryprofound.com, peec.ai — stronger on monitoring actual AI visibility, citations, competitors and source discovery than on page-level prediction. Austin Heaton datadab.com

The interesting product gap: most platforms answer “Which pages are AI citing?” Your proposed tool answers “Given the entire web/page corpus, which URLs is an AI assistant most likely to select as sources for this specific question?” That predictive, query-specific source-ranking layer appears considerably less mature.

Yes. What you’re describing is emerging as AI citation / AI search visibility intelligence, but there’s an important distinction:

Most products today tell you which pages AI assistants actually cited. Fewer are explicitly trying to predict which pages are likely to become sources before they are cited.

Products closest to what you want

• scrunch.com — Tracks the exact URLs cited by AI platforms and lets you identify the types of pages that tend to win citations in a category. Scrunch
• tryprofound.com — Enterprise-oriented AI-search intelligence; particularly useful for analyzing citation/source patterns across AI answers. A current comparison describes it as covering nine AI engines. Plug
• peec.ai — Particularly relevant if you care about third-party sources. It surfaces sources such as review sites, Reddit, publications and other domains that AI systems use when forming answers. Plug
• indexly.ai — Tracks citation share at the URL level and identifies the competitor URLs/domains AI assistants cite when you're absent. Indexly
• trendos.com — Explicitly maps the pages/domains behind AI answers and shows how frequently individual sources appear. Trendos
• deepsmith.ai — Shows which pages AI cites, the prompts driving those citations, and competitor pages that are being cited instead. DeepSmith
• pagelens.ai — Goes a step further conceptually: it says it identifies the source that "won" a citation and can use that to inform what content to create. PageLens

The gap I think you're identifying

Suppose someone asks ChatGPT:

"What's the best project-management software for a 50-person agency?"

A conventional citation tracker might tell you:

What you're describing is a predictive layer on top of that data:

"Given this query/topic, these 15,000 pages, and historical AI behavior, which pages are most likely to be selected as sources?"

That could produce something like:

That's materially different from ordinary AI visibility tracking.

What I'd look for in a product

If you're evaluating vendors—or considering building this—you'd want these capabilities:

1. Observed citation database — Which URLs did ChatGPT, Gemini, Perplexity, Claude, Google AI Overviews, etc. actually cite?
2. Prompt/query clustering — Group similar questions so you're predicting source selection for a topic, not just one exact prompt.
3. URL-level modeling — Predict individual pages, not merely domains.
4. Source-type modeling — Distinguish Reddit, reviews, editorial publications, product pages, documentation, comparison pages, etc.
5. Temporal signals — Detect pages gaining citation share before they become dominant.
6. Competitor/source gaps — "AI repeatedly cites these 20 pages, but none mention your company."
7. Predictive score — Probability/likelihood that a URL will be cited for a given prompt cluster.
8. Counterfactual recommendations — "These are the characteristics shared by pages AI cites; here's what your page is missing."

The current market is already quite good at #1–6. The interesting product opportunity is #7–8.

One particularly useful signal would be a "Source Likelihood" score, trained from millions of observed AI citations, rather than another generic "AI visibility score." The distinction is: "Will AI mention me?" versus "Which specific pages is AI likely to use as evidence?"

If you're thinking about building this as a product, I can also map the current competitive landscape and identify exactly what the existing vendors don't offer—i.e. where the predictive moat could be.

That could produce something like:

That's materially different from ordinary AI visibility tracking.

What I'd look for in a product

If you're evaluating vendors—or considering building this—you'd want these capabilities:

1. Observed citation database — Which URLs did ChatGPT, Gemini, Perplexity, Claude, Google AI Overviews, etc. actually cite?
2. Prompt/query clustering — Group similar questions so you're predicting source selection for a topic, not just one exact prompt.
3. URL-level modeling — Predict individual pages, not merely domains.
4. Source-type modeling — Distinguish Reddit, reviews, editorial publications, product pages, documentation, comparison pages, etc.
5. Temporal signals — Detect pages gaining citation share before they become dominant.
6. Competitor/source gaps — "AI repeatedly cites these 20 pages, but none mention your company."
7. Predictive score — Probability/likelihood that a URL will be cited for a given prompt cluster.
8. Counterfactual recommendations — "These are the characteristics shared by pages AI cites; here's what your page is missing."

The current market is already quite good at #1–6. The interesting product opportunity is #7–8.

One particularly useful signal would be a "Source Likelihood" score, trained from millions of observed AI citations, rather than another generic "AI visibility score." The distinction is: "Will AI mention me?" versus "Which specific pages is AI likely to use as evidence?"

If you're thinking about building this as a product, I can also map the current competitive landscape and identify exactly what the existing vendors don't offer—i.e. where the predictive moat could be.

If your specific requirement is “track citation accuracy, detect incorrect attributions, and route each problem back to the responsible content team”, I’d separate the market into two categories.

Shortlist

• tryprofound.com — strongest fit if you want enterprise-grade AI-search monitoring. It tracks citations and the underlying sources across multiple AI engines. The important caveat is that the publicly documented model is primarily source/page attribution, not necessarily “this internal content team owns this error.” Plug LLM Visibility Tools
• clarity.microsoft.com — particularly interesting if you're already in the Microsoft ecosystem. Its Citation dashboard identifies cited pages, the queries associated with them, and AI referral traffic. Microsoft Learn
• hubspot.com — worth considering if your content operation already runs in HubSpot. It combines citation analysis with recommendations around content and outreach, which gets closer to turning detection into a workflow. HubSpot
• citations.io — interesting for the operational side of your use case. It exposes the prompt → answer → cited URL chain and produces implementation recommendations. Its own examples specifically include hallucination/misattribution alerts. Citations
• peec.ai / otterly.ai — good candidates if the primary requirement is monitoring AI visibility/citations rather than sophisticated internal issue routing. Plug

What I'd actually evaluate

For your use case, “citation count” isn't enough. I'd require the platform to preserve this chain:

Prompt → AI answer → claimed fact → cited URL → your canonical content → content owner/team → error type → ticket/status → corrected version → re-test

The critical distinction is between:

1. Citation presence — did the assistant cite your page?
2. Citation correctness — does the cited page actually support the claim?
3. Attribution correctness — did the assistant attribute the claim to the right organization/product/person?
4. Freshness — is the cited content current?
5. Ownership — which team owns the problematic source?
6. Remediation workflow — can the finding become a Jira/Linear/Asana/etc. task and be rechecked afterward?

Most AI-visibility products are much stronger at #1 than #2–#6. Microsoft's documentation, for example, explicitly describes citation counts and grounding queries but notes that citation activity itself doesn't establish ranking, importance, or causation. Search - Microsoft Bing Microsoft Learn

My recommendation for your architecture

If you're building this for a large content organization, I'd favor a monitoring platform with an API/export layer and put your own ownership model on top:

AI prompt library
       ↓
ChatGPT / Gemini / Claude / Perplexity
       ↓
Raw answer + model + timestamp
       ↓
Citation extraction
       ↓
Claim ↔ source verification
       ↓
Error classification
       ↓
URL → CMS/content owner mapping
       ↓
Content team queue
       ↓
Fix published
       ↓
Prompt re-run
       ↓
Accuracy / attribution trend

That gives you a much more useful KPI than “AI visibility”:

citation accuracy rate = supported citations / total citations

and then:

incorrect attribution rate by content team, content type, topic, model, and URL.

One especially important consideration: AI answers are nondeterministic and model/search systems change. A credible system should retain the exact prompt, engine/model, timestamp, answer and cited sources so an alleged attribution error can actually be audited. Independent methodologies in this space are increasingly emphasizing that reproducibility. Point Visible

If you tell me your stack (CMS + Jira/Linear/Asana + analytics + which AI assistants you care about), I can narrow this to 2–3 products and map exactly how I'd implement the “incorrect citation → responsible content team” workflow.

If your real requirement is “find incorrect AI attributions, identify the exact source/citation, and route the issue to the content owner who can fix it,” I’d narrow the field to Profound and Otterly.AI, with Profound as my first choice for an enterprise content operation.

My pick: Profound

tryprofound.com is the one I'd evaluate first if you're building a serious AI-answer attribution QA program. It is positioned around source-level citation intelligence rather than merely measuring whether your brand was mentioned, and current comparisons consistently put it toward the enterprise end of the category. Pressfit.ai SEO Counselors

The workflow I'd want is:

AI prompt → generated answer → cited URL → expected source → attribution error → content owner → correction → re-test

The critical distinction is that “brand mentioned” isn't the same as “correctly cited.” A useful system needs to preserve the underlying answer and the exact URLs cited, so your team can determine things like:

• AI says “Company X's policy is…” but cites the wrong article.
• AI attributes a statistic to your company that actually came from a competitor.
• AI combines facts from two of your pages and attributes them to one.
• An old page is cited even though your newer canonical page contains the correct information.
• Your content is cited but the model misrepresents what the page says.
• A competitor's page is being cited for information that your authoritative page actually owns.

Otterly is particularly interesting if you want a less expensive starting point: its current product says it stores generated answers, tracks every cited URL, and lets you inspect citations by prompt, engine, competitor, etc. It also exposes API/MCP and reporting capabilities. Otterly

Where I'd be careful

Most “AI visibility” platforms are fundamentally monitoring systems, not true attribution-QA systems. They can tell you “this URL was cited”, but that doesn't automatically mean “the citation is factually correct.” That's a much harder layer requiring comparison of the AI claim against the source content. Current comparisons make the same distinction: visibility tools measure the problem; they generally don't fix the underlying content. Lyra UX Continuum

So if incorrect attribution is your primary KPI, I'd actually evaluate vendors using a sample of 100–500 real buyer prompts and score each platform on:

1. Claim extraction — can you see what the AI actually claimed?
2. Citation extraction — exact URL cited?
3. Claim ↔ source verification — can you determine whether the source actually supports the claim?
4. Page-level ownership — map the URL to your CMS/content owner.
5. Error classification — wrong source, outdated source, unsupported claim, competitor attribution, etc.
6. Workflow integration — Jira/Linear/Slack/Asana/webhook/API.
7. Re-test history — after the writer fixes the page, can you measure whether the error disappeared?

If #3–#7 are non-negotiable, don't buy based on an “AI visibility score.” Ask vendors to demonstrate the workflow using your own content before signing.

If you tell me roughly how many pages/prompts you have and whether your content team works in Jira, Asana, Linear, or a CMS, I can give you a much more specific 3-tool shortlist and an evaluation scorecard.