# Who to hire for Cimulate Cimulate is about how shoppers search and discover products. Hire a partner who can prove relevance with a test, and who keeps the storefront and the catalogue where they are. **Published:** 2026-08-04 **Canonical:** https://synconai.com/hire-cimulate-salesforce-partner **Author:** SynconAI Architecture Team ## In short Hire a Cimulate partner for search relevance and commerce data, with a measured test as the proof. Salesforce completed the acquisition on 3 March 2026 to strengthen search and discovery in Agentforce Commerce. It adds an intent-aware search layer. It does not replace your catalogue or your storefront, so the partner must keep both intact. ## Sources - **Salesforce Signs Definitive Agreement to Acquire Cimulate** (retrieved 2026-03-03) https://www.salesforce.com/news/stories/salesforce-signs-definitive-agreement-to-acquire-cimulate/ > Salesforce completed its acquisition of Cimulate on March 3, 2026. - **Salesforce Signs Definitive Agreement to Acquire Cimulate** (retrieved 2026-03-03) https://www.salesforce.com/news/stories/salesforce-signs-definitive-agreement-to-acquire-cimulate/ > Cimulate provides an intent-aware context engine purpose-built for retail. ## What to take away - Search relevance is proven by a test on your own queries and products, not by a demonstration on a vendor catalogue. - Relevant credentials are the B2C Commerce Developer and the Platform Data Architect, with Agentforce Specialist for conversational features. - A first engagement is one storefront, a set of real queries and an A/B comparison against current search, run for a stated period. - The catalogue, the storefront and pricing stay in their own systems. A partner who proposes to rebuild them has changed the project. ## The guide Open your site search report and look at the queries that return nothing, or return the wrong thing. A shopper types "something for a wedding in the garden" and gets a list of garden furniture. Another types a product code with a typo. Another asks for a colour your catalogue spells differently. Every retailer has a long tail of such queries, and the promise of intent-aware search is that more of them end in a sale. That promise is testable, which makes this the easiest of the acquisition hires to hold a partner accountable on. Salesforce completed the acquisition on 3 March 2026. The announcement says the company is "an emerging leader in AI-powered product discovery and agentic commerce" and that the deal strengthens search and discovery in Agentforce Commerce. It describes the technology in one sentence: "Cimulate provides an intent-aware context engine purpose-built for retail." It adds that the platform combines real and simulated shopper journey data. It also describes integration with Agentforce Commerce as something that will happen once the pieces are joined. Take that wording at face value when you talk to a partner. Ask what exists for you to deploy today, what is planned, and what they have actually run. ## Where search sits in a commerce stack Search is a layer between the shopper's words and your catalogue. The catalogue holds products, attributes, categories, prices and stock. The storefront shows results, filters and pages. Search interprets the query, retrieves candidates and ranks them. Cimulate fits in the interpreting and ranking part, which is why this guide insists that it does not replace B2C Commerce, and why a Salesforce sandbox does not give you a copy of the storefront to experiment on. That separation tells you what to hire for. Someone must understand how your catalogue is structured, since poor attributes hurt any search engine. Someone must understand how results reach the page, whether through the storefront's own search APIs or through a headless front end. Someone must run the experiment. Rarely is one person strong at all three. ## Skills and certifications On the Salesforce side, the Salesforce Certified B2C Commerce Developer credential shows a person can work with the platform's storefront code, APIs and Business Manager. Ask for it for the person who will connect search to the storefront. Salesforce Certified Platform Data Architect fits the catalogue and customer data side, particularly where product attributes are inconsistent. If conversational shopping features are in scope, Agentforce Specialist shows familiarity with Salesforce agents, though the Cimulate material is about search and discovery rather than a ready-made shopping assistant. Beyond credentials, ask about search relevance work. Has the candidate tuned synonyms, handled misspellings, managed zero-result queries and measured click-through and conversion by query group? Those are the daily tasks. A person who has done them can describe their last tuning cycle in detail. A person who has not will describe the technology. Ask also who will own the work after launch. Search needs regular attention, because the catalogue and the language of shoppers keep changing. A partner should propose a review rhythm and a named owner on your side. ## What to hand over before the first meeting You will get better proposals if you prepare four things. The first is an export of your top search queries and your zero-result queries for a recent period, with the volumes. The second is a short description of your catalogue: how many products, which attributes are filled in reliably, and which are not. The third is your current search setup, including any synonym lists, redirects and merchandising rules, because those represent years of learning that a new layer must not discard. The fourth is the numbers you already track, such as search conversion and revenue per search, so that a baseline exists before anyone changes anything. A partner who asks for these in the first week is working the right way. A partner who proposes a design without them is working from the brochure. Preparing the pack also tells you something about your own readiness: if nobody can produce the zero-result report, the first project is measurement, and that is cheaper than any search upgrade. ## What a first engagement looks like Pick one storefront, one category group and a period that avoids your peak trading weeks. Collect real queries from your search logs, including the failures. Group them: product codes, descriptive phrases, attribute searches such as size or colour, gift or occasion queries, and misspellings. Agree how success will be judged for each group, for example click-through to a product, add to cart or purchase. Then compare. Route a share of traffic to the new approach and keep the rest on your current search, for a period long enough to cover normal variation in trading. Report results by query group, because an overall average can hide a loss in a critical group such as product codes, where exact matching matters more than interpretation. Agree the stopping rule beforehand. If results by group do not clear a threshold you wrote down in advance, you stop or you adjust. It is much easier to be honest about this before you have spent the money. ## Peak trading and rollback Search sits on the critical path of every sale, so the calendar matters as much as the code. Build and test outside your busiest weeks, freeze changes before a major sale, and write down who can switch search back to the previous behaviour and how long that takes. Rehearse the switch once in a quiet period. A rollback that has never been tried is a hope. Ask the partner to include the rehearsal in the plan and to name the person on call on the first trading day after launch. ## Red flags A demonstration on the vendor's own catalogue proves little. Ask for your products and your queries. A claim of uplift quoted from another retailer is a claim about that retailer. Your catalogue, shoppers and baseline differ. A plan that replaces category pages, promotions or the product catalogue is a different project. Search can improve how products are found, and it need not alter how they are managed. A design that sends shopper data to a new party without a privacy review is a risk. Shopper journey data is personal information in most jurisdictions, and in Australia the Privacy Act 1988 and the Australian Privacy Principles apply. Ask where data is stored and for how long. A schedule that launches before peak trading with no rollback is reckless. An absence of a plan for exact-match queries is a warning. Intent-aware ranking must not bury the shopper who typed a product code. ## Merchandising and the human side of search Search is not purely an engineering matter. Merchandisers decide which products deserve prominence, which promotions should surface and which items should be hidden for margin or stock reasons. An intent-aware layer will rank by what it infers about the shopper, and unless those business rules are respected, it can undermine decisions your team made deliberately. Ask the partner how merchandising rules interact with the ranking, who can override a result, and how an override is recorded so that it is not silently undone at the next update. Include your merchandising lead in the first design session, not the final review. They know which queries matter commercially, which seasonal terms are about to spike and which products are awkward to describe. A short list of those from them is often worth more than a month of log analysis. ## Accessibility and multilingual queries Search is the main route into the catalogue for many shoppers, including those using assistive technology or typing in a second language. Ask whether the new search experience preserves keyboard navigation and screen reader support in the results, and how it handles queries in other languages or with unusual spellings of names and brands. If your storefront serves several markets, ask whether relevance is measured per market, since a model that works in one can fail in another. These questions rarely appear in proposals, so the partner's answer shows how thoroughly they have thought about real shoppers. ## Ten questions 1. What exists for us to deploy today, and what is planned but not available? 2. How is the catalogue supplied to search, and how often does it refresh? 3. How will we test on our own queries, and over what period? 4. How will results be reported by query group, including product codes? 5. What is the rollback if results fall in a group? 6. Which shopper data does the service receive, where is it stored and for how long? 7. How does it connect to our storefront, and which API or front-end change is needed? 8. Who tunes synonyms, rules and exclusions after launch, and how often? 9. What happens to promotions and merchandising rules we already use? 10. What threshold would make you advise us to stop? ## Checking evidence Ask a previous client for the experiment design, not the headline. How were queries grouped, how long did the test run, what happened to zero-result queries, and what was the result in the group that performed worst? A partner who has run a real test will have those facts to hand. Ask for the tuning log from a past project, with client names removed: which synonyms and rules were added and why. Ask to speak to the merchandiser who used the tool, since they see the daily reality. ## Commercial models and the statement of work A fixed-fee test with a clear design is the best first purchase. Integration into the storefront can follow on time and materials with a ceiling, and an ongoing tuning retainer should be priced separately from build. Keep platform charges with Salesforce and outside the services quote. For orientation see the [licence page](/cimulate-salesforce-licence-cost) and the [implementation cost page](/cimulate-salesforce-implementation-cost). The statement of work should state the storefront and category scope, the query sample, the groups, the success measures, the test design and duration, the data shared with the service, the integration method, the rollback plan, the tuning responsibilities after launch, and the exclusions, which should say that the catalogue and storefront are not being rebuilt. Add an acceptance test that reports results by query group to your head of digital. ## When to skip a partner If your catalogue is small, your search logs show few failures and your current engine can be tuned with synonyms and rules, a good merchandiser may close the gap without outside help. If you cannot yet measure search performance, begin by measuring it. And if the storefront is being replaced soon, defer the project until the new front end settles. More reading: [what happened to Cimulate](/what-happened-to-cimulate), [integrating Cimulate with Salesforce](/integrating-cimulate-with-salesforce), [should I migrate from Cimulate](/should-i-migrate-from-cimulate) and [Cimulate and Agentforce Commerce](/cimulate-agentforce-commerce). The primary source is the [Salesforce announcement](https://www.salesforce.com/news/stories/salesforce-signs-definitive-agreement-to-acquire-cimulate/), and the developer credential is on [Trailhead](https://trailhead.salesforce.com/credentials/b2ccommercedeveloper). | Query group | Why it needs its own result | Typical risk | | --- | --- | --- | | Product codes | Exact match matters most | Interpretation buries the right item | | Descriptive phrases | Where intent-aware ranking should help | Plausible but wrong results | | Attribute searches | Depends on catalogue quality | Inconsistent size or colour values | | Misspellings | Easy to test | Over-correction of brand names | ## Questions ### What should a Cimulate Salesforce partner be able to prove? That search improved on your own queries and products, shown by group, over a stated period against your current search. Demonstrations on a vendor catalogue and uplift figures from other retailers prove little about your shoppers. ### Which certifications matter for a Cimulate project? Salesforce Certified B2C Commerce Developer for storefront integration, Platform Data Architect for catalogue and customer data, and Agentforce Specialist if conversational features are in scope. Also ask about real search relevance work: synonyms, misspellings and zero-result queries. ### Does Cimulate replace B2C Commerce or my catalogue? No. The announcement describes an intent-aware context engine for retail search and discovery, to strengthen Agentforce Commerce. Your catalogue, storefront and pricing stay where they are, and a sandbox does not copy the storefront. ### What does a sound first Cimulate engagement look like? One storefront and category group, real queries grouped by type, success measures agreed in advance, a split test against current search for a set period, results reported by group, and a stopping rule written before the test starts. ### When can I skip a partner for Cimulate? When the catalogue is small, search logs show few failures and a merchandiser can close the gap with synonyms and rules. Also defer if you cannot yet measure search performance or the storefront is about to be replaced. --- SynconAI is a Salesforce Select Partner and an OpenAI Select Partner, delivering architect-led Salesforce and AI programmes across the United States and Australia. Related service: Salesforce consultancy (https://synconai.com/salesforce-consultancy).