Quick answer: Our 2026 Summer Release introduced the MCP connector, upgraded contact data, buyer intent pulled from documents and improved bid qualification reports. During the live session on 18 August, you sent in more questions than we could cover live. This article works through the ones we didn't get to. The most common questions were about how the MCP connector works, which AI models sit behind the platform, how your data is kept secure, and what it all costs. Here are the answers.
The MCP connector brings Stotles into the AI assistant your team already uses, so you can ask for market data in plain language and get it back inside your chat. It works with Claude, ChatGPT, Microsoft 365 Copilot, or any MCP-compatible client. You connect once, then the AI decides when to call Stotles based on what you ask and weaves the answer into its response. No developer work needed.
Today the connector is read-only. It can search across notices, buyers, suppliers, frameworks, and buyer-published documents, and it can pull buyer profiles with award history and top suppliers. It works on keywords and filters, so it does not yet apply your saved signal settings.
The MCP connector is currently in Beta and will be available soon. You can sign up to the waitlist to be the first to hear when it launches officially.
The connector is included on all paid plans and is not part of the Free plan. What you can reach through it scales with your plan:
In practice, the connector only surfaces the data your plan already gives you. On Basic, that means tenders and awards only, without spend data, frameworks, buyer or competitor profiles, documents, or decision-maker contacts. Growth and Expert open up the full set.
If you want MCP access without a full Stotles subscription, that is possible too. We handle those on a case-by-case basis, speak to a member of our team to discuss it.
Yes. The connector is read-only and only ever returns public market data. When your AI assistant pulls up Stotles data, it interacts with a specific set of endpoints that return procurement and market data, for example a search for cybersecurity tenders. That market data is the only thing that passes from Stotles back to you.
We never make one customer's inputs or outputs available to another customer, directly or indirectly. So what your team searches for stays with your team, and competitors cannot see it.
No. We do not train, fine-tune, or contribute data to any AI model, and no customer or user data is used for training. We use established external models purely for inference through their APIs. Your inputs and outputs are not used to improve any model, and we have not opted in to any provider using API data for training.
The same holds for anything you upload. Content you add to support a bid is not used to train any model and is not accessible to third parties or to Stotles staff. Your data is never sold.
We use AI to analyse public procurement and market data and turn it into summaries, insights, and recommendations. It never makes decisions on its own. The system retrieves relevant data points, then uses a model to interpret them and suggest an answer, for example flagging a likely incumbent supplier from past awards.
In short: Stotles AI produces summaries, insights, and recommended actions, and a person always reviews and acts on them. Outputs are cited back to the source data so accuracy can be checked.
There is no single model behind Stotles. We continually evaluate models from different providers and pick the best performer for each specific task, including OpenAI, Google Gemini and Anthropic. We run evaluations that test which model, prompt, and tool combination produces the best outcome for a given job, then use that. The mix changes as models improve, and we do not build or train our own models.
Bid Studio automatically extracts bid requirements from the tender documents you provide, so you are not left hunting through the bid pack for scope buried in different sections. It reads the published tender documents and pulls out the requirements to support your bid or no-bid decision.
For drafting, yes, it can help you answer standard questions. Bid Studio recommends content, phrasing, ideas, and statistics from your previous bid submissions and content library, and stotles corpus of market intelligence matched to the question being asked and your organisation's own history and context.
Yes, our contact data is GDPR compliant. We process on the basis of legitimate interest rather than opt-in, we honour all opt-outs, and we handle TPS screening so you do not have to. We work with our data partners to keep the data as accurate and fresh as possible.
Decision-maker contact data is refreshed around every month. You can find more detail on our data, including how we source and handle it, in our Trust Centre and on our data page.
It depends on the data type. Procurement notices published by government are refreshed as often as every hour. Buyer intent documents vary by document type, since annual reports are published annually while meeting minutes come out monthly. For documents, the longest refresh typically lands within around two to three months depending on the source, and the shortest is daily.
Yes. The core of this is procurement data, so you can see what your competitors have won and when those contracts expire, which lets you start conversations early. You can also see where competitors are mentioned, which matters when buyers procure through resellers or when there is a prime and subcontractor relationship. That shows up across supplier searches and supplier profiles.
Invoice spend data adds another view, showing who your competitors are transacting with. This all relies on the work we do to clean, enrich, and deduplicate organisations, so you get one clear view of any organisation, whether it is a buyer, supplier, competitor, or partner. It works on a push or pull basis. Get notified when a competitor is mentioned, or pull the information yourself from their profile or through the MCP connector.
Both. We have an off-the-shelf SaaS platform that you can customise and personalise, with a free version and paid plans. We also build custom solutions, such as data pipelines that connect with CRMs or other systems, and relevance and scoring models that match market data to your own organisational data.
We are SOC 2 Type 2 certified, fully audited across the company, and we run annual penetration testing through a CREST-approved provider.
Although SOC 2 originated with the AICPA, the Trust Services Criteria aren't US-specific.
We regularly review our capacity to expand our certifications in line with industry standards and customer requirements. Thus far, SOC 2 has been the most sought, but if other certificates are essential to working with you, please do reach out and let us know.
Our full SOC 2 Type 2 report is available in our Trust Centre, and we're happy to walk your security team through it.
There is a free plan, so you can start at no cost. However, you will need to sign up with a business email address rather than a personal one. Sign up for free here.
Paid plans build from there: Basic runs month to month with no annual commitment, and Growth and Expert are built for larger teams. Paid pricing is typically a platform fee plus per-user pricing.
There is no revenue threshold for using Stotles. The right plan depends on how you plan to use it rather than the size of your business, and starting on the free plan is a good way to see the value before you commit. Some features shown at the Summer Release, such as the MCP connector and decision-maker contacts, sit on the paid plans.
Watch the full recording on-demand here.
