--- title: How to Stay in Control When Connecting AI to Your Systems description: Connecting AI to your CRM, HR or support tools? Learn the real risks, the integration mistakes to avoid, and how to stay in control as AI use grows. image: https://resources.trustkeith.co/hubfs/Resources%20Card%20template%2c%20Guides%20(54).png --- [![642bcbad7735490755580d8d\_trust-keith-logo](https://resources.trustkeith.co/hubfs/642bcbad7735490755580d8d_trust-keith-logo.svg "642bcbad7735490755580d8d_trust-keith-logo")](https://www.trustkeith.co/) - Product - [Privacy OS](https://www.trustkeith.co/product/privacy-os) - [Dedicated DPO](https://www.trustkeith.co/product/dedicated-dpo) - [Automated Data Discovery](https://www.trustkeith.co/product/automated-data-discovery) - [Intelligent Workflows](https://www.trustkeith.co/product/intelligent-workflows) - [Staff Training](https://www.trustkeith.co/product/staff-training) - [Global Risks & Controls](https://www.trustkeith.co/product/global-risks-controls) - [Proportional Policies](https://www.trustkeith.co/product/proportional-policies) - [Monitoring & Reporting](https://www.trustkeith.co/product/monitoring-reporting) - Solutions - [By Industry](https://www.trustkeith.co/solutions/industry) - [By Team](https://www.trustkeith.co/solutions/team) - [By Use Case](https://www.trustkeith.co/solutions/use-case) - [By Privacy Regulation](https://www.trustkeith.co/solutions/regulations) - [How it works](https://www.trustkeith.co/how-it-works) - [Customers](https://www.trustkeith.co/customers) - [Pricing](https://www.trustkeith.co/pricing) - [Resources](https://resources.trustkeith.co) - [Company](https://www.trustkeith.co/company) [![Talk to an expert](https://hubspot-no-cache-eu1-prod.s3.amazonaws.com/cta/default/24879685/859c2149-73c5-4ae5-a465-6c639d24614f.png)](https://hubspot-cta-redirect-eu1-prod.s3.amazonaws.com/cta/redirect/24879685/859c2149-73c5-4ae5-a465-6c639d24614f) [![Trust Keith](https://resources.trustkeith.co/hs-fs/hubfs/raw_assets/public/Trustkeith_March2023/images/trust_keith.png?width=500&name=trust_keith.png "Trust Keith")](https://www.trustkeith.co/) - Product - [Privacy OS](https://www.trustkeith.co/product/privacy-os) - [Dedicated DPO](https://www.trustkeith.co/product/dedicated-dpo) - [Automated Data Discovery](https://www.trustkeith.co/product/automated-data-discovery) - [Intelligent Workflows](https://www.trustkeith.co/product/intelligent-workflows) - [Staff Training](https://www.trustkeith.co/product/staff-training) - [Global Risks & Controls](https://www.trustkeith.co/product/global-risks-controls) - [Proportional Policies](https://www.trustkeith.co/product/proportional-policies) - [Monitoring & Reporting](https://www.trustkeith.co/product/monitoring-reporting) - Solutions - [By Industry](https://www.trustkeith.co/solutions/industry) - [By Team](https://www.trustkeith.co/solutions/team) - [By Use Case](https://www.trustkeith.co/solutions/use-case) - [By Privacy Regulation](https://www.trustkeith.co/solutions/regulations) - [How it works](https://www.trustkeith.co/how-it-works) - [Customers](https://www.trustkeith.co/customers) - [Pricing](https://www.trustkeith.co/pricing) - [Resources](https://resources.trustkeith.co) - [Company](https://www.trustkeith.co/company) [![Talk to an expert](https://hubspot-no-cache-eu1-prod.s3.amazonaws.com/cta/default/24879685/859c2149-73c5-4ae5-a465-6c639d24614f.png)](https://hubspot-cta-redirect-eu1-prod.s3.amazonaws.com/cta/redirect/24879685/859c2149-73c5-4ae5-a465-6c639d24614f) [![Talk to an expert](https://hubspot-no-cache-eu1-prod.s3.amazonaws.com/cta/default/24879685/859c2149-73c5-4ae5-a465-6c639d24614f.png)](https://hubspot-cta-redirect-eu1-prod.s3.amazonaws.com/cta/redirect/24879685/859c2149-73c5-4ae5-a465-6c639d24614f) ![Menu Icon](https://resources.trustkeith.co/hubfs/raw_assets/public/Trustkeith_March2023/images/icon-menu.svg) ![Menu close](https://resources.trustkeith.co/hubfs/raw_assets/public/Trustkeith_March2023/images/icon-x.svg) ![](https://resources.trustkeith.co/hubfs/Resources%20Card%20template%2c%20Guides%20(54).png) # Connecting AI to Your Systems: How to Stay in Control Published by [Annabelle Ilsley](https://resources.trustkeith.co/author/annabelle-ilsley) on Oct 2, 2026, 11:34:47 AM Plugging an AI agent into your CRM, HR system or support inbox is where the big productivity gains are, but's also where an agent gets access to a massive amount of the data you hold... Trust Keith recently teamed up with [StackOne](https://www.stackone.com/), the AI integration platform, to dig into how to intergrate AI into your business systems securely and compliantly, and how to stay in control as usage grows. We've pulled the key takeaways into this guide, but you can [watch the full recording here \>\>](https://www.youtube.com/watch?v=s5j91jXl4nE) [![Free AI policy template from Trust Keith](https://resources.trustkeith.co/hs-fs/hubfs/Blog%20assets/ai-policy-template.png?width=200&height=222&name=ai-policy-template.png)](https://resources.trustkeith.co/free-download-ai-policy-template) Free Download | AI Policy Template Good guardrails only work once they're made official. Your team can't stick to rules they've never seen. Get your approved tools, data rules, agent access and review dates written down in one place with our free AI policy template, ready to make your own. [Download the free template](https://resources.trustkeith.co/free-download-ai-policy-template) --- ### What we'll cover: - [Where AI adoption actually stands right now](https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems#where-ai-adoption-stands) - [Why businesses still hesitate to connect AI to their systems](https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems#why-businesses-hesitate) - [What's at risk when AI can reach your CRM, HR or support tools](https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems#whats-at-risk) - [The 4 integration mistakes we see most often](https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems#integration-mistakes) - [How to stay in control as AI use grows](https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems#stay-in-control) - [What "good" looks like](https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems#what-good-looks-like) - [Common questions about connecting AI to business systems](https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems#common-questions) - [Key takeaways](https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems#key-takeaways) ---   ## **A quick intro to our experts** **Nick Harkin, AI and Data Protection Expert @ Trust Keith** ![Icons (98)](https://resources.trustkeith.co/hs-fs/hubfs/Icons%20(98).png?width=75&height=75&name=Icons%20(98).png)Nick is a DPO at Trust Keith with extensive experience across data protection, cyber risk, and AI governance. He works with businesses to help them navigate the practical realities of using AI safely, securely, and compliantly. **Guillaume Lebedel, Co-Founder & CTO @ StackOne** ![Icons - 2026-09-01T163612.601](https://resources.trustkeith.co/hs-fs/hubfs/Icons%20-%202026-09-01T163612.601.png?width=75&height=75&name=Icons%20-%202026-09-01T163612.601.png)Co-founder and CTO of StackOne, the Tools Gateway for AI agents. StackOne connects agents to enterprise business systems and governs what each one is allowed to do, so companies can expand AI usage without losing control of their data. ---   ## **Where AI adoption actually stands right now** Guillaume sees two very different groups. The first is tech-first businesses. Engineering drives a lot of the strategy, AI started there, and it's been spreading across the rest of the company since last year. They tend to move fast, and sometimes that means safety and controls come second. The second is more traditional or larger businesses. They started with "read" use cases like enterprise search and chat-style questions about internal information. Now they're asking how to let AI safely *take action* as well. And our experts have really seen this evolve. A year ago, some businesses still said "we just won't use AI." But that isn't realistic any more. > "People in your company are using AI. If you don't have an enterprise AI solution, they're typing it into their phones instead, and you don't have any oversight of it that way." **Nick Harkin**, AI and Data Protection Expert, Trust Keith The type of use has changed too. A year ago, most teams were pasting prompts into ChatGPT and copying the output into emails. Now they want their CRM talking to their HR system, with an agent doing the work in between. > "The question's moved on now from what someone's been pasting into the chat to what this thing can reach, and what actions it can do on its own without us even knowing about it." **Nick Harkin**, AI and Data Protection Expert, Trust Keith ---   ## **Why are businesses still hesitating to connect AI to their systems?** Most of the hesitation comes down to four things. **Visibility.** If an AI tool runs with one person's login, can it see their whole mailbox? The whole CRM? Every lead? Often nobody's checked. **The vendor.** Is the vendor training on your data? Who are its sub-processors? Where is the data processed, and does any of that match what your privacy notice and data protection agreements say? **Accountability.** When an agent gets something wrong, who owns it? Agents can't be held responsible the way a person can. Someone in the business has to be accountable. **Liability.** Some actions are now taken by AI on a user's behalf, and that user doesn't always understand the consequences. That leaves real questions about liability, for the business and the employee, that need answering before AI use scales. For fast-moving scale-ups, these questions often get skipped until an enterprise customer's security questionnaire or an investor's due diligence list lands. Then it turns into a scramble through old emails and meeting notes to work out what was actually done. (If that's on the horizon, [these are the privacy questions investors ask most](https://resources.trustkeith.co/5-privacy-questions-investors-ask-and-how-to-answer-them).) [![trust keith webinar](https://resources.trustkeith.co/hs-fs/hubfs/Blog%20Banners%20(21).png?width=1024&height=200&name=Blog%20Banners%20(21).png)](https://resources.trustkeith.co/office-hours) ---   ## **What's at risk when AI can reach your CRM, HR or support tools?** ### **AI taking the wrong action** A low-impact version is an email going out with content nobody approved. A high-impact version is refusing someone a service, a loan or a job because of something AI decided. And you might not even know it happened. AI generates a huge volume of logs, and most systems give you very little ability to review them at scale. ### **Customer data ending up in the wrong conversation** Picture a support chatbot connected to the whole company CRM. A customer asks a question, and the bot helpfully explains how a similar case was solved for "Joe Smith at 123 Fake Street." That's one customer's personal data handed to another, and possibly a commitment the business never intended to make. Poorly built support agents have done exactly this, exposing internal pricing, sales strategy and other customers' data. > "As you make those support agents more and more powerful, connected to every part of your system and CRM, it is very important you still tightly control what's going into any particular conversation." **Guillaume Lebedel**, Co-Founder & CTO, StackOne ### **Personal data spreading where you can't find it** If an agent connects to systems you didn't expect, or enriches profiles by pulling data from elsewhere, you can lose track of where someone's data actually lives. That makes it very hard to respond properly to a [data subject access request](https://resources.trustkeith.co/how-to-respond-to-a-dsar-request-in-7-simple-steps), and very hard to defend your [records of processing](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/accountability-and-governance/documentation/). ### **Prompt injection** [Prompt injection](https://genai.owasp.org/llmrisk/llm01-prompt-injection/) is when someone hides instructions in content your AI will read, like an email, a document or a meeting transcript, so the AI treats them as genuine. > "Someone sending you an email with some instruction that your agent can potentially take as gospel, such as sharing data you shouldn't be sharing. It's a bit like the old-school SQL injection, but prompt injection is much harder to defend against. And the more you connect things to AI, the more you open those attack vectors." **Guillaume Lebedel**, Co-Founder & CTO, StackOne At the serious end, that's an attempt to pull data out of the business. At the milder end, it's a candidate hiding text in their CV to get past an AI screening tool. ### **Too much output to properly review** Not every risk is a compliance one. AI now produces a huge volume of code, documents and emails. When review can't keep up, quality slips and teams lose track of what's actually going out the door. ### **A couple of real-life examples** **AI agreeing to things it shouldn't.** There have been plenty of reports of AI chatbots being haggled into deals or prices the business never meant to offer. Even when the business doesn't have to honour them, having to walk them back publicly does real damage to its reputation. **The note-taker that stayed on.** A candidate recently wrote about leaving a video interview while the hiring team stayed in the room. The AI note-taker kept recording the team's discussion about him and emailed the summary to everyone afterwards, including him. He called it the most useful feedback he's ever had. The company probably felt differently. ---   ## **The 4 integration mistakes we see most often** ### **1. Connecting AI with an admin or founder login** > "The agent can inherit everything that a person can see. So if you've linked your sales assistant to your founder's mailbox because that's where the sales requests are coming in, that's also linking up to your HR and your finance threads." **Nick Harkin**, AI and Data Protection Expert, Trust Keith Give each agent its own dedicated account with a narrow scope. ### **2. "It's just a pilot"** Pilots on live customer data have a habit of turning into production without anyone noticing. Treat a pilot as real processing from day one. That means guardrails, and an [AI DPIA](https://resources.trustkeith.co/a-guide-to-ai-dpias-when-to-run-them-and-how-to-get-them-right) where the risk calls for one. ### **3. Using consumer tiers instead of business tiers** Free and personal versions of AI tools can use your inputs to train their models, and they don't come with a proper data processing agreement. Get onto an enterprise or business tier before anything touches customer or employee data. ### **4. Ignoring free-text fields** Support tickets, notes and comment boxes are where personal data hides. One ticket could contain health details, payment information or a complaint about a named colleague. Strip out or mask that data before it reaches the model. It's also worth deciding what counts as sensitive *for your business*, beyond personal data, and which teams should be able to put it into an agent at all. And keep an approved list of connectors and add-ons, rather than letting people install them from anywhere. > "If it's a Wild West where people can install any random MCPs from the internet, it is going to lead to pain at some point." **Guillaume Lebedel**, Co-Founder & CTO, StackOne The good news is that the controls you already use to stop malware being installed, like device management, work here too. [![trust keith newsletter](https://resources.trustkeith.co/hs-fs/hubfs/Blog%20Banners%20(22).png?width=1024&height=200&name=Blog%20Banners%20(22).png)](https://www.trustkeith.co/resources) ---   ## **How to stay in control as AI use grows** ### **Enable it, don't gatekeep it** Block AI outright and people go around you with personal accounts. That's shadow AI, and you have no oversight of it. > "If you hinder it or gatekeep too much, they're going to find workarounds and use their own ChatGPT subscription, because the technology is just so good." **Guillaume Lebedel**, Co-Founder & CTO, StackOne Be proactive instead. Tell people which tools are approved and how to connect them safely, and listen to how teams actually want to use AI. (Trust Keith covers shadow AI in more depth in [this guide to adopting AI without losing control](https://resources.trustkeith.co/how-to-adopt-ai-without-losing-control-of-privacy-security-and-operational-risk).) ### **Make the right thing the easy thing** > "The trick to getting people to use AI responsibly is you make it easier to be compliant than not compliant. As soon as you make it easy for people, they will do the right thing." **Nick Harkin**, AI and Data Protection Expert, Trust Keith In practice, that means four things: - A short list of approved tools - A one-page guide to what can go into which tool - A longer guide for the detail - One named person (or one mailbox) to ask ### **Use a traffic light system** For the one-pager, Nick recommends sorting data into three categories: | | What it covers | What to do | | --- | --- | --- | | 🟢 Green | No personal data | Go ahead | | 🟠 Amber | Contains personal data | Check first: why am I putting this in, where did it come from, what am I doing with it? | | 🔴 Red | Special category data, or anything that makes a decision with a legal, financial or significant impact on a person | Not allowed by default. Needs a DPIA and sign-off from the right person | It's worth having a red list for *tools* too: tools nobody should use because they haven't been approved or have known privacy or security concerns. It also gives new joiners a clear "this is how we do AI here" from day one. ### **Plug AI into the processes you already have** > "Fundamentally, AI isn't any different to any other tool or process that you implement. It doesn't need its own processes. It should hook into your Article 30s and your procurement process anyway." **Nick Harkin**, AI and Data Protection Expert, Trust Keith That means building it into what already exists: - Your procurement and new-tool onboarding - Your Article 30 records of processing - Your monthly ops meetings - Your new starter training ### **Write your guardrails first** > "Write your guardrails first. Don't try to backwards-engineer your guardrails once things start going wrong. Businesses who write the guardrails first say yes to more use cases, because the answer to 'is this okay to do?' already exists." **Nick Harkin**, AI and Data Protection Expert, Trust Keith Requests then land in "yes" or "yes, if we do this" rather than defaulting to "no, let us check." ### **Review on triggers, not just the calendar** AI tools change weekly, not yearly. For anything touching personal or commercially sensitive data, Nick recommends a review at least every six months. Also do a light review whenever: - A new data source is connected - The purpose changes - The vendor adds a new sub-processor And keep evidence of every review. > "If you can't show you've done it, you haven't done it as far as the regulator is concerned." **Nick Harkin**, AI and Data Protection Expert, Trust Keith The [ICO's guidance on AI and data protection](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/artificial-intelligence/guidance-on-ai-and-data-protection/) is a useful reference point. ---   ## **What does "good" look like?** > "Great looks like when your team comes up with use cases independently and is able to do them safely." **Guillaume Lebedel**, Co-Founder & CTO, StackOne The other signal is your business numbers: revenue, customer feedback and efficiency, not just the absence of breaches. Nick's measure is the questions people start asking. For example: "What if we combined our event outreach data with the customers who came from it, and tracked what they cost to service and whether they renewed?" When people ask that *and* how to do it safely and ethically, you're in a great place. ---   ## **Common questions about connecting AI to business systems** ### **Can you control what data goes into Claude or ChatGPT when you connect it to Slack, Google Drive or Gmail?** In Guillaume's view, not to the level most businesses need with the native connectors alone. Their governance and policy controls are limited. His recommendation is to add a connectivity or agent platform on top that lets you set your own access controls and data policies. ### **How do regulated financial services businesses judge which AI integrations are safe?** > "There is no list of safe integrations. You need to be judging each one on its connector, its vendor and what it does, and keep that evidence." **Nick Harkin**, AI and Data Protection Expert, Trust Keith In practice, that means looking at: - What data it can access - Who the vendor is, and what the contract says - What the agent can actually *do* - Whether it only optimises internal processes, or affects consumer outcomes like servicing, complaints or credit ### **How do you stop people switching their brains off and approving whatever AI suggests?** You can't guarantee anyone stops to think. You can design the moment so that stopping is easy and skipping it is the visible, harder option. Nick suggests prompts at the right point, like "this field may contain customer data," with a named contact for questions. Don't expect people to have memorised the policy. StackOne takes it a step further. When personal data is detected, a human has to actively approve before the agent continues, which reminds them that the output is their responsibility. > "Everything that the AI outputs on their behalf is on them, and they should be able to stand behind every single sentence in a Notion page they've shared or an email they've sent a customer." **Guillaume Lebedel**, Co-Founder & CTO, StackOne ---   ## **Key takeaways** > "Treat these agents as a super keen new employee at quite a junior level. They're really keen, but you need to make sure that you're narrowing their focus, so they can access what they need to, and separating read versus write access versus action." **Nick Harkin**, AI and Data Protection Expert, Trust Keith Nick's four practical steps: 1. **Know your data.** 2. **Know the vendor and its terms.** 3. **Limit your agent's access.** 4. **Keep a light record** for regulators, investors, customers and partners, so you're never rebuilding it under pressure. > "It's never been more exciting. Despite all the risks, it is an incredible time, and I think we need to learn to enjoy ourselves throughout that journey. Doing it safely, of course." **Guillaume Lebedel**, Co-Founder & CTO, StackOne [![trust keith free risk assessment](https://resources.trustkeith.co/hs-fs/hubfs/Blog%20Banners%20(30).png?width=1024&height=200&name=Blog%20Banners%20(30).png)](https://resources.trustkeith.co/data-privacy-risk-assessment) ---   ## **Looking for a bit more support with your AI governance?** At Trust Keith, we help businesses stay on top of their compliance, from AI governance and beyond. If you'd like to find out how we could support your team, or you'd just like a chat about where you're at with AI, get in touch. [![trust keith contact us](https://resources.trustkeith.co/hs-fs/hubfs/CTAs%20(13).png?width=248&height=62&name=CTAs%20(13).png)](https://www.trustkeith.co/talk-to-us) [![Trust Keith](https://resources.trustkeith.co/hs-fs/hubfs/blog%20cta%20(7).png?width=3750&height=1250&name=blog%20cta%20(7).png)](https://www.trustkeith.co/)   ## Related Articles ## [Hiring With AI: How to Keep Recruitment Fast, Fair and Compliant](https://resources.trustkeith.co/how-to-stay-compliant-when-hiring-with-ai-trust-keith) AI has quickly become part of most hiring processes. It’s summarising interviews, drafting job... [Read More](https://resources.trustkeith.co/how-to-stay-compliant-when-hiring-with-ai-trust-keith) ## [How to Adopt AI Without Losing Control of Privacy, Security, and Operational Risk](https://resources.trustkeith.co/how-to-adopt-ai-without-losing-control-of-privacy-security-and-operational-risk) AI isn't coming, it's already here. And in most businesses, it's running well ahead of any kind of... [Read More](https://resources.trustkeith.co/how-to-adopt-ai-without-losing-control-of-privacy-security-and-operational-risk) ## [Should you trust ChatGPT with your data?](https://resources.trustkeith.co/should-you-trust-chatgpt-with-your-data) In the era of operational efficiency, Chat GPT is fast becoming a staple in the tech startup's tech... [Read More](https://resources.trustkeith.co/should-you-trust-chatgpt-with-your-data) ## Curious how Trust Keith can help take data compliance off your plate? [![Talk to an expert](https://hubspot-no-cache-eu1-prod.s3.amazonaws.com/cta/default/24879685/859c2149-73c5-4ae5-a465-6c639d24614f.png)](https://hubspot-cta-redirect-eu1-prod.s3.amazonaws.com/cta/redirect/24879685/859c2149-73c5-4ae5-a465-6c639d24614f) [![Trust Keith](https://resources.trustkeith.co/hs-fs/hubfs/raw_assets/public/Trustkeith_March2023/images/trust_keith01.png?width=268&name=trust_keith01.png "Trust Keith")](https://www.trustkeith.co/) [![Talk to an expert](https://hubspot-no-cache-eu1-prod.s3.amazonaws.com/cta/default/24879685/859c2149-73c5-4ae5-a465-6c639d24614f.png)](https://hubspot-cta-redirect-eu1-prod.s3.amazonaws.com/cta/redirect/24879685/859c2149-73c5-4ae5-a465-6c639d24614f) - Features - [Comply with data regulations](https://www.trustkeith.co/product/comply-with-regulations) - [Discover your data risks](https://www.trustkeith.co/product/discover-data-risks) - [Expert Advice](https://www.trustkeith.co/product/expert-advice) - [Govern your data protection programme](https://www.trustkeith.co/product/govern-data-protection) - [Manage breaches and incidents](https://www.trustkeith.co/product/manage-breaches-and-incidents) - [Train your staff](https://www.trustkeith.co/product/train-your-staff) - Product - [How it works](https://www.trustkeith.co/how-it-works) - [Marketplace](https://www.trustkeith.co/marketplace) - [Pricing](https://www.trustkeith.co/pricing) - [Customers](https://www.trustkeith.co/customers) - Trust Keith - [Company](https://www.trustkeith.co/company) - [Talk to us](https://www.trustkeith.co/talk-to-us) - [LinkedIn](https://www.linkedin.com/company/trustkeith/) - [Privacy policy](https://trustkeith.trustkeith.co/) © Trust Keith 2026. Trust Keith Ltd is a company registered in England and Wales number 12283797. ```json { "@context" : "https://schema.org", "@id" : "https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems#blogposting", "@type" : "BlogPosting", "author" : { "@type" : "Person", "name" : "Annabelle Ilsley" }, "dateModified" : "2026-10-02T10:40:46+0000", "datePublished" : "2026-10-02T10:34:47+0000", "description" : "Connecting AI to your CRM, HR or support tools? Learn the real risks, the integration mistakes to avoid, and how to stay in control as AI use grows.", "headline" : "Connecting AI to Your Systems: How to Stay in Control", "inLanguage" : "en-GB", "isPartOf" : { "@id" : "https://resources.trustkeith.co/#blog", "@type" : "Blog", "name" : "Trust Keith Resources Blog", "publisher" : { "@id" : "https://www.trustkeith.co/#organization" } }, "mainEntityOfPage" : { "@id" : "https://resources.trustkeith.co/how-to-stay-in-control-when-connecting-ai-to-your-systems", "@type" : "WebPage" }, "publisher" : { "@id" : "https://www.trustkeith.co/#organization", "@type" : "Organization", "logo" : { "@type" : "ImageObject", "url" : "https://cdn.prod.website-files.com/5f462e8160aa877216f7d4b7/642bcbad7735490755580d8d_trust-keith-logo.svg" }, "name" : "Trust Keith", "url" : "https://www.trustkeith.co/" } } ```