The alternatives to make use of AI in workflow automation are many and various, however one of many easiest methods to make use of AI to save lots of time and improve your group’s security posture is by constructing an automatic SMS evaluation service.
Workflow automation platform Tines gives a great instance of find out how to do it. The seller not too long ago launched their first native AI options, and security groups have already began sharing the AI-enhanced workflows they’ve constructed utilizing the platform.
Tines’ library of pre-built workflows contains AI-enhanced pre-built workflows for normalizing alerts, creating circumstances, and figuring out which phishing emails require escalations.
Let’s take a more in-depth take a look at their SMS evaluation workflow, which, like all of their pre-built workflows, is free to entry and import, and can be utilized with a free Neighborhood Version account.
Right here, we’ll share an summary of the workflow, and a step-by-step information for getting it up and operating.
The issue – SMS rip-off messages focused at workers
Staff typically obtain rip-off SMS messages designed to steal login credentials or set up malware on gadgets. These messages can result in compromised accounts, unauthorized entry to delicate data, and potential breaches that might trigger vital hurt to the group’s operations and fame.
Responding to every report manually may be time-consuming and error-prone, making it troublesome for security groups to maintain up with the amount of threats.
The answer – an automatic SMS evaluation service
 |
| An internet web page with a type inviting workers to submit the message for evaluation |
Utilizing workflow automation, security groups can create an SMS evaluation service that any worker throughout the group can use.
It operates as a easy self-service mannequin. When workers obtain a suspicious message, they go to the net web page and submit a screenshot of the message or a URL utilizing the supplied type. This way kicks off the workflow they usually get a response inside a few seconds. The response features a clear and detailed evaluation of the message and a few really useful subsequent steps.
This is an instance of the picture they may submit:
This is an instance of the outcome they may obtain:
 |
| The automated evaluation obtained by the worker |
The textual content reads:
Hello! Thanks for reporting the message. I’ve analyzed the content material and have a number of observations:
The message claims that your Amazon account has been locked resulting from a number of failed login makes an attempt. It is a widespread tactic utilized in phishing scams, the place attackers attempt to trick you into clicking on a malicious hyperlink to “get better” your account.
The message has a way of urgency, which is one other purple flag for phishing makes an attempt. Respectable firms sometimes do not demand rapid motion on this method.
The hyperlink supplied within the message (hxxp://s953909557/servweb) seems to be suspicious. I’ve defanged the hyperlink, however I might strongly advise in opposition to clicking on it, because it might result in a malicious web site designed to steal your login credentials or set up malware in your machine.
Moreover, the message doesn’t seem like from any of our senior executives [executive names here]. This might be an try at CEO fraud, the place somebody is impersonating a senior chief to request delicate data or carry out unauthorized actions.
In abstract, this message reveals a number of traits of a typical phishing rip-off and needs to be handled with warning. I like to recommend that you don’t click on on the supplied hyperlink and as a substitute contact Amazon immediately by means of their official web site or customer support channels to confirm the standing of your account.
Please let me know when you’ve got some other questions or considerations. I am right here to assist make sure the security of our group.
Listed here are among the key advantages of launching a service like this:
- It promotes a tradition of cybersecurity by making it simple for workers to be vigilant and reinforces security-conscious decision-making
- It reduces handbook, repetitive work for the security staff
- It improves the velocity and accuracy of risk detection
 |
| The pre-built workflow because it seems within the Tines library |
Workflow overview
This workflow makes use of Tines Pages to create an automatic SMS evaluation service that anybody within the group can use.
Instruments used:
- Tines – a workflow automation and orchestration platform that is well-liked with security groups. It is potential to make use of the free Neighborhood Version of Tines to construct and run this workflow if you do not have a paid account. AI have to be enabled in your tenant. Observe that AI motion utilization is predicated on a credit system however there’s a free credit score allowance on all accounts.
- OCR – a free instrument that parses photographs and multi-page PDF paperwork and returns the extracted textual content ends in a JSON format. Professional plans with bigger utilization allowances are additionally accessible.
The workflow is kicked off by a submission on a Tines web page, which features a type the place customers can submit a picture of an SMS message or a related URL.
The workflow then makes use of OCR to extract the textual content. If the picture exceeds the file measurement restrict, it is resized utilizing the Automated Mode transformation motion, which invokes a small piece of Python code that has been generated by AI in Tines.
The workflow additionally fetches the picture if the supplied enter is a URL. If the picture was uploaded, it renames the picture to match the required format.
As soon as the textual content is extracted, it is then despatched to the AI motion for evaluation. The AI immediate asks the language mannequin to investigate it for potential rip-off indicators and defang any hyperlinks.
This is the AI immediate the Tines staff used to create the workflow:
You’re a digital Safety Analyst analyzing a suspicious SMS reported to you. The screenshot of the SMS has been OCR’d by you.
Reply to the person submitting the SMS with the evaluation. You ought to be analyzing it for tone and for widespread scams like phishing, romance scams, faux bill, faux tickets, and dozens of others.
As that is an inside instrument, the first fear is CEO Fraud the place somebody is likely to be impersonating a senior govt. The Senior Executives on this firm are [provide executive names and titles here].
In case you are together with any hyperlinks within the response that could be suspicious, be sure you defang them.
Start with:
“Hello! Thanks for reporting the Message…”
The AI motion types a response to the person together with evaluation – whether or not or not the message seems to be malicious – and really useful subsequent steps – do not click on the hyperlink, and so forth.
If, for some purpose, the evaluation fails, the person will obtain a message prompting them to attempt once more or contact the security staff.
Configuring the workflow – step-by-step information
 |
| The Tines Neighborhood Version sign-up type |
1. Log into Tines or create a brand new account.
2. Guarantee AI is enabled in your tenant. For this, you might want to be the tenant proprietor. Choose the account settings drop-down within the high left of your display screen, and verify the field to show AI on.
 |
| The OCR House sign-up type |
 |
| Including a brand new credential in Tines |
3. Create your OCR credential. Arrange an OCR API account if you do not have one already and get the API key to your account. From the credentials web page, choose New credential. You’ll then be prompted to decide on the credential kind (on this case, Textual content) and full the required fields. Identify the credential “ocr_space” to routinely join the credential to the workflow.
 |
| Importing a narrative from the library to your tenant |
4. Navigate to the pre-built workflow within the library.
 |
| The workflow on Tines’ drag-and-drop canvas |
5. Choose import. This could take you straight to your new pre-built workflow.
 |
| Modifying the Tines web page |
 |
| Customizing the AI immediate |
6. Configure your actions. For instance, you might prefer to edit the format of the Tines web page that kicks off the workflow, and customise the AI immediate with the names of executives at your organization.
7. Check the workflow. Submit a picture by way of the shape to check your workflow.
8. Publish your workflow and share the Web page URL together with your desired customers.
Constructing in different automation platforms
You possibly can use one other no-code automation platform to construct an identical service, though it is value noting that among the options on this workflow are distinctive to Tines:
- Pages: This workflow is kicked off by a submission to a type on an internet web page, and the output is delivered by way of the identical internet web page. That is constructed utilizing Tines’ Pages function.
- Various: Obtain data and ship outcomes by way of e-mail.
- The AI Motion: Tines’ AI motion is exclusive in that it permits customers to immediately entry and use a language mannequin at any level of their workflow whereas providing sturdy security guardrails. There isn’t any coaching, logging, inspecting, or storing of knowledge that enters or exits the language mannequin.
- Various: Connect with an exterior LLM like ChatGPT for evaluation, however you’ll want to consider the security and privateness options of no matter mannequin you intend to make use of if delicate information might be handed by means of.
- Occasion Rework in Automated Mode: This function makes use of build-time AI to compose Python code primarily based on the steering and the enter the builder gives. When you save your modifications, the code is locked in place. Which means that when the motion runs, solely the code executes, and no AI is concerned.
- Various: Write Python code manually to rework your information.
If you would like to discover AI in Tines for your self or check out this workflow, you may join a free account together with AI performance.
