DISSENTING OPINION BY
KLEIN, J.:¶ 1 I believe that the loss in this case is not covered by the Commercial General Liability (CGL) policy that Upright Materials Handling, Inc. purchased. Instead, it would be covered under a Products Hazard and Completed Operations policy, which Upright did not purchase until after this accident. I also believe that The West American Insurance Company used exactly the language of the “products-completed operations hazard” exclusion that this Court told insurance companies they should use to avoid covering failure to warn. See Harford Mutual Insurance Co.16 v. Moorhead, 396 Pa.Super. 234, 578 A.2d 492 (1990). If there is no coverage, there is no “bad faith” in refusing to make payment. Therefore, I must dissent.
¶ 2 In this case, the insurance policy in question specifically excludes failure to warn claims from coverage. The definitions of ‘Tour product” and ‘Tour work” includes: “The providing of or failure to provide warnings or instructions.” See policy definitions. The majority states:
Regarding West American’s specific averment that the cases cited by the trial court are distinguishable because the wording of the exclusion in this case contained the phrase “[t]he providing of or failure to provide warnings or instruction,” we disagree. While the insurance policy at issue refers to “failure to warn,” the warnings must still be complete in order to fall within the ambit of the exclusion. Under the policy, the operation is not completed until the work has been put to its intended use at the site by someone other than another contractor on the same project. Here the installation of the backup alarm’s wiring was not completed at the time of the installation because of the negligent failure to warn, and, therefore, the operation was not complete.
Majority at 89.
¶ 3 If the circular logic of the majority and trial court is true, then a failure to *97warn can never be excluded. That theory is that if there is no warning: (a) there is negligence so the “products” exclusion does not apply; and (b) the product is not completed so the “completed operations” exclusion does not apply. If the majority and the trial court are correct, then even the clear and unambiguous language of the instant policy, following a specific instruction from our Court, is still unenforceable. That does not make sense. Here, in addition to the “products” exclusion and the “completed operations” exclusion, there is a specific exclusion for failure to warn. There is no reason, public policy or otherwise, that an insurance company should not be able to offer a general liability insurance policy at lower rates without having to charge more to cover possible lawsuits for failure to warn.
Types of exclusions
¶ 4 There are two types of exclusion relevant to this lawsuit: products hazard, addressing strict liability claims, and completed operations, addressing negligence in the installation process.
¶ 5 A CGL policy is just that, an insurance policy that provides coverage for general liability. In other words, the CGL is designed to provide coverage if the insured drops a hammer on another person’s foot during the installation process or if a person is injured on the insured’s premises, but it is not designed to provide coverage if the product itself is defective or if one of the insured’s employees installs the product incorrectly. Commonly, such a policy contains certain exclusions for occurrences such as are involved in products liability situations or negligence involved in the installation of the product.17
¶ 6 The exclusions referred to above are known as “products hazards” and “completed operations.” A products hazard exclusion operates to exclude strict liability claims from coverage. A leading case in products hazard litigation is Harford, supra.18 In that case, the trial court drew a distinction between a strict liability claim for failure to warn and a general negligence claim of failure to warn. Because the products hazard language was recognized as excluding only strict liability claims, the trial court found, and our Court affirmed, that a claim of failure to warn based on negligence was not excluded. In Harford, our Court specifically instructed insurers that if they wanted to exclude such negligent failure to warn claims from coverage, they had to do so explicitly.
¶ 7 A completed operations exclusion operates to exclude negligence in the installation of the product. A leading case in completed operations litigation is Friestad v. Travelers Indemnity Co., 260 Pa.Super. 178, 393 A.2d 1212 (1978). In Friestad, the negligent installation of a furnace led to a fire that destroyed Freistad’s property. Our Court stated that a completed operations exclusion would properly exclude such a claim from coverage.
¶ 8 Keystone Spray Equipment, Inc. v. Regis Ins. Co., 767 A.2d 572 (Pa.Super.2001), presents a twist in the case law. In Keystone, our Court held that under a completed operations exclusion, the instal*98lation could not be “completed” if proper warnings had not been given. Thus, the completed operations exclusion, which generally guards against negligence in the installation process, could not apply because the operation had never been completed. The stated rational used in Keystone was that the danger from a failure to warn could not be discovered until an injury had occurred. Thus, it would be unfair to allow an insurer to disclaim.
¶ 9 The insurance policy in question contains a hybrid “products hazard/completed operations” exclusion, encompassing both types of exclusion. In addition, it contains the language Harford directed insurers to use if they wanted to exclude negligent failure to warn claims from coverage. In fact, the policy uses the Harford language so that it applies to products hazard and completed operations.
¶ 10 If the majority and the trial court are correct in the interpretation of the policy and application.of prior case law, as noted the result would be that a failure to warn claim can never be excluded from coverage in a CGL policy. Neither the products exclusion nor the completed operations exclusion would work to avoid coverage, and specific language excluding coverage for failure to warn would be ignored.
Application of the law
¶ 11 One of the problems in understanding the application of these exclusions is that prior case law addressed each exclusion separately. That may be because each exclusion represented a type of coverage that was available separately or because the specific lawsuits that interpreted the exclusions only presented the trial court with an issue related to one specific exclusion. This becomes problematic in our present case because the exclusions are grouped together, making a hybrid “products hazard/completed operations” exclusion. Distinctions drawn in prior case law between the types of excluded claims are not relevant here because both types of exclusion are present.
¶ 12 I believe that the ruling of the trial court and affirmed by the majority contradicts the explicit directive made by our Court in Harford. Harford dealt with the application of a products hazard exclusion. A wine making supplier was sued over sulphur strips, which are (or were) used to kill bacteria in fermenting vessels. When a particular user of the strips lit one of the strips and placed it in a barrel, the strip ignited the alcohol fumes that were in the barrel and the barrel exploded, injuring several people. In the underlying lawsuit, it was claimed that the supplier negligently failed to inform the user of the dangers of using the strips. The insurer, Harford, sought to escape coverage by virtue of the products hazard exclusion. Our Court held against Harford, explaining that the products hazard exclusion dealt with strict liability claims, not negligence claims and if Harford wanted to exclude such negligence claims, it must expressly do so. Specifically, our Court stated:
The problem for Harford is that the failure to warn claims are not essentially negligence claims or essentially product liability claims. Rather, they are essentially both and fall in a region of analytical overlap between two commonly distinct theories of liability. As the drafter of the contract, it was incumbent upon Harford to expressly include that region of overlap in the exclusion if that was its intent.
Harford, 578 A.2d at 503. [Emphasis supplied.]
¶ 13 Thus, our Court instructed insurers to specifically include such negligence in the exclusion if those insurers wanted to forego covering negligent failure to warn claims. That is exactly what West American did in this case. West American took *99our Court at its word and specifically included language excluding warnings or failures to warn. Despite following our instructions, the trial court and the majority are now telling insurers that even if such claims are expressly excluded from coverage, that language means nothing. Therefore, I believe the view of the trial court and majority squarely conflicts with Harford.
¶ 14 Keystone, supra, does not change the result. Keystone found that the failure to warn meant that the installation process had not been completed, therefore the “completed operations” exclusion (which otherwise provides exclusion from coverage for negligently providing a service in conjunction with a product) could not apply. There are a number of problems with that argument, however. First and foremost is the fact that the insurance policy in Keystone did not contain the limiting language found in the instant case.19 By applying Keystone to the facts in this case, we are essentially completely ignoring the language of the insurance contract, because there is no situation where a failure to give warning exclusion would be valid. If a supplier or manufacturer fails to give a proper warning, then the product can never be completed, thus a failure to warn can never be part of a completed operations exclusion. This interpretation makes little, if any, sense after our Court specifically told insurers to include failure to warn in the exclusions.
¶ 15 Further, the rules of contract interpretation are well settled. The parties have a right to make their own contract, and it is not the function of the court to rewrite it or give it a construction in conflict with the accepted and plain meaning of the language used. Meeting House Lane, Ltd. v. Melso, 427 Pa.Super. 118, 628 A.2d 854, 857-58 (1993). One part of a contract cannot be interpreted so as to annul another part, and a contract must be construed, if possible, to give effect to all of its terms. Id. (emphasis added). When interpreting a contract of insurance it is necessary to consider the intent of the parties as manifested by the language of the instrument. Where the policy language is clear, the contract will be applied as written. Pempkowski v. State Farm Mutual Auto. Ins. Co., 451 Pa.Super. 61, 678 A.2d 398, 400 (1996) (emphasis added). When interpreting a contract, a court must construe it as written, giving effect to the clear language and plain meaning of the words. Solomon v. U.S. Healthcare Systems of Pa., 797 A.2d 346, 349 (Pa.Super.2002) (emphasis added).
¶ 16 There has been no claim and no finding that the limiting language in question, referring to warnings or failure to give warnings, is against public policy. No such argument could be credibly made given that our Court directed insurers to use the language, if they so desired. As such, the language violates no prior case law. The language violates no statutory provisions. The policy language is clear and unambiguous. Therefore, there is no reason to apply Keystone, which did not involve an interpretation of the language in question in such a manner as to ignore that language. By effectively writing the language out of the contract, we violate the standard rules of contract interpretation.
¶ 17 The application of Keystone to this situation creates a legal Catch-22. In essence we are telling insurers they can limit exposure on a CGL by specifically excluding products hazards and completed opera*100tions (and specifically failure to warn claims) and then telling insurers that they cannot enforce the exclusion because doing so indicates the operation was not completed.
¶ 18 The history of the application of the products-completed operations hazard analysis found in the trial court opinion and cited by the majority is a thorough examination of this insurance policy exclusion. I have no disagreement with the analysis, as far as it goes. However, I believe that facts found in this case compel a different result.
¶ 19 Initially, there was a distinction between an exclusion for strict liability failure to warn under Restatement, Torts § 402A and negligent failure to warn. Therefore, it was held that excluding strict liability from coverage did not exclude coverage for negligence. See Harford, supra. The situation is further complicated between the difference in coverage for the “product” and for “completed operations.”
¶20 Keystone, supra, relied upon by both the trial court and the majority, takes one view of the application of the exclusion. Basically, the ‘products’ portion of the exclusion covers the thing itself, in this case, a forklift and backup alarm. The “completed hazards” portion of the exclusion applies to the service of the thing. Often, completed hazards refer to the installation of the product.
In Friestad v. Travelers Indemnity Company, 260 Pa.Super. 178, 393 A.2d 1212 (1978), we considered an identical provision after the insured negligently installed a furnace which later caused a fire that destroyed the home in which it was installed. We distinguished the “products” and “completed operations” exclusions by stating that the “principal thrust of the completed operations is the insured’s provision of a service, while the principal thrust of the products hazard in the insured’s manufacture or sale of a product.” Id. at 1213.
Keystone, 767 A.2d at 575.
¶ 21 Sometimes an exclusion may cover a product but not “completed operations” if there is still work to be done, for example on installation. However, that is not the instant situation. Whether the accident occurred because of the product itself or because of later work, the language of the exclusion clearly provides that there is coverage for neither. In Friestad, the installation of the furnace was allegedly negligently performed, thereby causing a fire. Our Court determined that the alleged negligence fell within the ambit of the completed operations exclusion. Therefore, a negligent act does not automatically negate the application of the exclusion.
¶ 22 The cases seem to have carved out an odd distinction in determining what the “completed operations” exclusion means. Tangible negligence, such as failing to tighten a screw properly on a furnace, is excludable under the “completed operations” exclusion. See Friestad. However, intangible negligence, such as failing to warn a person not to stick his hand in a moving conveyor belt, is not excludable. See Keystone. This is a fine hair to split.
¶ 23 Apparently, at one time “products” and “completed operations” were separate exclusions. In Harford the policy only contained the “products” exclusion. In Harford, it was held that the products exclusion only covered the product itself, not the separate negligent act of failing to warn. Thus, because the policy did not address that aspect of negligence, the policy was construed against the insurer and coverage was deemed appropriate.
While we find that the instant exclusion clearly precludes claims which allege the product was defective for failing to include instructions or warnings, the best *101that can be said for this exclusion with regard to negligent failure to warn claims is that reasonable men could differ as to its proper scope.
To hold for Harford, here, would require this Court to find that a failure to warn negligence claim, which ostensibly asserts negligence in failing to reasonably instruct or warn, is essentially a product liability claim for failure to properly manufacture (with necessary warnings) and was intended to be excluded by the Products Hazard exclusion.
Harford, 578 A.2d at 502-03 (emphasis in the original).
¶ 24 It seems that in Harford, there was a distinction between exclusions for strict liability under 402A and negligence, where there was the negligent failure to warn. The distinction was necessary not because negligent acts cannot, at times, be excludable. This is proven by our Court instructing insurers to use language to specifically exclude negligent failure to warn. Rather, the specific act complained of, a failure to warn, is found in both a strict liability claim (which is properly excludable) and a negligence claim (which was not specifically alluded to in the exclusionary language). The differentiation between negligent failure to warn and strict liability failure to warn is also a fine distinction, given how close the factual predicates are for a claim of failure to warn, but the distinction was determined to be legally supportable.
¶ 25 Keystone lifted the distinction from the “products” exclusion and grafted it onto the “completed operations” exclusion. This appears to be an odd solution given that the products exclusion relied on a strict liability/negligence distinction and that distinction is not involved in the “completed operations” exclusion, which is designed to apply to the negligent performance of a service. Keystone attempts to explain the rationale by stating:
Moreover, the Eastcoast Court indicated, in dictum, that this rule applies equally to negligent misrepresentation and negligent failure to warn theories because the underlying policy is the same to prevent an insurer from refusing coverage for an injury caused by negligence at the time of installation simply because the injury cannot be discovered until the installation is complete and the installed equipment is put into service.
Keystone, 767 A.2d at 575 (citing Eastcoast Equipment Co. v. Maryland Casualty Co., 207 Pa.Super. 383, 218 A.2d 91 (1966)). What this reasoning fails to recognize is that there is no real difference between a negligent failure to warn (which is not subject to exclusion) and negligent failure to tighten a screw (which is subject to exclusion).20 Both acts of negligence may be undiscoverable until “after the installation is complete and the installed equipment is put into service.” Yet case law clearly indicates that the “tangible” act of negligence is properly addressed by the completed operations exclusion. As long as “tangible” negligence is properly ex-cludable, the fact that an act of negligence may not be immediately discoverable, the rationale relied on in Keystone, cannot logically be a differentiating factor.
¶ 26 I believe the rationale used in Keystone is fundamentally flawed and at some point should be revisited by an en banc panel of this Court or the Pennsylvania Supreme Court.21 However, even if Key*102stone remains good law, it does not apply to this case because neither Keystone, nor any other prior case cited, involves the specific exclusion found in this insurance policy. Applying Keystone to the present facts improperly conflicts with prior and settled case law. See Harford, supra; Meeting House Lane, supra; Pempkowski, supra; Solomon, supra.
¶27 Another problem with the trial court and majority solution is that it is transforming a CGL policy into another form of insurance, a products hazard and completed operations policy. There is insurance available to cover the types of negligence and strict liability claims excluded from this CGL policy. Evidence presented in this matter indicates that such products hazard and completed operations coverage was originally offered to Upright and rejected. Upright purchased that coverage after the accident that is the basis for this lawsuit.
¶ 28 The transformation of one type of coverage into another is disapproved of by our courts. In the realm of motor vehicle insurance, an insured may not convert un-derinsured motorist coverage into liability coverage. Rudloff v. Nationwide Mut. Ins. Co., 806 A.2d 1270 (Pa.Super.2002). A CGL cannot be converted into professional services coverage. Atlantic Mut. Ins. Co. v. Gula, 926 A.2d 449 (Pa.Super.2007). A CGL cannot be converted into a performance bond. Kvaerner Metals Div. of Kvaerner U.S., Inc. v. Commercial Union, 589 Pa. 317, 908 A.2d 888 (2006). Our Supreme Court stated:
Such claims simply do not present the fortuity contemplated by the ordinary definition of “accident” or its common judicial construction in this context. To hold otherwise would be to covert a policy for insurance into a performance bond. We are unwilling to do so, especially since such protections are already available for the protection of contractors.
Kvaerner, 908 A.2d at 899 (emphasis added).
Relying on that reasoning, our Court then stated:
The CGL is not meant'to insure the actual quality of the thing provided; in the construction business there are other ways for that, notably a performance bond. Here, the CGL is not meant to be converted into professional negligence insurance. The CGL policy is not designed to insure the quality of the case management service provided. The record demonstrates that there is specific insurance that does cover the possibility of negligent provision of case management. Novaeon has such coverage from St. Paul Fire and Marine. Unfortunately, that coverage was not purchased for the time frame applicable to this claim. The fact that Novaeon did not have the proper coverage in place for this claim does not mean that any other policy that covered Novaeon must provide coverage for professional negligence. The person who suffers a loss due to a house fire cannot convert his automobile insurance into a homeowner’s policy to pay for the repairs to his home.
Atlantic Mutual, 926 A.2d at 453.
¶ 29 Coverage for product hazards and completed operations was available to Upright. It chose not to purchase that cover*103age. This is the same scenario presented in Atlantic Mutual. Upright, and by extension Bombar, who obtained bad faith rights from Upright, cannot be heard to complain that West American acted badly in failing to cover a claim for which Upright bought no coverage. Further, Upright/Bombar are not allowed to convert a general liability policy into specific coverage for products hazards and completed operations.
¶30 I believe that: (a) our Court in Harford, supra, specifically instructed insurers that if they wanted to exclude failure to warn and like claims from CGL coverage, to use the exclusionary language used in the instant policy; (b) we are bound to give effect to language in an insurance contract as long as that language does not violate public policy, statutory requirements or prior case law; (c) this exclusionary language does not violate public policy, statutory requirements or prior case law; and (d) Keystone did not address the specific language of this policy and does not require a different result. Therefore, to rule as the trial court and the majority would, in effect, judicially negate otherwise proper contract language. I would find the exclusionary language in question to be valid and enforceable. Because I believe the exclusion to be enforceable, I would find West American did not act in bad faith in refusing to defend and indemnify the underlying claim.
¶ 31 Therefore, I dissent.
. “Harford” is not a typographical error. There is a "Harford Mutual Insurance Company” as well as a "Hartford Mutual Insurance Company.”
. There is no requirement that these types of claims be included from a CGL policy nor is there a requirement that the claims be excluded. It is a question of policy language. Also, it appears to be a method of allowing insureds to customize their insurance policies to fit their particular needs. Just as an automobile owner can purchase limited tort coverage, underinsured or uninsured motorist coverage to fit specific needs and budget, the business owner can choose to purchase separate coverage for product liability claims and/or completed operations claims.
. A detailed analysis of the case law cited here is provided in the next section.
. Other problems with applying Keystone to this factual situation will be discussed later in this dissent.
. It also ignores the fact the original reason for the distinction was because of the overlap of failure to warn in strict liability and negligence. This overlap does not and cannot exist between negligence and negligence.
. It is possible that the result of Keystone is proper and merely the rationale is incorrect. *102If Keystone is viewed as the flip side of Har-ford, then coverage may well have been appropriate. By this I mean, Harford did not specifically exclude the negligent failure to warn in a strict liability exclusion. In Keystone, the real problem might have been the failure in include strict liability language in the negligence driven completed operations exclusion.